diff options
Diffstat (limited to 'debian/additions/innotop')
| -rw-r--r-- | debian/additions/innotop/changelog.innotop | 441 | ||||
| -rw-r--r-- | debian/additions/innotop/innotop | 12250 | ||||
| -rw-r--r-- | debian/additions/innotop/innotop.1 | 2200 |
3 files changed, 14891 insertions, 0 deletions
diff --git a/debian/additions/innotop/changelog.innotop b/debian/additions/innotop/changelog.innotop new file mode 100644 index 00000000..18165952 --- /dev/null +++ b/debian/additions/innotop/changelog.innotop @@ -0,0 +1,441 @@ +Changelog for innotop: + +2017-01-20: version 1.11.4 + * add SUM function for ONLY_FULL_GROUP_BY + +2017-01-20: version 1.11.3 + + * Undisplay handlersocket's threads in hide_inactive + * fix runtime error regarding redundant sprintf argument #122 + * added sort on connection-name in display M, after sql/io running, seconds behind master and channel_name + * fixed bug that removed value for cxn and channel_name columns in display M + * added sort on replication delay, so that the replication-display will sort on slave_sql_running, timelag (in minutes) and channel_name. + * support for MariaDB 10.0 in InnoDB row (issue 93) + +2013-07-12: version 1.9.1 + + Bugs fixed: + * Support of MySQL 5.6 was broken on some pages (issue 82, 83) + * Deadlock clearing transactions is now not included in binary log + (issue 84) + * New spec file with requirements and build requirements for + CentOS/RHEL and Fedora + +2012-09-07: version 1.9.0 + + Changes: + * A new Health Dashboard (A) mode is the default mode. + * Added a new InnoDB Locked (K) mode. + * Added a new 'spark' config variable for sparklines. + * Added a new fuzzy_time formatting function. + * Added "query distill" summarizing. + * Handled more types of errors connecting to the server. + * Displayed some data more compactly. + + Bugs fixed: + * Double-quotes were used to terminate strings in SQL (issue 57). + * T mode didn't show InnoDB transaction times (issue 67). + * Killing a query didn't suggest the longest-running one automatically. + * Connections weren't closed on exit (issue 64). + * Q mode didn't have connections in its header (issue 63). + * Connections and server groups were poorly handled (issue 68). + * The RPM spec file was buggy (issue 59). + * Event filters were defined wrong (issue 54). + +2012-02-25: version 1.8.1 + + Bugs fixed: + * Various parsing errors with MySQL 5.5 (issue 23, 45, 47, 51, 52, 53). + * RPM spec file prevented building on CentOS 5.5 using mock (issue 44). + * Tests worked only from the test subdirectory (issue 43). + +2010-11-06: version 1.8.0 + + Changes: + * Don't re-fetch SHOW VARIABLES every iteration; it's too slow on many hosts. + * Add a filter to remove EVENT threads in SHOW PROCESSLIST (issue 32). + * Add a timestamp to output in -n mode, when -t is specified (issue 37). + * Add a new U mode, for Percona/MariaDB USER_STATISTICS (issue 39). + * Add support for millisecond query time in Percona Server (issue 39). + * Display a summary of queries executed in Query List mode (issue 26). + + Bugs fixed: + * Made config-file reading more robust (issue 41). + * Hostname parsing wasn't standards compliant (issue 30). + * MKDEBUG didn't work on some Perl versions (issue 22). + * Don't try to get InnoDB status if have_innodb != YES (issue 33). + * Status text from the InnoDB plugin wasn't parsed correctly (issue 36). + * Transaction ID from InnoDB plugin wasn't subtracted correctly (issue 38). + * Switching modes and pressing ? for help caused a crash (issue 40). + +2009-09-06: version 1.7.2 + + Changes: + * add support for --socket + + Bugs fixed: + * remove cxn from $meta->{group_by} if there's only one connection displayed + * fix for issue 19 - cxn column won't become visible when viewing two + connections after having viewed one connection + * suppress errors resulting from the addition of a 'BACKGROUND THREAD' + section in the output of 'show innodb status' + * possible fix for issue 22 - Useless use of a constant in void context + * small change to set_to_tbl() around hiding the cxn column if there + aren't two or more connections + +2009-03-09: version 1.7.1 + + Changes: + * Don't display the CXN column if only one connection is active in + the current view + * the 'state' column is now visible by default in Query List mode + + Bugs fixed: + * fixed bug where trying to aggregate the time column would result + in a crash if the time column had an undef value in it, which is + the case when a thread is in the 'Connect' state + * updated innotop.spec file to reflect current version + +2009-02-23: version 1.7.0 + + Changes: + * supports a central config (/etc/innotop/innotop.conf) + * changed the default home directory config to ~/.innotop/innotop.conf + (away from .ini) + * embedded InnoDBParser.pm into innotop so it can be run with no + installation + * no longer writes a new config file by default + * added --skipcentral (skip reading central config) and --write (write + a config if none were loaded at start-up) + * if no config file is loaded, connect to a MySQL database on + localhost using mysql_read_default_group=client + * embedded maatkit's DSNParser.pm and added support for --user, + --password, --host, --port + * changed default mode from T (InnoDB Transactions) to Q (Query List) + * in addition to connected threads, now displays running and cached + threads in statusbar + * don't load connections from a config file if any DSN information or + a username or password is specified on the command-line + + Bugs fixed: + * fixed bug preventing utilization of command-line options that + override default config settings if no config file was loaded + * fixed a bug where migrating from an old version of the config will + delete ~/innotop.ini, if it exists. Now uses File::Temp::tempfile(). + +2007-11-09: version 1.6.0 + + * S mode crashed on non-numeric values. + * New user-defined columns crashed upon restart. + * Added --color option to control terminal coloring. + +2007-09-18: version 1.5.2 + + * Added the ability to monitor InnoDB status from a file. + * Changed W mode to L mode; it monitors all locks, not just lock waits. + +2007-09-16: version 1.5.1 + + * Added C (Command Summary) mode. + * Fixed a bug in the 'avg' aggregate function. + +2007-09-10: version 1.5.0 + + Changes: + * Added plugin functionality. + * Added group-by functionality. + * Moved the configuration file to a directory. + * Enhanced filtering and sorting on pivoted tables. + * Many small bug fixes. + +2007-07-16: version 1.4.3 + + Changes: + * Added standard --version command-line option + * Changed colors to cyan instead of blue; more visible on dark terminals. + * Added information to the filter-choosing dialog. + * Added column auto-completion when entering a filter expression. + * Changed Term::ReadKey from optional to mandatory. + * Clarified username in password prompting. + * Ten thousand words of documentation! + + Bugs fixed: + * innotop crashed in W mode when InnoDB status data was truncated. + * innotop didn't display errors in tables if debug was enabled. + * The colored() subroutine wasn't being created in non-interactive mode. + * Don't prompt to save password except the first time. + +2007-05-03: version 1.4.2 + + This version contains all changes to the trunk until revision 239; some + changes in revisions 240:250 are included. + + MAJOR CHANGES: + + * Quick-filters to easily filter any column in any display + * Compatibility with MySQL 3.23 through 6.0 + * Improved error handling when a server is down, permissions denied, etc + * Use additional SHOW INNODB STATUS information in 5.1.x + * Make all modes use tables consistently, so they can all be edited, + filtered, colored and sorted consistently + * Combine V, G and S modes into S mode, with v, g, and s hot-keys + * Let DBD driver read MySQL option files; permit connections without + user/pass/etc + * Compile SQL-like expressions into Perl subroutines; eliminate need to + know Perl + * Do not save all config data to config file, only save user's customizations + * Rewritten and improved command-line option handling + * Added --count, --delay, and other command-line options to support + run-and-exit operation + * Improve built-in variable sets + * Improve help screen with three-part balanced-column layout + * Simplify table-editor and improve hotkey support + * Require Perl to have high-resolution time support (Time::HiRes) + * Help the user choose a query to analyze or kill + * Enable EXPLAIN, show-full-query in T mode just like Q mode + * Let data-extraction access current, previous and incremental data sets + all at once + + MINOR CHANGES: + + * Column stabilizing for Q mode + * New color rules for T, Q, W modes + * Apply slave I/O filter to Q mode + * Improve detection of server version and other meta-data + * Make connection timeout a config variable + * Improve cross-version-compatible SQL syntax + * Get some information from the DBD driver instead of asking MySQL for it + * Improved error messages + * Improve server group creation/editing + * Improve connection/thread killing + * Fix broken key bindings and restore previously mapped hot-keys for + choosing columns + * Some documentation updates (but not nearly enough) + * Allow the user to specify graphing char in S mode (formerly G mode) + * Allow easy switching between variable sets in S mode + * Bind 'n' key globally to choose the 'next' server connection + * Bind '%' key globally to filter displayed tables + * Allow aligning columns on the decimal place for easy readability + * Add hide_hdr config variable to hide column headers in tables + * Add a feature to smartly run PURGE MASTER LOGS in Replication mode + * Enable debug mode as a globally configurable variable + * Improve error messages when an expression or filter doesn't compile or has + a run-time error; die on error when debug is enabled + * Allow user-configurable delays after executing SQL (to let the server + settle down before taking another measurement) + * Add an expression to show how long until a transaction is finished + * Add skip_innodb as a global config variable + * Add '%' after percentages to help disambiguate (user-configurable) + * Add column to M mode to help see how fast slave is catching up to master + + BUG FIXES: + + * T and W modes had wrong value for wait_status column + * Error tracking on connections didn't reset when the connection recovered + * wait_timeout on connections couldn't be set before MySQL 4.0.3 + * There was a crash on 3.23 when wiping deadlocks + * Lettercase changes in some result sets (SHOW MASTER/SLAVE STATUS) between + MySQL versions crashed innotop + * Inactive connections crashed innotop upon access to DBD driver + * set_precision did not respect user defaults for number of digits + * --inc command-line option could not be negated + * InnoDB status parsing was not always parsing all needed information + * S mode (formerly G mode) could crash trying to divide non-numeric data + * M table didn't show Slave_open_temp_tables variable; incorrect lettercase + * DBD drivers with broken AutoCommit would crash innotop + * Some key bindings had incorrect labels + * Some config-file loading routines could load data for things that didn't + exist + * Headers printed too often in S mode + * High-resolution time was not used even when the user had it + * Non-interactive mode printed blank lines sometimes + * Q-mode header and statusbar showed different QPS numbers + * Formulas for key-cache and query-cache hit ratios were wrong + * Mac OS "Darwin" machines were mis-identified as Microsoft Windows + * Some multiplications crashed when given undefined input + * The commify transformation did not check its input and could crash + * Specifying an invalid mode on the command line or config file could crash + innotop + +2007-03-29: version 1.4.1 + + * More tweaks to display of connection errors. + * Fixed a problem with skip-innodb in MySQL 5.1. + * Fix a bug with dead connections in single-connection mode. + * Fix a regex to allow parsing more data from truncated deadlocks. + * Don't load active cxns from the config file if the cxn isn't defined. + +2007-03-03: version 1.4.0 + + * Further tweak error handling and display of connection errors + * More centralization of querying + * Fix forking so it doesn't kill all database connections + * Allow user to run innotop without permissions for GLOBAL variables and status + +2007-02-11: version 1.3.6 + + * Handle some connection failures so innotop doesn't crash because of one server. + * Enable incremental display in more modes. + * Tweaks to colorizing, color editor, and default color rules. + * Tweaks to default sorting rules. + * Use prepared statements for efficiency. + * Bug fixes and code cleanups. + * Data storage is keyed on clock ticks now. + +2007-02-03: version 1.3.5 + + * Bug fixes. + * More tools for editing configuration from within innotop. + * Filters and transformations are constrained to valid values. + * Support for colorizing rows. + * Sorting by multiple columns. + * Compress headers when display is very wide. + * Stabilize and limit column widths. + * Check config file formats when upgrading so upgrades go smoothly. + * Make D mode handle many connections at once. + * Extract simple expressions from data sets in column src property. + This makes innotop more awk-ish. + +2007-01-16: version 1.3 + + * Readline support. + * Can be used unattended, or in a pipe-and-filter mode + where it outputs tab-separated data to standard output. + * You can specify a config file on the command line. + Config files can be marked read-only. + * Monitor multiple servers simultaneously. + * Server groups to help manage many servers conveniently. + * Monitor master/slave status, and control slaves. + * Columns can have user-defined expressions as their data sources. + * Better configuration tools. + * InnoDB status information is merged into SHOW VARIABLES and + SHOW STATUS information, so you can access it all together. + * High-precision time support in more places. + * Lots of tweaks to make things display more readably and compactly. + * Column transformations and filters. + +2007-01-16: version 1.0.1 + * NOTE: innotop is now hosted at Sourceforge, in Subversion not CVS. + The new project homepage is http://sourceforge.net/projects/innotop/ + * Tweak default T/Q mode sort columns to match what people expect. + * Fix broken InnoDBParser.pm documentation (and hence man page). + +2007-01-06: version 1.0 + * NOTE: innotop is now hosted at Sourceforge, in Subversion not CVS. + The new project homepage is http://sourceforge.net/projects/innotop/ + * Prevent control characters from freaking terminal out. + * Set timeout to keep busy servers from closing connection. + * There is only one InnoDB insert buffer. + * Make licenses clear and consistent. + +2006-11-14: innotop 0.1.160, InnoDBParser version 1.69 + * Support for ANSI color on Microsoft Windows (more readable, compact + display; thanks Gisbert W. Selke). + * Better handling of $ENV{HOME} on Windows. + * Added a LICENSE file to the package as per Gentoo bug: + http://bugs.gentoo.org/show_bug.cgi?id=147600 + +2006-11-11: innotop 0.1.157, InnoDBParser version 1.69 + * Add Microsoft Windows support. + +2006-10-19: innotop 0.1.154, InnoDBParser version 1.69 + * Add O (Open Tables) mode + * Add some more checks to handle incomplete InnoDB status information + +2006-09-30: innotop 0.1.152, InnoDBParser version 1.69 + * Figured out what was wrong with package $VERSION variable: it wasn't + after the package declaration! + +2006-09-28: innotop 0.1.152, InnoDBParser version 1.67 + * Make more efforts towards crash-resistance and tolerance of completely + messed-up inputs. If innotop itself is broken, it is now much harder to + tell, because it just keeps on running without complaining. + * Fix a small bug parsing out some information and displaying it. + +2006-09-05: innotop 0.1.149, InnoDBParser version 1.64 + * Try to find and eliminate any parsing code that assumes pattern matches + will succeed. + +2006-09-05: innotop 0.1.149, InnoDBParser version 1.62 + * Make innotop crash-resistant, so I can declare it STABLE finally. + * Instead of using SQL conditional comments, detect MySQL version. + +2006-08-22: innotop 0.1.147, InnoDBParser version 1.60 + * Fix some innotop bugs with undefined values, bad formatting etc. + +2006-08-19: innotop 0.1.146, InnoDBParser version 1.60 + * Make innotop handle some unexpected NULL values in Q mode. + * Add OS wait information to W mode, so it is now "everything that waits." + * Center section captions better. + * Make R mode more readable and compact. + * Make InnoDBParser parse lock waits even when they've been waiting 0 secs. + +2006-08-12: innotop 0.1.139, InnoDBParser version 1.59 + * Add more documentation + * Tweak V mode to show more info in less space. + * Fix a bug in G mode. + +2006-08-10: innotop 0.1.132, InnoDBParser version 1.58 + * Handle yet more types of FK error... it will never end! + * Handle some special cases when DEADLOCK info truncated + * Add a bit more FK info to F mode in innotop + * More tests added to the test suite + +2006-08-07: innotop 0.1.131, InnoDBParser version 1.55 + * Fix another issue with configuration + * Handle another type of FK error + +2006-08-03: innotop 0.1.130, InnoDBParser version 1.54 + * Fix an issue loading config file + * Add heap_no to 'D' (InnoDB Deadlock) mode to ease deadlock debugging. + +2006-08-02: innotop 0.1.128, InnoDBParser version 1.54 + * Parse lock wait information from the TRANSACTION section. + * Even more OS-specific parsing... pain in the butt... + * Add 'W' (InnoDB Lock Wait) mode. + * Fix some minor display issues with statusbar. + +2006-08-02: innotop 0.1.125, InnoDBParser version 1.50 + * Don't try to get references to Perl built-in functions like time() + * Handle more OS-specific variations of InnoDB status text + * Add some more information to various places in innotop + +2006-08-01: innotop 0.1.123, InnoDBParser version 1.47 + + * Enhance S and G modes: clear screen and re-print headers + * Don't crash when deadlock data is truncated + * Make Analyze mode say how to get back to whatever you came from + * Display 'nothing to display' when there is nothing + * Add ability to read InnoDB status text from a file (mostly helps test) + * Add table of Wait Array Information in Row Op/Semaphore mode + * Add table of lock information in InnoDB deadlock mode + * Ensure new features in upgrades don't get masked by existing config files + * Tweak default column choices for T mode + * Enhance foreign key parsing + * Enhance physical record and data tuple parsing + * Enhance lock parsing (handle old-style and new-style formats) + +2006-07-24: innotop 0.1.112, InnoDBParser version 1.36 + + * InnoDBParser enhancements for FK error messages. + * A fix to innotop to prevent it from crashing while trying to display a FK + error message. + * Some minor cosmetic changes to number formatting in innotop. + +2006-07-22: innotop 0.1.106, InnoDBParser version 1.35 + + * InnoDBParser is much more complete and accurate. + * Tons of bug fixes. + * Add partitions to EXPLAIN mode. + * Enhance Q mode header, add T mode header. + * Share some configuration variables across modes. + * Add formatted time columns to Q, T modes. + * Add command-line argument parsing. + * Turn off echo when asking for password. + * Add option to specify port when connecting. + * Let display-optimized-query display multiple notes. + * Lots of small improvements, such as showing more info in statusbar. + +2006-07-02: innotop 0.1.74, InnoDBParser version 1.24 + + * Initial release for public consumption. diff --git a/debian/additions/innotop/innotop b/debian/additions/innotop/innotop new file mode 100644 index 00000000..2bc09091 --- /dev/null +++ b/debian/additions/innotop/innotop @@ -0,0 +1,12250 @@ +#!/usr/bin/perl + +# vim: tw=160:nowrap:expandtab:tabstop=3:shiftwidth=3:softtabstop=3 + +# This program is copyright (c) 2006 Baron Schwartz, baron at xaprb dot com. +# Maintainers since 2013 : Kenny Gryp - Frédéric Descamps +# Feedback and improvements are gratefully received. +# +# THIS PROGRAM IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED +# WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF +# MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. +# +# This program is free software; you can redistribute it and/or modify it under +# the terms of the GNU General Public License as published by the Free Software +# Foundation, version 2; OR the Perl Artistic License. On UNIX and similar +# systems, you can issue `man perlgpl' or `man perlartistic' to read these + +# You should have received a copy of the GNU General Public License along with +# this program; if not, write to the Free Software Foundation, Inc., 51 Franklin +# Street, Fifth Floor, Boston, MA 02110-1335 USA + +use strict; +use warnings FATAL => 'all'; + +our $VERSION = '1.11.4'; + +# Find the home directory; it's different on different OSes. +our $homepath = $ENV{HOME} || $ENV{HOMEPATH} || $ENV{USERPROFILE} || '.'; + +# Configuration files +our $default_home_conf = "$homepath/.innotop/innotop.conf"; +our $default_central_conf = "/etc/innotop/innotop.conf"; +our $conf_file = ""; + +## Begin packages ## + +package DSNParser; + +use DBI; +use Data::Dumper; +$Data::Dumper::Indent = 0; +$Data::Dumper::Quotekeys = 0; +use English qw(-no_match_vars); + +use constant MKDEBUG => $ENV{MKDEBUG} || 0; + +# Defaults are built-in, but you can add/replace items by passing them as +# hashrefs of {key, desc, copy, dsn}. The desc and dsn items are optional. +# You can set properties with the prop() sub. Don't set the 'opts' property. +sub new { + my ( $class, @opts ) = @_; + my $self = { + opts => { + A => { + desc => 'Default character set', + dsn => 'charset', + copy => 1, + }, + D => { + desc => 'Database to use', + dsn => 'database', + copy => 1, + }, + F => { + desc => 'Only read default options from the given file', + dsn => 'mariadb_read_default_file', + copy => 1, + }, + h => { + desc => 'Connect to host', + dsn => 'host', + copy => 1, + }, + p => { + desc => 'Password to use when connecting', + dsn => 'password', + copy => 1, + }, + P => { + desc => 'Port number to use for connection', + dsn => 'port', + copy => 1, + }, + S => { + desc => 'Socket file to use for connection', + dsn => 'mariadb_socket', + copy => 1, + }, + u => { + desc => 'User for login if not current user', + dsn => 'user', + copy => 1, + }, + }, + }; + foreach my $opt ( @opts ) { + if (MKDEBUG) { + _d('Adding extra property ' . $opt->{key}); + } + $self->{opts}->{$opt->{key}} = { desc => $opt->{desc}, copy => $opt->{copy} }; + } + return bless $self, $class; +} + +# Recognized properties: +# * autokey: which key to treat a bareword as (typically h=host). +# * dbidriver: which DBI driver to use; assumes mysql, supports Pg. +# * required: which parts are required (hashref). +# * setvars: a list of variables to set after connecting +sub prop { + my ( $self, $prop, $value ) = @_; + if ( @_ > 2 ) { + MKDEBUG && _d("Setting $prop property"); + $self->{$prop} = $value; + } + return $self->{$prop}; +} + +sub parse { + my ( $self, $dsn, $prev, $defaults ) = @_; + if ( !$dsn ) { + MKDEBUG && _d('No DSN to parse'); + return; + } + MKDEBUG && _d("Parsing $dsn"); + $prev ||= {}; + $defaults ||= {}; + my %given_props; + my %final_props; + my %opts = %{$self->{opts}}; + my $prop_autokey = $self->prop('autokey'); + + # Parse given props + foreach my $dsn_part ( split(/,/, $dsn) ) { + if ( my ($prop_key, $prop_val) = $dsn_part =~ m/^(.)=(.*)$/ ) { + # Handle the typical DSN parts like h=host, P=3306, etc. + $given_props{$prop_key} = $prop_val; + } + elsif ( $prop_autokey ) { + # Handle barewords + MKDEBUG && _d("Interpreting $dsn_part as $prop_autokey=$dsn_part"); + $given_props{$prop_autokey} = $dsn_part; + } + else { + MKDEBUG && _d("Bad DSN part: $dsn_part"); + } + } + + # Fill in final props from given, previous, and/or default props + foreach my $key ( keys %opts ) { + MKDEBUG && _d("Finding value for $key"); + $final_props{$key} = $given_props{$key}; + if ( !defined $final_props{$key} + && defined $prev->{$key} && $opts{$key}->{copy} ) + { + $final_props{$key} = $prev->{$key}; + MKDEBUG && _d("Copying value for $key from previous DSN"); + } + if ( !defined $final_props{$key} ) { + $final_props{$key} = $defaults->{$key}; + MKDEBUG && _d("Copying value for $key from defaults"); + } + } + + # Sanity check props + foreach my $key ( keys %given_props ) { + die "Unrecognized DSN part '$key' in '$dsn'\n" + unless exists $opts{$key}; + } + if ( (my $required = $self->prop('required')) ) { + foreach my $key ( keys %$required ) { + die "Missing DSN part '$key' in '$dsn'\n" unless $final_props{$key}; + } + } + + return \%final_props; +} + +sub as_string { + my ( $self, $dsn ) = @_; + return $dsn unless ref $dsn; + return join(',', + map { "$_=" . ($_ eq 'p' ? '...' : $dsn->{$_}) } + grep { defined $dsn->{$_} && $self->{opts}->{$_} } + sort keys %$dsn ); +} + +sub usage { + my ( $self ) = @_; + my $usage + = "DSN syntax is key=value[,key=value...] Allowable DSN keys:\n" + . " KEY COPY MEANING\n" + . " === ==== =============================================\n"; + my %opts = %{$self->{opts}}; + foreach my $key ( sort keys %opts ) { + $usage .= " $key " + . ($opts{$key}->{copy} ? 'yes ' : 'no ') + . ($opts{$key}->{desc} || '[No description]') + . "\n"; + } + if ( (my $key = $self->prop('autokey')) ) { + $usage .= " If the DSN is a bareword, the word is treated as the '$key' key.\n"; + } + return $usage; +} + +# Supports PostgreSQL via the dbidriver element of $info, but assumes MySQL by +# default. +sub get_cxn_params { + my ( $self, $info ) = @_; + my $dsn; + my %opts = %{$self->{opts}}; + my $driver = $self->prop('dbidriver') || ''; + if ( $driver eq 'Pg' ) { + $dsn = 'DBI:Pg:dbname=' . ( $info->{D} || '' ) . ';' + . join(';', map { "$opts{$_}->{dsn}=$info->{$_}" } + grep { defined $info->{$_} } + qw(h P)); + } + else { + $dsn = 'DBI:MariaDB:' . ( $info->{D} || '' ) . ';' + . join(';', map { "$opts{$_}->{dsn}=$info->{$_}" } + grep { defined $info->{$_} } + qw(F h P S A)) + . ';mariadb_read_default_group=client'; + } + MKDEBUG && _d($dsn); + return ($dsn, $info->{u}, $info->{p}); +} + + +# Fills in missing info from a DSN after successfully connecting to the server. +sub fill_in_dsn { + my ( $self, $dbh, $dsn ) = @_; + my $vars = $dbh->selectall_hashref('SHOW VARIABLES', 'Variable_name'); + my ($user, $db) = $dbh->selectrow_array('SELECT USER(), DATABASE()'); + $user =~ s/@.*//; + $dsn->{h} ||= $vars->{hostname}->{Value}; + $dsn->{S} ||= $vars->{'socket'}->{Value}; + $dsn->{P} ||= $vars->{port}->{Value}; + $dsn->{u} ||= $user; + $dsn->{D} ||= $db; +} + +sub get_dbh { + my ( $self, $cxn_string, $user, $pass, $opts ) = @_; + $opts ||= {}; + my $defaults = { + AutoCommit => 0, + RaiseError => 1, + PrintError => 0, + }; + @{$defaults}{ keys %$opts } = values %$opts; + my $dbh; + my $tries = 2; + while ( !$dbh && $tries-- ) { + eval { + MKDEBUG && _d($cxn_string, ' ', $user, ' ', $pass, ' {', + join(', ', map { "$_=>$defaults->{$_}" } keys %$defaults ), '}'); + $dbh = DBI->connect($cxn_string, $user, $pass, $defaults); + # Immediately set character set and binmode on STDOUT. + if ( my ($charset) = $cxn_string =~ m/charset=(\w+)/ ) { + my $sql = "/*!40101 SET NAMES $charset*/"; + MKDEBUG && _d("$dbh: $sql"); + $dbh->do($sql); + MKDEBUG && _d('Enabling charset for STDOUT'); + if ( $charset eq 'utf8' ) { + binmode(STDOUT, ':utf8') + or die "Can't binmode(STDOUT, ':utf8'): $OS_ERROR"; + } + else { + binmode(STDOUT) or die "Can't binmode(STDOUT): $OS_ERROR"; + } + } + }; + if ( !$dbh && $EVAL_ERROR ) { + MKDEBUG && _d($EVAL_ERROR); + if ( !$tries ) { + die $EVAL_ERROR; + } + } + } + # If setvars exists and it's MySQL connection, set them + my $setvars = $self->prop('setvars'); + if ( $cxn_string =~ m/mysql/i && $setvars ) { + my $sql = "SET $setvars"; + MKDEBUG && _d("$dbh: $sql"); + eval { + $dbh->do($sql); + }; + if ( $EVAL_ERROR ) { + MKDEBUG && _d($EVAL_ERROR); + } + } + MKDEBUG && _d('DBH info: ', + $dbh, + Dumper($dbh->selectrow_hashref( + 'SELECT DATABASE(), CONNECTION_ID(), VERSION()/*!50038 , @@hostname*/')), + ' Connection info: ', ($dbh->{mariadb_hostinfo} || 'undef'), + ' Character set info: ', + Dumper($dbh->selectall_arrayref( + 'SHOW VARIABLES LIKE "character_set%"', { Slice => {}})), + ' $DBD::MariaDB::VERSION: ', $DBD::MariaDB::VERSION, + ' $DBI::VERSION: ', $DBI::VERSION, + ); + return $dbh; +} + +# Tries to figure out a hostname for the connection. +sub get_hostname { + my ( $self, $dbh ) = @_; + if ( my ($host) = ($dbh->{mariadb_hostinfo} || '') =~ m/^(\w+) via/ ) { + return $host; + } + my ( $hostname, $one ) = $dbh->selectrow_array( + 'SELECT /*!50038 @@hostname, */ 1'); + return $hostname; +} + +# Disconnects a database handle, but complains verbosely if there are any active +# children. These are usually $sth handles that haven't been finish()ed. +sub disconnect { + my ( $self, $dbh ) = @_; + MKDEBUG && $self->print_active_handles($dbh); + $dbh->disconnect; +} + +sub print_active_handles { + my ( $self, $thing, $level ) = @_; + $level ||= 0; + printf("# Active %sh: %s %s %s\n", ($thing->{Type} || 'undef'), "\t" x $level, + $thing, (($thing->{Type} || '') eq 'st' ? $thing->{Statement} || '' : '')) + or die "Cannot print: $OS_ERROR"; + foreach my $handle ( grep {defined} @{ $thing->{ChildHandles} } ) { + $self->print_active_handles( $handle, $level + 1 ); + } +} + +sub _d { + my ($package, undef, $line) = caller 0; + @_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; } + map { defined $_ ? $_ : 'undef' } + @_; + # Use $$ instead of $PID in case the package + # does not use English. + print "# $package:$line $$ ", @_, "\n"; +} + +1; + +package InnoDBParser; + +use Data::Dumper; +$Data::Dumper::Sortkeys = 1; +use English qw(-no_match_vars); +use List::Util qw(max); +use POSIX qw(strftime); + +# Some common patterns +my $d = qr/(\d+)/; # Digit +my $f = qr/(\d+\.\d+)/; # Float +my $t = qr/((?:\d+ \d+)|(?:[A-Fa-f0-9]+))/; # Transaction ID +my $i = qr/((?:\d{1,3}\.){3}\d+)/; # IP address +my $n = qr/([^`\s]+)/; # MySQL object name +my $w = qr/(\w+)/; # Words +my $fl = qr/([\w\.\/]+) line $d/; # Filename and line number +my $h = qr/((?:0x)?[0-9a-f]*)/; # Hex +my $s = qr/(\d{6} .?\d:\d\d:\d\d)/; # InnoDB timestamp + +# If you update this variable, also update the SYNOPSIS in the pod. +my %innodb_section_headers = ( + "TRANSACTIONS" => "tx", + "BUFFER POOL AND MEMORY" => "bp", + "SEMAPHORES" => "sm", + "LOG" => "lg", + "ROW OPERATIONS" => "ro", + "INSERT BUFFER AND ADAPTIVE HASH INDEX" => "ib", + "FILE I/O" => "io", + "LATEST DETECTED DEADLOCK" => "dl", + "LATEST FOREIGN KEY ERROR" => "fk", + "BACKGROUND THREAD" => "bt", +); + +my %parser_for = ( + tx => \&parse_tx_section, + bp => \&parse_bp_section, + sm => \&parse_sm_section, + lg => \&parse_lg_section, + ro => \&parse_ro_section, + ib => \&parse_ib_section, + io => \&parse_io_section, + dl => \&parse_dl_section, + fk => \&parse_fk_section, +); + +my %fk_parser_for = ( + Transaction => \&parse_fk_transaction_error, + Error => \&parse_fk_bad_constraint_error, + Cannot => \&parse_fk_cant_drop_parent_error, +); + +# A thread's proc_info can be at least 98 different things I've found in the +# source. Fortunately, most of them begin with a gerunded verb. These are +# the ones that don't. +my %is_proc_info = ( + 'After create' => 1, + 'Execution of init_command' => 1, + 'FULLTEXT initialization' => 1, + 'Reopen tables' => 1, + 'Repair done' => 1, + 'Repair with keycache' => 1, + 'System lock' => 1, + 'Table lock' => 1, + 'Thread initialized' => 1, + 'User lock' => 1, + 'copy to tmp table' => 1, + 'discard_or_import_tablespace' => 1, + 'end' => 1, + 'got handler lock' => 1, + 'got old table' => 1, + 'init' => 1, + 'key cache' => 1, + 'locks' => 1, + 'malloc' => 1, + 'query end' => 1, + 'rename result table' => 1, + 'rename' => 1, + 'setup' => 1, + 'statistics' => 1, + 'status' => 1, + 'table cache' => 1, + 'update' => 1, +); + +sub new { + bless {}, shift; +} + +# Parse the status and return it. +# See srv_printf_innodb_monitor in innobase/srv/srv0srv.c +# Pass in the text to parse, whether to be in debugging mode, which sections +# to parse (hashref; if empty, parse all), and whether to parse full info from +# locks and such (probably shouldn't unless you need to). +sub parse_status_text { + my ( $self, $fulltext, $debug, $sections, $full, $mysqlversion ) = @_; + + die "I can't parse undef" unless defined $fulltext; + $fulltext =~ s/[\r\n]+/\n/g; + + $sections ||= {}; + die '$sections must be a hashref' unless ref($sections) eq 'HASH'; + + my %innodb_data = ( + got_all => 0, # Whether I was able to get the whole thing + ts => '', # Timestamp the server put on it + last_secs => 0, # Num seconds the averages are over + sections => {}, # Parsed values from each section + ); + + if ( $debug ) { + $innodb_data{'fulltext'} = $fulltext; + } + + # Get the most basic info about the status: beginning and end, and whether + # I got the whole thing (if there has been a big deadlock and there are + # too many locks to print, the output might be truncated) + + my $time_text; + if ( ($mysqlversion =~ /^5\.[67]\./) || ($mysqlversion =~ /^10\.[0-9]\./) ) { + ( $time_text ) = $fulltext =~ m/^([0-9-]* [0-9:]*) [0-9a-fx]* INNODB MONITOR OUTPUT/m; + $innodb_data{'ts'} = [ parse_innodb_timestamp_56( $time_text ) ]; + } else { + ( $time_text ) = $fulltext =~ m/^$s INNODB MONITOR OUTPUT$/m; + $innodb_data{'ts'} = [ parse_innodb_timestamp( $time_text ) ]; + } + + $innodb_data{'timestring'} = ts_to_string($innodb_data{'ts'}); + ( $innodb_data{'last_secs'} ) = $fulltext + =~ m/Per second averages calculated from the last $d seconds/; + + ( my $got_all ) = $fulltext =~ m/END OF INNODB MONITOR OUTPUT/; + $innodb_data{'got_all'} = $got_all || 0; + + # Split it into sections. Each section begins with + # ----- + # LABEL + # ----- + my %innodb_sections; + my @matches = $fulltext + =~ m#\n(---+)\n([A-Z /]+)\n\1\n(.*?)(?=\n(---+)\n[A-Z /]+\n\4\n|$)#gs; + while ( my ( $start, $name, $text, $end ) = splice(@matches, 0, 4) ) { + $innodb_sections{$name} = [ $text, $end ? 1 : 0 ]; + } + + # Just for sanity's sake, make sure I understand what to do with each + # section. + eval { + foreach my $section ( keys %innodb_sections ) { + my $header = $innodb_section_headers{$section}; + if ( !$header && $debug ) { + warn "Unknown section $section in $fulltext\n"; + } + + # The last section in the file is a special case, because instead of + # ending with the beginning of another section, it ends with the end of + # the file. So this section is complete if the entire file is + # complete. In different versions of InnoDB, various sections are + # last. + if ( $innodb_sections{$section}->[0] =~ s/\n---+\nEND OF INNODB.+\n=+$// ) { + $innodb_sections{$section}->[1] ||= $innodb_data{'got_all'}; + } + + if ( $header && $section ) { + $innodb_data{'sections'}->{ $header } + ->{'fulltext'} = $innodb_sections{$section}->[0]; + $innodb_data{'sections'}->{ $header } + ->{'complete'} = $innodb_sections{$section}->[1]; + } + else { + _debug( $debug, "header = " . ($header || 'undef') . ", section = " . ($section || 'undef')) if $debug; + } + } + }; + if ( $EVAL_ERROR ) { + _debug( $debug, $EVAL_ERROR); + } + + # ################################################################ + # Parse the detailed data out of the sections. + # ################################################################ + eval { + foreach my $section ( keys %parser_for ) { + if ( defined $innodb_data{'sections'}->{$section} + && (!%$sections || (defined($sections->{$section} && $sections->{$section})) )) { + $parser_for{$section}->( + $innodb_data{'sections'}->{$section}, + $innodb_data{'sections'}->{$section}->{'complete'}, + $debug, + $full, + $mysqlversion) + or delete $innodb_data{'sections'}->{$section}; + } + else { + delete $innodb_data{'sections'}->{$section}; + } + } + }; + if ( $EVAL_ERROR ) { + _debug( $debug, $EVAL_ERROR); + } + + return \%innodb_data; +} + +# Parses the status text and returns it flattened out as a single hash. +sub get_status_hash { + my ( $self, $fulltext, $debug, $sections, $full, $mysqlversion ) = @_; + + # Parse the status text... + my $innodb_status + = $self->parse_status_text($fulltext, $debug, $sections, $full, $mysqlversion ); + + # Flatten the hierarchical structure into a single list by grabbing desired + # sections from it. + return + (map { 'IB_' . $_ => $innodb_status->{$_} } qw(timestring last_secs got_all)), + (map { 'IB_bp_' . $_ => $innodb_status->{'sections'}->{'bp'}->{$_} } + qw( writes_pending buf_pool_hit_rate total_mem_alloc buf_pool_reads + awe_mem_alloc pages_modified writes_pending_lru page_creates_sec + reads_pending pages_total buf_pool_hits writes_pending_single_page + page_writes_sec pages_read pages_written page_reads_sec + writes_pending_flush_list buf_pool_size add_pool_alloc + dict_mem_alloc pages_created buf_free complete )), + (map { 'IB_tx_' . $_ => $innodb_status->{'sections'}->{'tx'}->{$_} } + qw( num_lock_structs history_list_len purge_done_for transactions + purge_undo_for is_truncated trx_id_counter complete )), + (map { 'IB_ib_' . $_ => $innodb_status->{'sections'}->{'ib'}->{$_} } + qw( hash_table_size hash_searches_s non_hash_searches_s + bufs_in_node_heap used_cells size free_list_len seg_size inserts + merged_recs merges complete )), + (map { 'IB_lg_' . $_ => $innodb_status->{'sections'}->{'lg'}->{$_} } + qw( log_ios_done pending_chkp_writes last_chkp log_ios_s + log_flushed_to log_seq_no pending_log_writes complete )), + (map { 'IB_sm_' . $_ => $innodb_status->{'sections'}->{'sm'}->{$_} } + qw( wait_array_size rw_shared_spins rw_excl_os_waits mutex_os_waits + mutex_spin_rounds mutex_spin_waits rw_excl_spins rw_shared_os_waits + waits signal_count reservation_count complete )), + (map { 'IB_ro_' . $_ => $innodb_status->{'sections'}->{'ro'}->{$_} } + qw( queries_in_queue n_reserved_extents main_thread_state + main_thread_proc_no main_thread_id read_sec del_sec upd_sec ins_sec + read_views_open num_rows_upd num_rows_ins num_rows_read + queries_inside num_rows_del complete )), + (map { 'IB_fk_' . $_ => $innodb_status->{'sections'}->{'fk'}->{$_} } + qw( trigger parent_table child_index parent_index attempted_op + child_db timestring fk_name records col_name reason txn parent_db + type child_table parent_col complete )), + (map { 'IB_io_' . $_ => $innodb_status->{'sections'}->{'io'}->{$_} } + qw( pending_buffer_pool_flushes pending_pwrites pending_preads + pending_normal_aio_reads fsyncs_s os_file_writes pending_sync_ios + reads_s flush_type avg_bytes_s pending_ibuf_aio_reads writes_s + threads os_file_reads pending_aio_writes pending_log_ios os_fsyncs + pending_log_flushes complete )), + (map { 'IB_dl_' . $_ => $innodb_status->{'sections'}->{'dl'}->{$_} } + qw( timestring rolled_back txns complete )); + +} + +sub ts_to_string { + my $parts = shift; + return sprintf('%02d-%02d-%02d %02d:%02d:%02d', @$parts); +} + +sub parse_innodb_timestamp { + my $text = shift; + my ( $y, $m, $d, $h, $i, $s ) + = $text =~ m/^(\d\d)(\d\d)(\d\d) +(\d+):(\d+):(\d+)$/; + die("Can't get timestamp from $text\n") unless $y; + $y += 2000; + return ( $y, $m, $d, $h, $i, $s ); +} + +sub parse_innodb_timestamp_56 { + my $text = shift; + my ( $y, $m, $d, $h, $i, $s ) + = $text =~ m/^(\d\d\d\d)-(\d\d)-(\d\d) +(\d+):(\d+):(\d+)$/; + die("Can't get timestamp from $text\n") unless $y; + return ( $y, $m, $d, $h, $i, $s ); +} + +sub parse_fk_section { + my ( $section, $complete, $debug, $full, $mysqlversion ) = @_; + my $fulltext = $section->{'fulltext'}; + + return 0 unless $fulltext; + + my ( $ts, $type ); + if ( ($mysqlversion =~ /^5.[67]\./) || ($mysqlversion =~ /^10.[0-9]\./) ) { + ( $ts, $type ) = $fulltext =~ m/^([0-9-]* [0-9:]*)\s[0-9a-fx]*\s+(\w+)/m; + $section->{'ts'} = [ parse_innodb_timestamp_56( $ts ) ]; + } else { + ( $ts, $type ) = $fulltext =~ m/^$s\s+(\w+)/m; + $section->{'ts'} = [ parse_innodb_timestamp( $ts ) ]; + } + + $section->{'timestring'} = ts_to_string($section->{'ts'}); + $section->{'type'} = $type; + + # Decide which type of FK error happened, and dispatch to the right parser. + if ( $type && $fk_parser_for{$type} ) { + $fk_parser_for{$type}->( $section, $complete, $debug, $fulltext, $full ); + } + + delete $section->{'fulltext'} unless $debug; + + return 1; +} + +sub parse_fk_cant_drop_parent_error { + my ( $section, $complete, $debug, $fulltext, $full ) = @_; + + # Parse the parent/child table info out + @{$section}{ qw(attempted_op parent_db parent_table) } = $fulltext + =~ m{Cannot $w table `(.*)/(.*)`}m; + @{$section}{ qw(child_db child_table) } = $fulltext + =~ m{because it is referenced by `(.*)/(.*)`}m; + + ( $section->{'reason'} ) = $fulltext =~ m/(Cannot .*)/s; + if ( !defined $section->{reason} ) { + ( $section->{'reason'} ) = $fulltext =~ m/(Trying to add .*)/s; + } + $section->{'reason'} =~ s/\n(?:InnoDB: )?/ /gm + if $section->{'reason'}; + + # Certain data may not be present. Make them '' if not present. + map { $section->{$_} ||= "" } + qw(child_index fk_name col_name parent_col); +} + +# See dict/dict0dict.c, function dict_foreign_error_report +# I don't care much about these. There are lots of different messages, and +# they come from someone trying to create a foreign key, or similar +# statements. They aren't indicative of some transaction trying to insert, +# delete or update data. Sometimes it is possible to parse out a lot of +# information about the tables and indexes involved, but often the message +# contains the DDL string the user entered, which is way too much for this +# module to try to handle. +sub parse_fk_bad_constraint_error { + my ( $section, $complete, $debug, $fulltext, $full ) = @_; + + # Parse the parent/child table and index info out + @{$section}{ qw(child_db child_table) } = $fulltext + =~ m{Error in foreign key constraint of table (.*)/(.*):$}m; + $section->{'attempted_op'} = 'DDL'; + + # FK name, parent info... if possible. + @{$section}{ qw(fk_name col_name parent_db parent_table parent_col) } + = $fulltext + =~ m/CONSTRAINT `?$n`? FOREIGN KEY \(`?$n`?\) REFERENCES (?:`?$n`?\.)?`?$n`? \(`?$n`?\)/; + + if ( !defined($section->{'fk_name'}) ) { + # Try to parse SQL a user might have typed in a CREATE statement or such + @{$section}{ qw(col_name parent_db parent_table parent_col) } + = $fulltext + =~ m/FOREIGN\s+KEY\s*\(`?$n`?\)\s+REFERENCES\s+(?:`?$n`?\.)?`?$n`?\s*\(`?$n`?\)/i; + } + $section->{'parent_db'} ||= $section->{'child_db'}; + + # Name of the child index (index in the same table where the FK is, see + # definition of dict_foreign_struct in include/dict0mem.h, where it is + # called foreign_index, as opposed to referenced_index which is in the + # parent table. This may not be possible to find. + @{$section}{ qw(child_index) } = $fulltext + =~ m/^The index in the foreign key in table is $n$/m; + + @{$section}{ qw(reason) } = $fulltext =~ m/:\s*([^:]+)(?= Constraint:|$)/ms; + $section->{'reason'} =~ s/\s+/ /g + if $section->{'reason'}; + + # Certain data may not be present. Make them '' if not present. + map { $section->{$_} ||= "" } + qw(child_index fk_name col_name parent_table parent_col); +} + +# see source file row/row0ins.c +sub parse_fk_transaction_error { + my ( $section, $complete, $debug, $fulltext, $full ) = @_; + + # Parse the txn info out + my ( $txn ) = $fulltext + =~ m/Transaction:\n(TRANSACTION.*)\nForeign key constraint fails/s; + if ( $txn ) { + $section->{'txn'} = parse_tx_text( $txn, $complete, $debug, $full ); + } + + # Parse the parent/child table and index info out. There are two types: an + # update or a delete of a parent record leaves a child orphaned + # (row_ins_foreign_report_err), and an insert or update of a child record has + # no matching parent record (row_ins_foreign_report_add_err). + + @{$section}{ qw(reason child_db child_table) } + = $fulltext =~ m{^(Foreign key constraint fails for table `(.*?)`?[/.]`?(.*)`:)$}m; + + @{$section}{ qw(fk_name col_name parent_db parent_table parent_col) } + = $fulltext + =~ m/CONSTRAINT `$n` FOREIGN KEY \(`$n`\) REFERENCES (?:`$n`\.)?`$n` \(`$n`\)/; + $section->{'parent_db'} ||= $section->{'child_db'}; + + # Special case, which I don't know how to trigger, but see + # innobase/row/row0ins.c row_ins_check_foreign_constraint + if ( $fulltext =~ m/ibd file does not currently exist!/ ) { + my ( $attempted_op, $index, $records ) + = $fulltext =~ m/^Trying to (add to index) `$n` tuple:\n(.*))?/sm; + $section->{'child_index'} = $index; + $section->{'attempted_op'} = $attempted_op || ''; + if ( $records && $full ) { + ( $section->{'records'} ) + = parse_innodb_record_dump( $records, $complete, $debug ); + } + @{$section}{qw(parent_db parent_table)} + =~ m/^But the parent table `$n`\.`$n`$/m; + } + else { + my ( $attempted_op, $which, $index ) + = $fulltext =~ m/^Trying to ([\w ]*) in (child|parent) table, in index `$n` tuple:$/m; + if ( $which ) { + $section->{$which . '_index'} = $index; + $section->{'attempted_op'} = $attempted_op || ''; + + # Parse out the related records in the other table. + my ( $search_index, $records ); + if ( $which eq 'child' ) { + ( $search_index, $records ) = $fulltext + =~ m/^But in parent table [^,]*, in index `$n`,\nthe closest match we can find is record:\n(.*)/ms; + $section->{'parent_index'} = $search_index; + } + else { + ( $search_index, $records ) = $fulltext + =~ m/^But in child table [^,]*, in index `$n`, (?:the record is not available|there is a record:\n(.*))?/ms; + $section->{'child_index'} = $search_index; + } + if ( $records && $full ) { + $section->{'records'} + = parse_innodb_record_dump( $records, $complete, $debug ); + } + else { + $section->{'records'} = ''; + } + } + } + + # Parse out the tuple trying to be updated, deleted or inserted. + my ( $trigger ) = $fulltext =~ m/^(DATA TUPLE: \d+ fields;\n.*)$/m; + if ( $trigger ) { + $section->{'trigger'} = parse_innodb_record_dump( $trigger, $complete, $debug ); + } + + # Certain data may not be present. Make them '' if not present. + map { $section->{$_} ||= "" } + qw(child_index fk_name col_name parent_table parent_col); +} + +# There are new-style and old-style record formats. See rem/rem0rec.c +# TODO: write some tests for this +sub parse_innodb_record_dump { + my ( $dump, $complete, $debug ) = @_; + return undef unless $dump; + + my $result = {}; + + if ( $dump =~ m/PHYSICAL RECORD/ ) { + my $style = $dump =~ m/compact format/ ? 'new' : 'old'; + $result->{'style'} = $style; + + # This is a new-style record. + if ( $style eq 'new' ) { + @{$result}{qw( heap_no type num_fields info_bits )} + = $dump + =~ m/^(?:Record lock, heap no $d )?([A-Z ]+): n_fields $d; compact format; info bits $d$/m; + } + + # OK, it's old-style. Unfortunately there are variations here too. + elsif ( $dump =~ m/-byte offs / ) { + # Older-old style. + @{$result}{qw( heap_no type num_fields byte_offset info_bits )} + = $dump + =~ m/^(?:Record lock, heap no $d )?([A-Z ]+): n_fields $d; $d-byte offs [A-Z]+; info bits $d$/m; + if ( $dump !~ m/-byte offs TRUE/ ) { + $result->{'byte_offset'} = 0; + } + } + else { + # Newer-old style. + @{$result}{qw( heap_no type num_fields byte_offset info_bits )} + = $dump + =~ m/^(?:Record lock, heap no $d )?([A-Z ]+): n_fields $d; $d-byte offsets; info bits $d$/m; + } + + } + else { + $result->{'style'} = 'tuple'; + @{$result}{qw( type num_fields )} + = $dump =~ m/^(DATA TUPLE): $d fields;$/m; + } + + # Fill in default values for things that couldn't be parsed. + map { $result->{$_} ||= 0 } + qw(heap_no num_fields byte_offset info_bits); + map { $result->{$_} ||= '' } + qw(style type ); + + my @fields = $dump =~ m/ (\d+:.*?;?);(?=$| \d+:)/gm; + $result->{'fields'} = [ map { parse_field($_, $complete, $debug ) } @fields ]; + + return $result; +} + +# New/old-style applies here. See rem/rem0rec.c +# $text should not include the leading space or the second trailing semicolon. +sub parse_field { + my ( $text, $complete, $debug ) = @_; + + # Sample fields: + # '4: SQL NULL, size 4 ' + # '1: len 6; hex 000000005601; asc V ;' + # '6: SQL NULL' + # '5: len 30; hex 687474703a2f2f7777772e737765657477617465722e636f6d2f73746f72; asc http://www.sweetwater.com/stor;...(truncated)' + my ( $id, $nullsize, $len, $hex, $asc, $truncated ); + ( $id, $nullsize ) = $text =~ m/^$d: SQL NULL, size $d $/; + if ( !defined($id) ) { + ( $id ) = $text =~ m/^$d: SQL NULL$/; + } + if ( !defined($id) ) { + ( $id, $len, $hex, $asc, $truncated ) + = $text =~ m/^$d: len $d; hex $h; asc (.*);(\.\.\.\(truncated\))?$/; + } + + die "Could not parse this field: '$text'" unless defined $id; + return { + id => $id, + len => defined($len) ? $len : defined($nullsize) ? $nullsize : 0, + 'hex' => defined($hex) ? $hex : '', + asc => defined($asc) ? $asc : '', + trunc => $truncated ? 1 : 0, + }; + +} + +sub parse_dl_section { + my ( $dl, $complete, $debug, $full, $mysqlversion ) = @_; + return unless $dl; + my $fulltext = $dl->{'fulltext'}; + return 0 unless $fulltext; + + my ( $ts ) = $fulltext =~ m/^$s$/m; + return 0 unless $ts; + + if ( ($mysqlversion =~ /^5\.[67]\./) || ($mysqlversion =~ /^10\.[0-9]\./) ) { + $dl->{'ts'} = [ parse_innodb_timestamp_56( $ts ) ]; + } + else { + $dl->{'ts'} = [ parse_innodb_timestamp( $ts ) ]; + } + $dl->{'timestring'} = ts_to_string($dl->{'ts'}); + $dl->{'txns'} = {}; + + my @sections + = $fulltext + =~ m{ + ^\*{3}\s([^\n]*) # *** (1) WAITING FOR THIS... + (.*?) # Followed by anything, non-greedy + (?=(?:^\*{3})|\z) # Followed by another three stars or EOF + }gmsx; + + + # Loop through each section. There are no assumptions about how many + # there are, who holds and wants what locks, and who gets rolled back. + while ( my ($header, $body) = splice(@sections, 0, 2) ) { + my ( $txn_id, $what ) = $header =~ m/^\($d\) (.*):$/; + next unless $txn_id; + $dl->{'txns'}->{$txn_id} ||= {}; + my $txn = $dl->{'txns'}->{$txn_id}; + + if ( $what eq 'TRANSACTION' ) { + $txn->{'tx'} = parse_tx_text( $body, $complete, $debug, $full ); + } + else { + push @{$txn->{'locks'}}, parse_innodb_record_locks( $body, $complete, $debug, $full ); + } + } + + @{ $dl }{ qw(rolled_back) } + = $fulltext =~ m/^\*\*\* WE ROLL BACK TRANSACTION \($d\)$/m; + + # Make sure certain values aren't undef + map { $dl->{$_} ||= '' } qw(rolled_back); + + delete $dl->{'fulltext'} unless $debug; + return 1; +} + +sub parse_innodb_record_locks { + my ( $text, $complete, $debug, $full ) = @_; + my @result; + + foreach my $lock ( $text =~ m/(^(?:RECORD|TABLE) LOCKS?.*$)/gm ) { + my $hash = {}; + @{$hash}{ qw(lock_type space_id page_no n_bits index db table txn_id lock_mode) } + = $lock + =~ m{^(RECORD|TABLE) LOCKS? (?:space id $d page no $d n bits $d index `?$n`? of )?table `$n(?:/|`\.`)$n` trx id $t lock.mode (\S+)}m; + ( $hash->{'special'} ) + = $lock =~ m/^(?:RECORD|TABLE) .*? locks (rec but not gap|gap before rec)/m; + $hash->{'insert_intention'} + = $lock =~ m/^(?:RECORD|TABLE) .*? insert intention/m ? 1 : 0; + $hash->{'waiting'} + = $lock =~ m/^(?:RECORD|TABLE) .*? waiting/m ? 1 : 0; + + # Some things may not be in the text, so make sure they are not + # undef. + map { $hash->{$_} ||= 0 } qw(n_bits page_no space_id); + map { $hash->{$_} ||= "" } qw(index special); + push @result, $hash; + } + + return @result; +} + +sub parse_tx_text { + my ( $txn, $complete, $debug, $full ) = @_; + + my ( $txn_id, $txn_status ) + = $txn + =~ m/^(?:---)?TRANSACTION $t, ([^\n0-9,]*[^\s\d])/m; + $txn_status =~ s/,$// if $txn_status; + my ( $active_secs) + = $txn + =~ m/^[^\n]*\b$d sec\b/m; + my ( $proc_no ) + = $txn + =~ m/process no $d/m; + my ( $os_thread_id ) + = $txn + =~ m/OS thread id $d/m; + my ( $thread_status, $thread_decl_inside ) + = $txn + =~ m/(?:OS thread id \d+|\d sec)(?: ([^,]+?))?(?:, thread declared inside InnoDB $d)?$/m; + + # Parsing the line that begins 'MySQL thread id' is complicated. The only + # thing always in the line is the thread and query id. See function + # innobase_mysql_print_thd in InnoDB source file sql/ha_innodb.cc. + my ( $thread_line ) = $txn =~ m/^((?:MariaDB|MySQL) thread id .*)$/m; + my ( $mysql_thread_id, $query_id, $hostname, $ip, $user, $query_status ); + + if ( $thread_line ) { + # These parts can always be gotten. + ( $mysql_thread_id, $query_id ) = $thread_line =~ m/^(?:MariaDB|MySQL) thread id $d, .*?query id $d/m; + + # If it's a master/slave thread, "Has (read|sent) all" may be the thread's + # proc_info. In these cases, there won't be any host/ip/user info + ( $query_status ) = $thread_line =~ m/(Has (?:read|sent) all .*$)/m; + if ( defined($query_status) ) { + $user = 'system user'; + } + + # It may be the case that the query id is the last thing in the line. + elsif ( $thread_line =~ m/query id \d+ / ) { + # The IP address is the only non-word thing left, so it's the most + # useful marker for where I have to start guessing. + ( $hostname, $ip ) = $thread_line =~ m/query id \d+(?: ([A-Za-z]\S+))? $i/m; + if ( defined $ip ) { + ( $user, $query_status ) = $thread_line =~ m/$ip $w(?: (.*))?$/; + } + else { # OK, there wasn't an IP address. + # There might not be ANYTHING except the query status. + ( $query_status ) = $thread_line =~ m/query id \d+ (.*)$/; + if ( $query_status !~ m/^\w+ing/ && !exists($is_proc_info{$query_status}) ) { + # The remaining tokens are, in order: hostname, user, query_status. + # It's basically impossible to know which is which. + ( $hostname, $user, $query_status ) = $thread_line + =~ m/query id \d+(?: ([A-Za-z]\S+))?(?: $w(?: (.*))?)?$/m; + if ( ($hostname || '') eq 'Slave' ) { + $hostname = ''; + $user = 'system user'; + $query_status = "Slave has $query_status"; + } + } + else { + $user = 'system user'; + } + } + } + } + + my ( $lock_wait_status, $lock_structs, $heap_size, $row_locks, $undo_log_entries ) + = $txn + =~ m/^(?:(\D*) )?$d lock struct\(s\), heap size $d(?:, $d row lock\(s\))?(?:, undo log entries $d)?$/m; + my ( $lock_wait_time ) + = $txn + =~ m/^------- TRX HAS BEEN WAITING $d SEC/m; + + my $locks; + # If the transaction has locks, grab the locks. + if ( $txn =~ m/^TABLE LOCK|RECORD LOCKS/ ) { + $locks = [parse_innodb_record_locks($txn, $complete, $debug, $full)]; + } + + my ( $tables_in_use, $tables_locked ) + = $txn + =~ m/^mysql tables in use $d, locked $d$/m; + my ( $txn_doesnt_see_ge, $txn_sees_lt ) + = $txn + =~ m/^Trx read view will not see trx with id >= $t, sees < $t$/m; + my $has_read_view = defined($txn_doesnt_see_ge); + # Only a certain number of bytes of the query text are included here, at least + # under some circumstances. Some versions include 300, some 600. + my ( $query_text ) + = $txn + =~ m{ + ^MySQL\sthread\sid\s[^\n]+\n # This comes before the query text + (.*?) # The query text + (?= # Followed by any of... + ^Trx\sread\sview + |^-------\sTRX\sHAS\sBEEN\sWAITING + |^TABLE\sLOCK + |^RECORD\sLOCKS\sspace\sid + |^(?:---)?TRANSACTION + |^\*\*\*\s\(\d\) + |\Z + ) + }xms; + if ( $query_text ) { + $query_text =~ s/\s+$//; + } + else { + $query_text = ''; + } + + my %stuff = ( + active_secs => $active_secs, + has_read_view => $has_read_view, + heap_size => $heap_size, + hostname => $hostname, + ip => $ip, + lock_structs => $lock_structs, + lock_wait_status => $lock_wait_status, + lock_wait_time => $lock_wait_time, + mysql_thread_id => $mysql_thread_id, + os_thread_id => $os_thread_id, + proc_no => $proc_no, + query_id => $query_id, + query_status => $query_status, + query_text => $query_text, + row_locks => $row_locks, + tables_in_use => $tables_in_use, + tables_locked => $tables_locked, + thread_decl_inside => $thread_decl_inside, + thread_status => $thread_status, + txn_doesnt_see_ge => $txn_doesnt_see_ge, + txn_id => $txn_id, + txn_sees_lt => $txn_sees_lt, + txn_status => $txn_status, + undo_log_entries => $undo_log_entries, + user => $user, + ); + $stuff{'fulltext'} = $txn if $debug; + $stuff{'locks'} = $locks if $locks; + + # Some things may not be in the txn text, so make sure they are not + # undef. + map { $stuff{$_} ||= 0 } qw(active_secs heap_size lock_structs + tables_in_use undo_log_entries tables_locked has_read_view + thread_decl_inside lock_wait_time proc_no row_locks); + map { $stuff{$_} ||= "" } qw(thread_status txn_doesnt_see_ge + txn_sees_lt query_status ip query_text lock_wait_status user); + $stuff{'hostname'} ||= $stuff{'ip'}; + + return \%stuff; +} + +sub parse_tx_section { + my ( $section, $complete, $debug, $full ) = @_; + return unless $section && $section->{'fulltext'}; + my $fulltext = $section->{'fulltext'}; + $section->{'transactions'} = []; + + # Handle the individual transactions + my @transactions = $fulltext =~ m/(---TRANSACTION [0-9A-Fa-f].*?)(?=\n---TRANSACTION|$)/gs; + foreach my $txn ( @transactions ) { + my $stuff = parse_tx_text( $txn, $complete, $debug, $full ); + delete $stuff->{'fulltext'} unless $debug; + push @{$section->{'transactions'}}, $stuff; + } + + # Handle the general info + @{$section}{ 'trx_id_counter' } + = $fulltext =~ m/^Trx id counter $t$/m; + @{$section}{ 'purge_done_for', 'purge_undo_for' } + = $fulltext =~ m/^Purge done for trx's n:o < $t undo n:o < $t$/m; + @{$section}{ 'history_list_len' } # This isn't present in some 4.x versions + = $fulltext =~ m/^History list length $d$/m; + @{$section}{ 'num_lock_structs' } + = $fulltext =~ m/^Total number of lock structs in row lock hash table $d$/m; + @{$section}{ 'is_truncated' } + = $fulltext =~ m/^\.\.\. truncated\.\.\.$/m ? 1 : 0; + + # Fill in things that might not be present + foreach ( qw(history_list_len) ) { + $section->{$_} ||= 0; + } + + delete $section->{'fulltext'} unless $debug; + return 1; +} + +# I've read the source for this section. +sub parse_ro_section { + my ( $section, $complete, $debug, $full ) = @_; + return unless $section && $section->{'fulltext'}; + my $fulltext = $section->{'fulltext'}; + + # Grab the info + @{$section}{ 'queries_inside', 'queries_in_queue' } + = $fulltext =~ m/^$d queries inside InnoDB, $d queries in queue$/m; + ( $section->{ 'read_views_open' } ) + = $fulltext =~ m/^$d read views open inside InnoDB$/m; + ( $section->{ 'n_reserved_extents' } ) + = $fulltext =~ m/^$d tablespace extents now reserved for B-tree/m; + @{$section}{ 'main_thread_proc_no', 'main_thread_id', 'main_thread_state' } + = $fulltext =~ m/^Main thread (?:process no. $d, )?id $d, state: (.*)$/m; + @{$section}{ 'num_rows_ins', 'num_rows_upd', 'num_rows_del', 'num_rows_read' } + = $fulltext =~ m/^Number of rows inserted $d, updated $d, deleted $d, read $d$/m; + @{$section}{ 'ins_sec', 'upd_sec', 'del_sec', 'read_sec' } + = $fulltext =~ m#^$f inserts/s, $f updates/s, $f deletes/s, $f reads/s$#m; + $section->{'main_thread_proc_no'} ||= 0; + + map { $section->{$_} ||= 0 } qw(read_views_open n_reserved_extents); + delete $section->{'fulltext'} unless $debug; + return 1; +} + +sub parse_lg_section { + my ( $section, $complete, $debug, $full ) = @_; + return unless $section; + my $fulltext = $section->{'fulltext'}; + + # Grab the info + ( $section->{ 'log_seq_no' } ) + = $fulltext =~ m/Log sequence number \s*(\d.*)$/m; + ( $section->{ 'log_flushed_to' } ) + = $fulltext =~ m/Log flushed up to \s*(\d.*)$/m; + ( $section->{ 'last_chkp' } ) + = $fulltext =~ m/Last checkpoint at \s*(\d.*)$/m; + @{$section}{ 'pending_log_writes', 'pending_chkp_writes' } + = $fulltext =~ m/$d pending log (?:writes|flushes), $d pending chkp writes/; + @{$section}{ 'log_ios_done', 'log_ios_s' } + = $fulltext =~ m#$d log i/o's done, $f log i/o's/second#; + + delete $section->{'fulltext'} unless $debug; + return 1; +} + +sub parse_ib_section { + my ( $section, $complete, $debug, $full ) = @_; + return unless $section && $section->{'fulltext'}; + my $fulltext = $section->{'fulltext'}; + + # Some servers will output ibuf information for tablespace 0, as though there + # might be many tablespaces with insert buffers. (In practice I believe + # the source code shows there will only ever be one). I have to parse both + # cases here, but I assume there will only be one. + @{$section}{ 'size', 'free_list_len', 'seg_size' } + = $fulltext =~ m/^Ibuf(?: for space 0)?: size $d, free list len $d, seg size $d/m; + @{$section}{ 'inserts', 'merged_recs', 'merges' } + = $fulltext =~ m/^$d inserts, $d merged recs, $d merges$/m; + if ( ! defined $section->{inserts} ) { + @{$section}{ 'inserts' } + = $fulltext =~ m/merged operations:\n insert $d,/s; + # This isn't really true, but it's not really important either. We already + # aren't supporting the 'delete' operations. + @{$section}{ 'merged_recs', 'merges' } = (0, 0); + } + + @{$section}{ 'hash_table_size', 'used_cells', 'bufs_in_node_heap' } + = $fulltext =~ m/^Hash table size $d(?:, used cells $d)?, node heap has $d buffer\(s\)$/m; + @{$section}{ 'hash_searches_s', 'non_hash_searches_s' } + = $fulltext =~ m{^$f hash searches/s, $f non-hash searches/s$}m; + + delete $section->{'fulltext'} unless $debug; + return 1; +} + +sub parse_wait_array { + my ( $text, $complete, $debug, $full ) = @_; + my %result; + + @result{ qw(thread waited_at_filename waited_at_line waited_secs) } + = $text =~ m/^--Thread $d has waited at $fl for $f seconds/m; + + # Depending on whether it's a SYNC_MUTEX,RW_LOCK_EX,RW_LOCK_SHARED, + # there will be different text output + if ( $text =~ m/^Mutex at/m ) { + $result{'request_type'} = 'M'; + @result{ qw( lock_mem_addr lock_cfile_name lock_cline lock_var) } + = $text =~ m/^Mutex at $h created file $fl, lock var $d$/m; + @result{ qw( waiters_flag )} + = $text =~ m/^waiters flag $d$/m; + } + else { + @result{ qw( request_type lock_mem_addr lock_cfile_name lock_cline) } + = $text =~ m/^(.)-lock on RW-latch at $h created in file $fl$/m; + @result{ qw( writer_thread writer_lock_mode ) } + = $text =~ m/^a writer \(thread id $d\) has reserved it in mode (.*)$/m; + @result{ qw( num_readers waiters_flag )} + = $text =~ m/^number of readers $d, waiters flag $d$/m; + @result{ qw(last_s_file_name last_s_line ) } + = $text =~ m/Last time read locked in file $fl$/m; + @result{ qw(last_x_file_name last_x_line ) } + = $text =~ m/Last time write locked in file $fl$/m; + } + + $result{'cell_waiting'} = $text =~ m/^wait has ended$/m ? 0 : 1; + $result{'cell_event_set'} = $text =~ m/^wait is ending$/m ? 1 : 0; + + # Because there are two code paths, some things won't get set. + map { $result{$_} ||= '' } + qw(last_s_file_name last_x_file_name writer_lock_mode); + map { $result{$_} ||= 0 } + qw(num_readers lock_var last_s_line last_x_line writer_thread); + + return \%result; +} + +sub parse_sm_section { + my ( $section, $complete, $debug, $full ) = @_; + return 0 unless $section && $section->{'fulltext'}; + my $fulltext = $section->{'fulltext'}; + + # Grab the info + @{$section}{ 'reservation_count', 'signal_count' } + = $fulltext =~ m/^OS WAIT ARRAY INFO: reservation count $d, signal count $d$/m; + @{$section}{ 'mutex_spin_waits', 'mutex_spin_rounds', 'mutex_os_waits' } + = $fulltext =~ m/^Mutex spin waits $d, rounds $d, OS waits $d$/m; + @{$section}{ 'rw_shared_spins', 'rw_shared_os_waits', 'rw_excl_spins', 'rw_excl_os_waits' } + = $fulltext =~ m/^RW-shared spins $d, OS waits $d; RW-excl spins $d, OS waits $d$/m; + if ( ! defined $section->{rw_shared_spins} ) { + @{$section}{ 'rw_shared_spins', 'rw_shared_os_waits'} + = $fulltext =~ m/^RW-shared spins $d, rounds \d+, OS waits $d$/m; + @{$section}{ 'rw_excl_spins', 'rw_excl_os_waits' } + = $fulltext =~ m/^RW-excl spins $d, rounds \d+, OS waits $d$/m; + } + + # Look for info on waits. + my @waits = $fulltext =~ m/^(--Thread.*?)^(?=Mutex spin|--Thread)/gms; + $section->{'waits'} = [ map { parse_wait_array($_, $complete, $debug) } @waits ]; + $section->{'wait_array_size'} = scalar(@waits); + + delete $section->{'fulltext'} unless $debug; + return 1; +} + +# I've read the source for this section. +sub parse_bp_section { + my ( $section, $complete, $debug, $full ) = @_; + return unless $section && $section->{'fulltext'}; + my $fulltext = $section->{'fulltext'}; + + # Grab the info + @{$section}{ 'total_mem_alloc', 'add_pool_alloc' } + = $fulltext =~ m/^Total memory allocated $d; in additional pool allocated $d$/m; + @{$section}{'dict_mem_alloc'} = $fulltext =~ m/Dictionary memory allocated $d/; + @{$section}{'awe_mem_alloc'} = $fulltext =~ m/$d MB of AWE memory/; + @{$section}{'buf_pool_size'} = $fulltext =~ m/^Buffer pool size\s*$d$/m; + @{$section}{'buf_free'} = $fulltext =~ m/^Free buffers\s*$d$/m; + @{$section}{'pages_total'} = $fulltext =~ m/^Database pages\s*$d$/m; + @{$section}{'pages_modified'} = $fulltext =~ m/^Modified db pages\s*$d$/m; + @{$section}{'pages_read', 'pages_created', 'pages_written'} + = $fulltext =~ m/^Pages read $d, created $d, written $d$/m; + @{$section}{'page_reads_sec', 'page_creates_sec', 'page_writes_sec'} + = $fulltext =~ m{^$f reads/s, $f creates/s, $f writes/s$}m; + @{$section}{'buf_pool_hits', 'buf_pool_reads'} + = $fulltext =~ m{Buffer pool hit rate $d / $d}m; + if ($fulltext =~ m/^No buffer pool page gets since the last printout$/m) { + @{$section}{'buf_pool_hits', 'buf_pool_reads'} = (0, 0); + @{$section}{'buf_pool_hit_rate'} = '--'; + } + else { + @{$section}{'buf_pool_hit_rate'} + = $fulltext =~ m{Buffer pool hit rate (\d+ / \d+)}m; + } + @{$section}{'reads_pending'} = $fulltext =~ m/^Pending reads $d/m; + @{$section}{'writes_pending_lru', 'writes_pending_flush_list', 'writes_pending_single_page' } + = $fulltext =~ m/^Pending writes: LRU $d, flush list $d, single page $d$/m; + + map { $section->{$_} ||= 0 } + qw(writes_pending_lru writes_pending_flush_list writes_pending_single_page + awe_mem_alloc dict_mem_alloc); + @{$section}{'writes_pending'} = List::Util::sum( + @{$section}{ qw(writes_pending_lru writes_pending_flush_list writes_pending_single_page) }); + + delete $section->{'fulltext'} unless $debug; + return 1; +} + +# I've read the source for this. +sub parse_io_section { + my ( $section, $complete, $debug, $full ) = @_; + return unless $section && $section->{'fulltext'}; + my $fulltext = $section->{'fulltext'}; + $section->{'threads'} = {}; + + # Grab the I/O thread info + my @threads = $fulltext =~ m<^(I/O thread \d+ .*)$>gm; + foreach my $thread (@threads) { + my ( $tid, $state, $purpose, $event_set ) + = $thread =~ m{I/O thread $d state: (.+?) \((.*)\)(?: ev set)?$}m; + if ( defined $tid ) { + $section->{'threads'}->{$tid} = { + thread => $tid, + state => $state, + purpose => $purpose, + event_set => $event_set ? 1 : 0, + }; + } + } + + # Grab the reads/writes/flushes info + @{$section}{ 'pending_normal_aio_reads', 'pending_aio_writes' } + = $fulltext =~ m/^Pending normal aio reads: $d(?: [^\]]*\])?, aio writes: $d/m; + @{$section}{ 'pending_ibuf_aio_reads', 'pending_log_ios', 'pending_sync_ios' } + = $fulltext =~ m{^ ibuf aio reads: $d, log i/o's: $d, sync i/o's: $d$}m; + @{$section}{ 'flush_type', 'pending_log_flushes', 'pending_buffer_pool_flushes' } + = $fulltext =~ m/^Pending flushes \($w\) log: $d; buffer pool: $d$/m; + @{$section}{ 'os_file_reads', 'os_file_writes', 'os_fsyncs' } + = $fulltext =~ m/^$d OS file reads, $d OS file writes, $d OS fsyncs$/m; + @{$section}{ 'reads_s', 'avg_bytes_s', 'writes_s', 'fsyncs_s' } + = $fulltext =~ m{^$f reads/s, $d avg bytes/read, $f writes/s, $f fsyncs/s$}m; + @{$section}{ 'pending_preads', 'pending_pwrites' } + = $fulltext =~ m/$d pending preads, $d pending pwrites$/m; + @{$section}{ 'pending_preads', 'pending_pwrites' } = (0, 0) + unless defined($section->{'pending_preads'}); + + delete $section->{'fulltext'} unless $debug; + return 1; +} + +sub _debug { + my ( $debug, $msg ) = @_; + if ( $debug ) { + die $msg; + } + else { + warn $msg; + } + return 1; +} + +1; + +# end_of_package InnoDBParser + +package main; + +use sigtrap qw(handler finish untrapped normal-signals); + +use Data::Dumper; +use DBI; +use English qw(-no_match_vars); +use File::Basename qw(dirname); +use File::Temp; +use Getopt::Long; +use List::Util qw(max min maxstr sum); +use POSIX qw(ceil); +use Time::HiRes qw(time sleep); +use Term::ReadKey qw(ReadMode ReadKey); + +# License and warranty information. {{{1 +# ########################################################################### + +my $innotop_license = <<"LICENSE"; + +This is innotop version $VERSION, a MySQL and InnoDB monitor. + +This program is copyright (c) 2006 Baron Schwartz. +Feedback and improvements are welcome. + +THIS PROGRAM IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED +WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF +MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + +This program is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free Software +Foundation, version 2; OR the Perl Artistic License. On UNIX and similar +systems, you can issue `man perlgpl' or `man perlartistic' to read these +licenses. + +You should have received a copy of the GNU General Public License along with +this program; if not, write to the Free Software Foundation, Inc., 51 Franklin +Street, Fifth Floor, Boston, MA 02110-1335 USA. +LICENSE + +# Configuration information and global setup {{{1 +# ########################################################################### + +# Really, really, super-global variables. +my @config_versions = ( + "000-000-000", "001-003-000", # config file was one big name-value hash. + "001-003-000", "001-004-002", # config file contained non-user-defined stuff. +); + +my $clear_screen_sub; +my $dsn_parser = new DSNParser(); + +# This defines expected properties and defaults for the column definitions that +# eventually end up in tbl_meta. +my %col_props = ( + hdr => '', + just => '-', + dec => 0, # Whether to align the column on the decimal point + num => 0, + label => '', + user => 0, + src => '', + tbl => '', # Helps when writing/reading custom columns in config files + minw => 0, + maxw => 0, + trans => [], + agg => 'first', # Aggregate function + aggonly => 0, # Whether to show only when tbl_meta->{aggregate} is true + agghide => 0, # Whether NOT to show when tbl_meta->{aggregate} is true +); + +# Actual DBI connections to MySQL servers. +my %dbhs; + +# Command-line parameters {{{2 +# ########################################################################### + +my @opt_spec = ( + { s => 'help', d => 'Show this help message' }, + { s => 'color|C!', d => 'Use terminal coloring (default)', c => 'color' }, + { s => 'config|c=s', d => 'Config file to read' }, + { s => 'nonint|n', d => 'Non-interactive, output tab-separated fields' }, + { s => 'count=i', d => 'Number of updates before exiting' }, + { s => 'delay|d=f', d => 'Delay between updates in seconds', c => 'interval' }, + { s => 'mode|m=s', d => 'Operating mode to start in', c => 'mode' }, + { s => 'inc|i!', d => 'Measure incremental differences', c => 'status_inc' }, + { s => 'spark=i', d => 'Length of status sparkline (default 10)', c=> 'spark' }, + { s => 'write|w', d => 'Write running configuration into home directory if no config files were loaded' }, + { s => 'skipcentral|s', d => 'Skip reading the central configuration file' }, + { s => 'version', d => 'Output version information and exit' }, + { s => 'user|u=s', d => 'User for login if not current user' }, + { s => 'password|p=s', d => 'Password to use for connection' }, + { s => 'askpass', d => 'Prompt for a password when connecting to MySQL'}, + { s => 'host|h=s', d => 'Connect to host' }, + { s => 'port|P=i', d => 'Port number to use for connection' }, + { s => 'socket|S=s', d => 'MySQL socket to use for connection' }, + { s => 'timestamp|t+', d => 'Print timestamp in -n mode (1: per iter; 2: per line)' }, +); + +# This is the container for the command-line options' values to be stored in +# after processing. Initial values are defaults. +my %opts = ( + n => !( -t STDIN && -t STDOUT ), # If in/out aren't to terminals, we're interactive +); +# Post-process... +my %opt_seen; +foreach my $spec ( @opt_spec ) { + my ( $long, $short ) = $spec->{s} =~ m/^(\w+)(?:\|([^!+=:]*))?/; + $spec->{k} = $short || $long; + $spec->{l} = $long; + $spec->{t} = $short; + $spec->{n} = $spec->{s} =~ m/!/; + $opts{$spec->{k}} = undef unless defined $opts{$spec->{k}}; + die "Duplicate option $spec->{k}" if $opt_seen{$spec->{k}}++; +} + +Getopt::Long::Configure('no_ignore_case', 'bundling'); +GetOptions( map { $_->{s} => \$opts{$_->{k}} } @opt_spec) or $opts{help} = 1; + +if ( $opts{version} ) { + print "innotop Ver $VERSION\n"; + exit(0); +} + +if ( $opts{c} and ! -f $opts{c} ) { + print $opts{c} . " doesn't exist. Exiting.\n"; + exit(1); +} + +if ( $opts{'askpass'} ) { + $opts{'p'} = noecho_password("Enter password "); +} + +if ( $opts{'help'} ) { + print "Usage: innotop <options> <innodb-status-file>\n\n"; + my $maxw = max(map { length($_->{l}) + ($_->{n} ? 4 : 0)} @opt_spec); + foreach my $spec ( sort { $a->{l} cmp $b->{l} } @opt_spec ) { + my $long = $spec->{n} ? "[no]$spec->{l}" : $spec->{l}; + my $short = $spec->{t} ? "-$spec->{t}" : ''; + printf(" --%-${maxw}s %-4s %s\n", $long, $short, $spec->{d}); + } + print <<USAGE; + +innotop is a MySQL and InnoDB transaction/status monitor, like 'top' for +MySQL. It displays queries, InnoDB transactions, lock waits, deadlocks, +foreign key errors, open tables, replication status, buffer information, +row operations, logs, I/O operations, load graph, and more. You can +monitor many servers at once with innotop. + +USAGE + exit(1); +} + +if ( defined($opts{p}) && length($opts{p}) == 0 ) { + print "Enter password: "; + ReadMode('noecho'); + $opts{p} = <STDIN>; + chomp($opts{p}); + ReadMode('normal'); +} + +# Meta-data (table definitions etc) {{{2 +# ########################################################################### + +# Expressions {{{3 +# Convenience so I can copy/paste these in several places... +# ########################################################################### +my %exprs = ( + Host => q{my $host = host || hostname || ''; ($host) = $host =~ m/^((?:[\d.]+(?=:|$))|(?:[a-z0-9-]+))/i; return $host || ''}, + Port => q{my ($p) = host =~ m/:(.*)$/; return $p || 0}, + OldVersions => q{dulint_to_int(IB_tx_trx_id_counter) - dulint_to_int(IB_tx_purge_done_for)}, + MaxTxnTime => q/max(map{ $_->{active_secs} } @{ IB_tx_transactions }) || 0/, + NumTxns => q{scalar @{ IB_tx_transactions } }, + DirtyBufs => q{ $cur->{IB_bp_pages_modified} / ($cur->{IB_bp_buf_pool_size} || 1) }, + BufPoolFill => q{ $cur->{IB_bp_pages_total} / ($cur->{IB_bp_buf_pool_size} || 1) }, + ServerLoad => q{ $cur->{Threads_connected}/(Questions||1)/Uptime_hires }, + Connection => q{ max_connections || $cur->{Threads_connected} }, + chcxn_2_cxn => q{ if ( defined($cur->{cxn}) ) { return $cur->{cxn}; } else { my ($cha, $conn) = split ("=",$cur->{chcxn}) ; return $conn; } }, + chcxn_2_ch => q{ if ( defined($cur->{channel_name}) ) { return $cur->{channel_name}; } else { my ($cha, $conn) = split ("=",$cur->{chcxn}) ; $cha = '' if ($cha = /no_channels/); return $cha || 'failed'; } }, + TxnTimeRemain => q{ defined undo_log_entries && defined $pre->{undo_log_entries} && undo_log_entries < $pre->{undo_log_entries} ? undo_log_entries / (($pre->{undo_log_entries} - undo_log_entries)/((active_secs-$pre->{active_secs})||1))||1 : 0}, + SlaveCatchupRate => ' defined $cur->{seconds_behind_master} && defined $pre->{seconds_behind_master} && $cur->{seconds_behind_master} < $pre->{seconds_behind_master} ? ($pre->{seconds_behind_master}-$cur->{seconds_behind_master})/($cur->{Uptime_hires}-$pre->{Uptime_hires}) : 0', + QcacheHitRatio => q{(Qcache_hits||0)/(((Com_select||0)+(Qcache_hits||0))||1)}, + QueryDetail => q{sprintf("%2d/%2d/%2d/%2d", + ((Com_select||0)+(Qcache_hits||0))/(Questions||1)*100, + ((Com_insert||0)+(Com_replace||0))/(Questions||1)*100, + (Com_update||0)/(Questions||1)*100, + (Com_delete||0)/(Questions||1)*100, +)}, +); + +# ########################################################################### +# Column definitions {{{3 +# Defines every column in every table. A named column has the following +# properties: +# * hdr Column header/title +# * label Documentation for humans. +# * num Whether it's numeric (for sorting). +# * just Alignment; generated from num, user-overridable in tbl_meta +# * minw, maxw Auto-generated, user-overridable. +# Values from this hash are just copied to tbl_meta, which is where everything +# else in the program should read from. +# ########################################################################### + +my %columns = ( + active_secs => { hdr => 'SecsActive', num => 1, label => 'Seconds transaction has been active', }, + add_pool_alloc => { hdr => 'Add\'l Pool', num => 1, label => 'Additional pool allocated' }, + attempted_op => { hdr => 'Action', num => 0, label => 'The action that caused the error' }, + awe_mem_alloc => { hdr => 'AWE Memory', num => 1, label => '[Windows] AWE memory allocated' }, + binlog_cache_overflow => { hdr => 'Binlog Cache', num => 1, label => 'Transactions too big for binlog cache that went to disk' }, + binlog_do_db => { hdr => 'Binlog Do DB', num => 0, label => 'binlog-do-db setting' }, + binlog_ignore_db => { hdr => 'Binlog Ignore DB', num => 0, label => 'binlog-ignore-db setting' }, + blocking_thread => { hdr => 'BThread', num => 1, label => 'Blocking thread' }, + blocking_query => { hdr => 'Blocking Query', num => 0, label => 'Blocking query' }, + blocking_rows_modified => { hdr => 'BRowsMod', num => 1, label => 'Blocking number rows modified' }, + blocking_age => { hdr => 'BAge', num => 1, label => 'Blocking age' }, + blocking_wait_secs => { hdr => 'BWait', num => 1, label => 'Blocking wait time' }, + blocking_user => { hdr => 'BUser', num => 0, label => 'Blocking user' }, + blocking_host => { hdr => 'BHost', num => 0, label => 'Blocking host' }, + blocking_db => { hdr => 'BDB', num => 0, label => 'Blocking database' }, + blocking_status => { hdr => 'BStatus', num => 0, label => 'Blocking thread status' }, + bps_in => { hdr => 'BpsIn', num => 1, label => 'Bytes per second received by the server', }, + bps_out => { hdr => 'BpsOut', num => 1, label => 'Bytes per second sent by the server', }, + buf_free => { hdr => 'Free Bufs', num => 1, label => 'Buffers free in the buffer pool' }, + buf_pool_hit_rate => { hdr => 'Hit Rate', num => 0, label => 'Buffer pool hit rate' }, + buf_pool_hits => { hdr => 'Hits', num => 1, label => 'Buffer pool hits' }, + buf_pool_reads => { hdr => 'Reads', num => 1, label => 'Buffer pool reads' }, + buf_pool_size => { hdr => 'Size', num => 1, label => 'Buffer pool size' }, + bufs_in_node_heap => { hdr => 'Node Heap Bufs', num => 1, label => 'Buffers in buffer pool node heap' }, + bytes_behind_master => { hdr => 'ByteLag', num => 1, label => 'Bytes the slave lags the master in binlog' }, + cell_event_set => { hdr => 'Ending?', num => 1, label => 'Whether the cell event is set' }, + cell_waiting => { hdr => 'Waiting?', num => 1, label => 'Whether the cell is waiting' }, + channel_name => { hdr => 'Channel', num => 0, label => 'The name of the replication channel' }, + child_db => { hdr => 'Child DB', num => 0, label => 'The database of the child table' }, + child_index => { hdr => 'Child Index', num => 0, label => 'The index in the child table' }, + child_table => { hdr => 'Child Table', num => 0, label => 'The child table' }, + cmd => { hdr => 'Cmd', num => 0, label => 'Type of command being executed', }, + cnt => { hdr => 'Cnt', num => 0, label => 'Count', agg => 'count', aggonly => 1 }, + connections => { hdr => 'Cxns', num => 1, label => 'Connections' }, + connect_retry => { hdr => 'Connect Retry', num => 1, label => 'Slave connect-retry timeout' }, + cxn => { hdr => 'CXN', num => 0, label => 'Connection from which the data came', }, + db => { hdr => 'DB', num => 0, label => 'Current database', }, + dict_mem_alloc => { hdr => 'Dict Mem', num => 1, label => 'Dictionary memory allocated' }, + dirty_bufs => { hdr => 'Dirty Buf', num => 1, label => 'Dirty buffer pool pages' }, + dl_txn_num => { hdr => 'Num', num => 0, label => 'Deadlocked transaction number', }, + event_set => { hdr => 'Evt Set?', num => 1, label => '[Win32] if a wait event is set', }, + exec_master_log_pos => { hdr => 'Exec Master Log Pos', num => 1, label => 'Exec Master Log Position' }, + executed_gtid_set => { hdr => 'Executed GTID Set', num => 0, label => 'Executed GTID Set', }, + fk_name => { hdr => 'Constraint', num => 0, label => 'The name of the FK constraint' }, + free_list_len => { hdr => 'Free List Len', num => 1, label => 'Length of the free list' }, + has_read_view => { hdr => 'Rd View', num => 1, label => 'Whether the transaction has a read view' }, + hash_searches_s => { hdr => 'Hash/Sec', num => 1, label => 'Number of hash searches/sec' }, + hash_table_size => { hdr => 'Size', num => 1, label => 'Number of non-hash searches/sec' }, + heap_no => { hdr => 'Heap', num => 1, label => 'Heap number' }, + heap_size => { hdr => 'Heap', num => 1, label => 'Heap size' }, + history_list_len => { hdr => 'History', num => 1, label => 'History list length' }, + host_and_domain => { hdr => 'Host', num => 0, label => 'Hostname/IP and domain' }, + host_and_port => { hdr => 'Host/IP', num => 0, label => 'Hostname or IP address, and port number', }, + hostname => { hdr => 'Host', num => 0, label => 'Hostname' }, + index => { hdr => 'Index', num => 0, label => 'The index involved', agghide => 1 }, + index_ref => { hdr => 'Index Ref', num => 0, label => 'Index referenced' }, + info => { hdr => 'Query', num => 0, label => 'Info or the current query', }, + insert_intention => { hdr => 'Ins Intent', num => 1, label => 'Whether the thread was trying to insert' }, + inserts => { hdr => 'Inserts', num => 1, label => 'Inserts' }, + io_bytes_s => { hdr => 'Bytes/Sec', num => 1, label => 'Average I/O bytes/sec' }, + io_flush_type => { hdr => 'Flush Type', num => 0, label => 'I/O Flush Type' }, + io_fsyncs_s => { hdr => 'fsyncs/sec', num => 1, label => 'I/O fsyncs/sec' }, + io_reads_s => { hdr => 'Reads/Sec', num => 1, label => 'Average I/O reads/sec' }, + io_writes_s => { hdr => 'Writes/Sec', num => 1, label => 'Average I/O writes/sec' }, + ip => { hdr => 'IP', num => 0, label => 'IP address' }, + is_name_locked => { hdr => 'Locked', num => 1, label => 'Whether table is name locked', }, + key_buffer_hit => { hdr => 'KCacheHit', num => 1, label => 'Key cache hit ratio', }, + key_len => { hdr => 'Key Length', num => 1, label => 'Number of bytes used in the key' }, + last_chkp => { hdr => 'Last Checkpoint', num => 0, label => 'Last log checkpoint' }, + last_errno => { hdr => 'Last Errno', num => 1, label => 'Last error number' }, + last_error => { hdr => 'Last Error', num => 0, label => 'Last error' }, + last_s_file_name => { hdr => 'S-File', num => 0, label => 'Filename where last read locked' }, + last_s_line => { hdr => 'S-Line', num => 1, label => 'Line where last read locked' }, + last_x_file_name => { hdr => 'X-File', num => 0, label => 'Filename where last write locked' }, + last_x_line => { hdr => 'X-Line', num => 1, label => 'Line where last write locked' }, + last_pct => { hdr => 'Pct', num => 1, label => 'Last Percentage' }, + last_total => { hdr => 'Last Total', num => 1, label => 'Last Total' }, + last_value => { hdr => 'Last Incr', num => 1, label => 'Last Value' }, + load => { hdr => 'Load', num => 1, label => 'Server load' }, + locked_count => { hdr => 'Lock', num => 1, label => 'Number of locked threads' }, + lock_cfile_name => { hdr => 'Crtd File', num => 0, label => 'Filename where lock created' }, + lock_cline => { hdr => 'Crtd Line', num => 1, label => 'Line where lock created' }, + lock_info => { hdr => 'Lock Info', num => 0, label => 'Lock information' }, + lock_mem_addr => { hdr => 'Addr', num => 0, label => 'The lock memory address' }, + lock_mode => { hdr => 'Mode', num => 0, label => 'The lock mode' }, + lock_structs => { hdr => 'LStrcts', num => 1, label => 'Number of lock structs' }, + lock_type => { hdr => 'Type', num => 0, label => 'The lock type' }, + lock_var => { hdr => 'Lck Var', num => 1, label => 'The lock variable' }, + lock_wait_time => { hdr => 'Wait', num => 1, label => 'How long txn has waited for a lock' }, + log_flushed_to => { hdr => 'Flushed To', num => 0, label => 'Log position flushed to' }, + log_ios_done => { hdr => 'IO Done', num => 1, label => 'Log I/Os done' }, + log_ios_s => { hdr => 'IO/Sec', num => 1, label => 'Average log I/Os per sec' }, + log_seq_no => { hdr => 'Sequence No.', num => 0, label => 'Log sequence number' }, + longest_sql => { hdr => 'SQL', num => 0, label => 'Longest-running SQL statement' }, + main_thread_id => { hdr => 'Main Thread ID', num => 1, label => 'Main thread ID' }, + main_thread_proc_no => { hdr => 'Main Thread Proc', num => 1, label => 'Main thread process number' }, + main_thread_state => { hdr => 'Main Thread State', num => 0, label => 'Main thread state' }, + master_file => { hdr => 'File', num => 0, label => 'Master file' }, + master_host => { hdr => 'Master', num => 0, label => 'Master server hostname' }, + master_log_file => { hdr => 'Master Log File', num => 0, label => 'Master log file' }, + master_port => { hdr => 'Master Port', num => 1, label => 'Master port' }, + master_pos => { hdr => 'Position', num => 1, label => 'Master position' }, + master_ssl_allowed => { hdr => 'Master SSL Allowed', num => 0, label => 'Master SSL Allowed' }, + master_ssl_ca_file => { hdr => 'Master SSL CA File', num => 0, label => 'Master SSL Cert Auth File' }, + master_ssl_ca_path => { hdr => 'Master SSL CA Path', num => 0, label => 'Master SSL Cert Auth Path' }, + master_ssl_cert => { hdr => 'Master SSL Cert', num => 0, label => 'Master SSL Cert' }, + master_ssl_cipher => { hdr => 'Master SSL Cipher', num => 0, label => 'Master SSL Cipher' }, + master_ssl_key => { hdr => 'Master SSL Key', num => 0, label => 'Master SSL Key' }, + master_user => { hdr => 'Master User', num => 0, label => 'Master username' }, + master_uuid => { hdr => 'Master UUID', num => 0, label => 'Master UUID', }, + max_txn => { hdr => 'MaxTxnTime', num => 1, label => 'MaxTxn' }, + max_query_time => { hdr => 'MaxSQL', num => 1, label => 'Longest running SQL' }, + merged_recs => { hdr => 'Merged Recs', num => 1, label => 'Merged records' }, + merges => { hdr => 'Merges', num => 1, label => 'Merges' }, + miss_rate => { hdr => 'Miss', num => 1, label => 'InnoDB buffer pool miss rate' }, + mutex_os_waits => { hdr => 'Waits', num => 1, label => 'Mutex OS Waits' }, + mutex_spin_rounds => { hdr => 'Rounds', num => 1, label => 'Mutex Spin Rounds' }, + mutex_spin_waits => { hdr => 'Spins', num => 1, label => 'Mutex Spin Waits' }, + mysql_thread_id => { hdr => 'ID', num => 1, label => 'MySQL connection (thread) ID', }, + name => { hdr => 'Name', num => 0, label => 'Variable Name' }, + n_bits => { hdr => '# Bits', num => 1, label => 'Number of bits' }, + non_hash_searches_s => { hdr => 'Non-Hash/Sec', num => 1, label => 'Non-hash searches/sec' }, + num_deletes => { hdr => 'Del', num => 1, label => 'Number of deletes' }, + num_deletes_sec => { hdr => 'Del/Sec', num => 1, label => 'Number of deletes' }, + num_inserts => { hdr => 'Ins', num => 1, label => 'Number of inserts' }, + num_inserts_sec => { hdr => 'Ins/Sec', num => 1, label => 'Number of inserts' }, + num_readers => { hdr => 'Readers', num => 1, label => 'Number of readers' }, + num_reads => { hdr => 'Read', num => 1, label => 'Number of reads' }, + num_reads_sec => { hdr => 'Read/Sec', num => 1, label => 'Number of reads' }, + num_res_ext => { hdr => 'BTree Extents', num => 1, label => 'Number of extents reserved for B-Tree' }, + num_rows => { hdr => 'Row Count', num => 1, label => 'Number of rows estimated to examine' }, + num_times_open => { hdr => 'In Use', num => 1, label => '# times table is opened', }, + num_txns => { hdr => 'Txns', num => 1, label => 'Number of transactions' }, + num_updates => { hdr => 'Upd', num => 1, label => 'Number of updates' }, + num_updates_sec => { hdr => 'Upd/Sec', num => 1, label => 'Number of updates' }, + 'open' => { hdr => 'Tbls', num => 1, label => 'Number of open tables' }, + os_file_reads => { hdr => 'OS Reads', num => 1, label => 'OS file reads' }, + os_file_writes => { hdr => 'OS Writes', num => 1, label => 'OS file writes' }, + os_fsyncs => { hdr => 'OS fsyncs', num => 1, label => 'OS fsyncs' }, + os_thread_id => { hdr => 'OS Thread', num => 1, label => 'The operating system thread ID' }, + p_aio_writes => { hdr => 'Async Wrt', num => 1, label => 'Pending asynchronous I/O writes' }, + p_buf_pool_flushes => { hdr => 'Buffer Pool Flushes', num => 1, label => 'Pending buffer pool flushes' }, + p_ibuf_aio_reads => { hdr => 'IBuf Async Rds', num => 1, label => 'Pending insert buffer asynch I/O reads' }, + p_log_flushes => { hdr => 'Log Flushes', num => 1, label => 'Pending log flushes' }, + p_log_ios => { hdr => 'Log I/Os', num => 1, label => 'Pending log I/O operations' }, + p_normal_aio_reads => { hdr => 'Async Rds', num => 1, label => 'Pending asynchronous I/O reads' }, + p_preads => { hdr => 'preads', num => 1, label => 'Pending p-reads' }, + p_pwrites => { hdr => 'pwrites', num => 1, label => 'Pending p-writes' }, + p_sync_ios => { hdr => 'Sync I/Os', num => 1, label => 'Pending synchronous I/O operations' }, + page_creates_sec => { hdr => 'Creates/Sec', num => 1, label => 'Page creates/sec' }, + page_no => { hdr => 'Page', num => 1, label => 'Page number' }, + page_reads_sec => { hdr => 'Reads/Sec', num => 1, label => 'Page reads per second' }, + page_writes_sec => { hdr => 'Writes/Sec', num => 1, label => 'Page writes per second' }, + pages_created => { hdr => 'Created', num => 1, label => 'Pages created' }, + pages_modified => { hdr => 'Dirty Pages', num => 1, label => 'Pages modified (dirty)' }, + pages_read => { hdr => 'Reads', num => 1, label => 'Pages read' }, + pages_total => { hdr => 'Pages', num => 1, label => 'Pages total' }, + pages_written => { hdr => 'Writes', num => 1, label => 'Pages written' }, + parent_col => { hdr => 'Parent Column', num => 0, label => 'The referred column in the parent table', }, + parent_db => { hdr => 'Parent DB', num => 0, label => 'The database of the parent table' }, + parent_index => { hdr => 'Parent Index', num => 0, label => 'The referred index in the parent table' }, + parent_table => { hdr => 'Parent Table', num => 0, label => 'The parent table' }, + part_id => { hdr => 'Part ID', num => 1, label => 'Sub-part ID of the query' }, + partitions => { hdr => 'Partitions', num => 0, label => 'Query partitions used' }, + pct => { hdr => 'Pct', num => 1, label => 'Percentage' }, + pending_chkp_writes => { hdr => 'Chkpt Writes', num => 1, label => 'Pending log checkpoint writes' }, + pending_log_writes => { hdr => 'Log Writes', num => 1, label => 'Pending log writes' }, + port => { hdr => 'Port', num => 1, label => 'Client port number', }, + possible_keys => { hdr => 'Poss. Keys', num => 0, label => 'Possible keys' }, + proc_no => { hdr => 'Proc', num => 1, label => 'Process number' }, + q_detail => { hdr => 'Se/In/Up/De%', num => 0, label => 'Detailed Query', }, + q_cache_hit => { hdr => 'QCacheHit', num => 1, label => 'Query cache hit ratio', }, + qps => { hdr => 'QPS', num => 1, label => 'How many queries/sec', }, + queries_in_queue => { hdr => 'Queries Queued', num => 1, label => 'Queries in queue' }, + queries_inside => { hdr => 'Queries Inside', num => 1, label => 'Queries inside InnoDB' }, + query_id => { hdr => 'Query ID', num => 1, label => 'Query ID' }, + query_status => { hdr => 'Query Status', num => 0, label => 'The query status' }, + query_text => { hdr => 'Query Text', num => 0, label => 'The query text' }, + questions => { hdr => 'Questions', num => 1, label => 'How many queries the server has gotten', }, + read_master_log_pos => { hdr => 'Read Master Pos', num => 1, label => 'Read master log position' }, + read_views_open => { hdr => 'Rd Views', num => 1, label => 'Number of read views open' }, + reads_pending => { hdr => 'Pending Reads', num => 1, label => 'Reads pending' }, + relay_log_file => { hdr => 'Relay File', num => 0, label => 'Relay log file' }, + relay_log_pos => { hdr => 'Relay Pos', num => 1, label => 'Relay log position' }, + relay_log_size => { hdr => 'Relay Size', num => 1, label => 'Relay log size' }, + relay_master_log_file => { hdr => 'Relay Master File', num => 0, label => 'Relay master log file' }, + replicate_do_db => { hdr => 'Do DB', num => 0, label => 'Replicate-do-db setting' }, + replicate_do_table => { hdr => 'Do Table', num => 0, label => 'Replicate-do-table setting' }, + replicate_ignore_db => { hdr => 'Ignore DB', num => 0, label => 'Replicate-ignore-db setting' }, + replicate_ignore_table => { hdr => 'Ignore Table', num => 0, label => 'Replicate-do-table setting' }, + replicate_wild_do_table => { hdr => 'Wild Do Table', num => 0, label => 'Replicate-wild-do-table setting' }, + replicate_wild_ignore_table => { hdr => 'Wild Ignore Table', num => 0, label => 'Replicate-wild-ignore-table setting' }, + request_type => { hdr => 'Type', num => 0, label => 'Type of lock the thread waits for' }, + reservation_count => { hdr => 'ResCnt', num => 1, label => 'Reservation Count' }, + retrieved_gtid_set => { hdr => 'Retrieved GTID Set', num => 0, label => 'Retrieved GTID Set', }, + row_locks => { hdr => 'RLocks', num => 1, label => 'Number of row locks' }, + rows_changed => { hdr => 'Changed', num => 1, label => 'Number of rows changed' }, + rows_changed_x_indexes => { hdr => 'Chg X Idx', num => 1, label => 'Number of rows changed X indexes' }, + rows_read => { hdr => 'Reads', num => 1, label => 'Number of rows read' }, + rows_read_from_indexes => { hdr => 'Reads Via Idx', num => 1, label => 'Number of rows read from indexes' }, + run => { hdr => 'Run', num => 1, label => 'Threads_running' }, + rw_excl_os_waits => { hdr => 'RW Waits', num => 1, label => 'R/W Excl. OS Waits' }, + rw_excl_spins => { hdr => 'RW Spins', num => 1, label => 'R/W Excl. Spins' }, + rw_shared_os_waits => { hdr => 'Sh Waits', num => 1, label => 'R/W Shared OS Waits' }, + rw_shared_spins => { hdr => 'Sh Spins', num => 1, label => 'R/W Shared Spins' }, + spark_qps => { hdr => 'QPS', num => 0, label => 'QPS Sparkline' }, + spark_run => { hdr => 'Run', num => 0, label => 'Threads_running Sparkline' }, + scan_type => { hdr => 'Type', num => 0, label => 'Scan type in chosen' }, + seg_size => { hdr => 'Seg. Size', num => 1, label => 'Segment size' }, + select_type => { hdr => 'Select Type', num => 0, label => 'Type of select used' }, + server_uuid => { hdr => 'Server UUID', num => 0, label => 'Server UUID', }, + signal_count => { hdr => 'Signals', num => 1, label => 'Signal Count' }, + size => { hdr => 'Size', num => 1, label => 'Size of the tablespace' }, + skip_counter => { hdr => 'Skip Counter', num => 1, label => 'Skip counter' }, + slave_catchup_rate => { hdr => 'Catchup', num => 1, label => 'How fast the slave is catching up in the binlog' }, + slave_io_running => { hdr => 'Slave-IO', num => 0, label => 'Whether the slave I/O thread is running' }, + slave_io_state => { hdr => 'Slave IO State', num => 0, label => 'Slave I/O thread state' }, + slave_open_temp_tables => { hdr => 'Temp', num => 1, label => 'Slave open temp tables' }, + slave_running => { hdr => 'Repl', num => 0, label => 'Slave running' }, + slave_sql_running => { hdr => 'Slave-SQL', num => 0, label => 'Whether the slave SQL thread is running' }, + sort_time => { hdr => 'SortTimeLag', num => 1, label => 'Time slave lags master sort' }, + slow => { hdr => 'Slow', num => 1, label => 'How many slow queries', }, + space_id => { hdr => 'Space', num => 1, label => 'Tablespace ID' }, + special => { hdr => 'Special', num => 0, label => 'Special/Other info' }, + state => { hdr => 'State', num => 0, label => 'Connection state', maxw => 18, }, + tables_in_use => { hdr => 'Tbl Used', num => 1, label => 'Number of tables in use' }, + tables_locked => { hdr => 'Tbl Lck', num => 1, label => 'Number of tables locked' }, + tbl => { hdr => 'Table', num => 0, label => 'Table', }, + thread => { hdr => 'Thread', num => 1, label => 'Thread number' }, + thread_decl_inside => { hdr => 'Thread Inside', num => 0, label => 'What the thread is declared inside' }, + thread_purpose => { hdr => 'Purpose', num => 0, label => "The thread's purpose" }, + thread_status => { hdr => 'Thread Status', num => 0, label => 'The thread status' }, + time => { hdr => 'Time', num => 1, label => 'Time since the last event', }, + time_behind_master => { hdr => 'TimeLag', num => 1, label => 'Time slave lags master' }, + timestring => { hdr => 'Timestring', num => 0, label => 'Time the event occurred' }, + total => { hdr => 'Total', num => 1, label => 'Total' }, + total_mem_alloc => { hdr => 'Memory', num => 1, label => 'Total memory allocated' }, + truncates => { hdr => 'Trunc', num => 0, label => 'Whether the deadlock is truncating InnoDB status' }, + txn_doesnt_see_ge => { hdr => "Txn Won't See", num => 0, label => 'Where txn read view is limited' }, + txn_id => { hdr => 'ID', num => 0, label => 'Transaction ID' }, + txn_sees_lt => { hdr => 'Txn Sees', num => 1, label => 'Where txn read view is limited' }, + txn_status => { hdr => 'Txn Status', num => 0, label => 'Transaction status' }, + txn_time_remain => { hdr => 'Remaining', num => 1, label => 'Time until txn rollback/commit completes' }, + undo_log_entries => { hdr => 'Undo', num => 1, label => 'Number of undo log entries' }, + undo_for => { hdr => 'Undo', num => 0, label => 'Undo for' }, + until_condition => { hdr => 'Until Condition', num => 0, label => 'Slave until condition' }, + until_log_file => { hdr => 'Until Log File', num => 0, label => 'Slave until log file' }, + uptime => { hdr => 'Uptime', num => 1, label => 'Uptime' }, + until_log_pos => { hdr => 'Until Log Pos', num => 1, label => 'Slave until log position' }, + used_cells => { hdr => 'Cells Used', num => 1, label => 'Number of cells used' }, + used_bufs => { hdr => 'Used Bufs', num => 1, label => 'Number of buffer pool pages used' }, + user => { hdr => 'User', num => 0, label => 'Database username', }, + value => { hdr => 'Value', num => 1, label => 'Value' }, + versions => { hdr => 'Versions', num => 1, label => 'Number of InnoDB MVCC versions unpurged' }, + victim => { hdr => 'Victim', num => 0, label => 'Whether this txn was the deadlock victim' }, + wait_array_size => { hdr => 'Wait Array Size', num => 1, label => 'Wait Array Size' }, + wait_status => { hdr => 'Lock Status', num => 0, label => 'Status of txn locks' }, + waited_at_filename => { hdr => 'File', num => 0, label => 'Filename at which thread waits' }, + waited_at_line => { hdr => 'Line', num => 1, label => 'Line at which thread waits' }, + waiters_flag => { hdr => 'Waiters', num => 1, label => 'Waiters Flag' }, + waiting => { hdr => 'Waiting', num => 1, label => 'Whether lock is being waited for' }, + waiting_thread => { hdr => 'WThread', num => 1, label => 'Waiting thread' }, + waiting_query => { hdr => 'Waiting Query', num => 0, label => 'Waiting query' }, + waiting_rows_modified => { hdr => 'WRowsMod', num => 1, label => 'Waiting number rows modified' }, + waiting_age => { hdr => 'WAge', num => 1, label => 'Waiting age' }, + waiting_wait_secs => { hdr => 'WWait', num => 1, label => 'Waiting wait time' }, + waiting_user => { hdr => 'WUser', num => 0, label => 'Waiting user' }, + waiting_host => { hdr => 'WHost', num => 0, label => 'Waiting host' }, + waiting_db => { hdr => 'WDB', num => 0, label => 'Waiting database' }, + when => { hdr => 'When', num => 0, label => 'Time scale' }, + writer_lock_mode => { hdr => 'Wrtr Lck Mode', num => 0, label => 'Writer lock mode' }, + writer_thread => { hdr => 'Wrtr Thread', num => 1, label => 'Writer thread ID' }, + writes_pending => { hdr => 'Writes', num => 1, label => 'Number of writes pending' }, + writes_pending_flush_list => { hdr => 'Flush List Writes', num => 1, label => 'Number of flush list writes pending' }, + writes_pending_lru => { hdr => 'LRU Writes', num => 1, label => 'Number of LRU writes pending' }, + writes_pending_single_page => { hdr => '1-Page Writes', num => 1, label => 'Number of 1-page writes pending' }, +); + +# Apply a default property or three. By default, columns are not width-constrained, +# aligned left, and sorted alphabetically, not numerically. +foreach my $col ( values %columns ) { + map { $col->{$_} ||= 0 } qw(num minw maxw); + $col->{just} = $col->{num} ? '' : '-'; +} + +# Filters {{{3 +# This hash defines every filter that can be applied to a table. These +# become part of tbl_meta as well. Each filter is just an expression that +# returns true or false. +# Properties of each entry: +# * func: the subroutine +# * name: the name, repeated +# * user: whether it's a user-defined filter (saved in config) +# * text: text of the subroutine +# * note: explanation +my %filters = (); + +# These are pre-processed to live in %filters above, by compiling them. +my %builtin_filters = ( + hide_self => { + text => <<' END', + return ( ($set->{info} || '') !~ m#/\*innotop\*/# ); + END + note => 'Removes the innotop processes from the list', + tbls => [qw(innodb_transactions processlist)], + }, + hide_inactive => { + text => <<' END', + return ( !defined($set->{txn_status}) || $set->{txn_status} ne 'not started' ) + && ( !defined($set->{cmd}) || $set->{cmd} !~ m/Sleep|Binlog Dump/ ) + && ( !defined($set->{state}) || $set->{state} !~ m/^handlersocket/ ) + && ( !defined($set->{info}) || $set->{info} =~ m/\S/ ); + END + note => 'Removes processes which are not doing anything', + tbls => [qw(innodb_transactions processlist)], + }, + hide_connect => { + text => <<' END', + return ( !defined($set->{cmd}) || $set->{cmd} !~ m/Connect/ ); + END + note => 'Removes the slave processes from the list', + tbls => [qw(processlist)], + }, + hide_slave_io => { + text => <<' END', + return !$set->{state} || $set->{state} !~ m/^(?:Waiting for master|read all relay)/; + END + note => 'Removes slave I/O threads from the list', + tbls => [qw(processlist slave_io_status)], + }, + hide_event => { + text => <<' END', + return (($set->{state} || '') !~ m/^Daemon/) || (($set->{info} || '') !~ m/\S/); + END + note => 'Removes idle event threads from the list', + tbls => [qw(processlist)], + }, + table_is_open => { + text => <<' END', + return $set->{num_times_open} + $set->{is_name_locked}; + END + note => 'Removes tables that are not in use or locked', + tbls => [qw(open_tables)], + }, + cxn_is_master => { + text => <<' END', + return $set->{master_file} ? 1 : 0; + END + note => 'Removes servers that are not masters', + tbls => [qw(master_status)], + }, + cxn_is_slave => { + text => <<' END', + return $set->{master_host} ? 1 : 0; + END + note => 'Removes servers that are not slaves', + tbls => [qw(slave_io_status slave_sql_status)], + }, + thd_is_not_waiting => { + text => <<' END', + return $set->{thread_status} !~ m#waiting for i/o request#; + END + note => 'Removes idle I/O threads', + tbls => [qw(io_threads)], + }, +); +foreach my $key ( keys %builtin_filters ) { + my ( $sub, $err ) = compile_filter($builtin_filters{$key}->{text}); + $filters{$key} = { + func => $sub, + text => $builtin_filters{$key}->{text}, + user => 0, + name => $key, # useful for later + note => $builtin_filters{$key}->{note}, + tbls => $builtin_filters{$key}->{tbls}, + } +} + +# Variable sets {{{3 +# Sets (arrayrefs) of variables that are used in S mode. They are read/written to +# the config file. +my %var_sets = ( + general => { + text => join( + ', ', + 'set_precision(Questions/Uptime_hires) as QPS', + 'set_precision(Com_commit/Uptime_hires) as Commit_PS', + 'set_precision((Com_rollback||0)/(Com_commit||1)) as Rollback_Commit', + 'set_precision((' + . join('+', map { "($_||0)" } + qw(Com_delete Com_delete_multi Com_insert Com_insert_select Com_replace + Com_replace_select Com_select Com_update Com_update_multi)) + . ')/(Com_commit||1)) as Write_Commit', + 'set_precision((Com_select+(Qcache_hits||0))/((' + . join('+', map { "($_||0)" } + qw(Com_delete Com_delete_multi Com_insert Com_insert_select Com_replace + Com_replace_select Com_select Com_update Com_update_multi)) + . ')||1)) as R_W_Ratio', + 'set_precision(Opened_tables/Uptime_hires) as Opens_PS', + 'percent($cur->{Open_tables}/($cur->{table_cache})) as Table_Cache_Used', + 'set_precision(Threads_created/Uptime_hires) as Threads_PS', + 'percent($cur->{Threads_cached}/($cur->{thread_cache_size}||1)) as Thread_Cache_Used', + 'percent($cur->{Max_used_connections}/($cur->{max_connections}||1)) as CXN_Used_Ever', + 'percent($cur->{Threads_connected}/($cur->{max_connections}||1)) as CXN_Used_Now', + ), + }, + commands => { + text => join( + ', ', + qw(Uptime Questions Com_delete Com_delete_multi Com_insert + Com_insert_select Com_replace Com_replace_select Com_select Com_update + Com_update_multi) + ), + }, + query_status => { + text => join( + ',', + qw( Uptime Select_full_join Select_full_range_join Select_range + Select_range_check Select_scan Slow_queries Sort_merge_passes + Sort_range Sort_rows Sort_scan) + ), + }, + innodb => { + text => join( + ',', + qw( Uptime Innodb_row_lock_current_waits Innodb_row_lock_time + Innodb_row_lock_time_avg Innodb_row_lock_time_max Innodb_row_lock_waits + Innodb_rows_deleted Innodb_rows_inserted Innodb_rows_read + Innodb_rows_updated) + ), + }, + txn => { + text => join( + ',', + qw( Uptime Com_begin Com_commit Com_rollback Com_savepoint + Com_xa_commit Com_xa_end Com_xa_prepare Com_xa_recover Com_xa_rollback + Com_xa_start) + ), + }, + key_cache => { + text => join( + ',', + qw( Uptime Key_blocks_not_flushed Key_blocks_unused Key_blocks_used + Key_read_requests Key_reads Key_write_requests Key_writes ) + ), + }, + query_cache => { + text => join( + ',', + "percent($exprs{QcacheHitRatio}) as Hit_Pct", + 'set_precision((Qcache_hits||0)/(Qcache_inserts||1)) as Hit_Ins', + 'set_precision((Qcache_lowmem_prunes||0)/Uptime_hires) as Lowmem_Prunes_sec', + 'percent(1-((Qcache_free_blocks||0)/(Qcache_total_blocks||1))) as Blocks_used', + qw( Qcache_free_blocks Qcache_free_memory Qcache_not_cached Qcache_queries_in_cache) + ), + }, + handler => { + text => join( + ',', + qw( Uptime Handler_read_key Handler_read_first Handler_read_next + Handler_read_prev Handler_read_rnd Handler_read_rnd_next Handler_delete + Handler_update Handler_write) + ), + }, + cxns_files_threads => { + text => join( + ',', + qw( Uptime Aborted_clients Aborted_connects Bytes_received Bytes_sent + Compression Connections Created_tmp_disk_tables Created_tmp_files + Created_tmp_tables Max_used_connections Open_files Open_streams + Open_tables Opened_tables Table_locks_immediate Table_locks_waited + Threads_cached Threads_connected Threads_created Threads_running) + ), + }, + prep_stmt => { + text => join( + ',', + qw( Uptime Com_dealloc_sql Com_execute_sql Com_prepare_sql Com_reset + Com_stmt_close Com_stmt_execute Com_stmt_fetch Com_stmt_prepare + Com_stmt_reset Com_stmt_send_long_data ) + ), + }, + innodb_health => { + text => join( + ',', + "$exprs{OldVersions} as OldVersions", + qw(IB_sm_mutex_spin_waits IB_sm_mutex_spin_rounds IB_sm_mutex_os_waits), + "$exprs{NumTxns} as NumTxns", + "$exprs{MaxTxnTime} as MaxTxnTime", + qw(IB_ro_queries_inside IB_ro_queries_in_queue), + "set_precision($exprs{DirtyBufs} * 100) as dirty_bufs", + "set_precision($exprs{BufPoolFill} * 100) as buf_fill", + qw(IB_bp_pages_total IB_bp_pages_read IB_bp_pages_written IB_bp_pages_created) + ), + }, + innodb_health2 => { + text => join( + ', ', + 'percent(1-((Innodb_buffer_pool_pages_free||0)/($cur->{Innodb_buffer_pool_pages_total}||1))) as BP_page_cache_usage', + 'percent(1-((Innodb_buffer_pool_reads||0)/(Innodb_buffer_pool_read_requests||1))) as BP_cache_hit_ratio', + 'Innodb_buffer_pool_wait_free', + 'Innodb_log_waits', + ), + }, + slow_queries => { + text => join( + ', ', + 'set_precision(Slow_queries/Uptime_hires) as Slow_PS', + 'set_precision(Select_full_join/Uptime_hires) as Full_Join_PS', + 'percent(Select_full_join/(Com_select||1)) as Full_Join_Ratio', + ), + }, +); + +# Server sets {{{3 +# Defines sets of servers between which the user can quickly switch. +my %server_groups; + +# Connections {{{3 +# This hash defines server connections. Each connection is a string that can be passed to +# the DBI connection. These are saved in the connections section in the config file. +my %connections; +# Defines the parts of connections. +my @conn_parts = qw(user have_user pass have_pass dsn savepass dl_table); + +# Graph widths {{{3 +# This hash defines the max values seen for various status/variable values, for graphing. +# These are stored in their own section in the config file. These are just initial values: +my %mvs = ( + Com_select => 50, + Com_insert => 50, + Com_update => 50, + Com_delete => 50, + Questions => 100, +); + +# ########################################################################### +# Valid Term::ANSIColor color strings. +# ########################################################################### +my %ansicolors = map { $_ => 1 } + qw( black blink blue bold clear concealed cyan dark green magenta on_black + on_blue on_cyan on_green on_magenta on_red on_white on_yellow red reset + reverse underline underscore white yellow); + +# ########################################################################### +# Valid comparison operators for color rules +# ########################################################################### +my %comp_ops = ( + '==' => 'Numeric equality', + '>' => 'Numeric greater-than', + '<' => 'Numeric less-than', + '>=' => 'Numeric greater-than/equal', + '<=' => 'Numeric less-than/equal', + '!=' => 'Numeric not-equal', + 'eq' => 'String equality', + 'gt' => 'String greater-than', + 'lt' => 'String less-than', + 'ge' => 'String greater-than/equal', + 'le' => 'String less-than/equal', + 'ne' => 'String not-equal', + '=~' => 'Pattern match', + '!~' => 'Negated pattern match', +); + +# ########################################################################### +# Valid aggregate functions. +# ########################################################################### +my %agg_funcs = ( + first => sub { + return $_[0] + }, + count => sub { + return 0 + @_; + }, + avg => sub { + my @args = grep { defined $_ } @_; + return (sum(map { m/([\d\.-]+)/g } @args) || 0) / (scalar(@args) || 1); + }, + sum => sub { + my @args = grep { defined $_ } @_; + return sum(@args); + } +); + +# ########################################################################### +# Valid functions for transformations. +# ########################################################################### +my %trans_funcs = ( + shorten => \&shorten, + secs_to_time => \&secs_to_time, + distill => \&distill, + no_ctrl_char => \&no_ctrl_char, + percent => \&percent, + commify => \&commify, + dulint_to_int => \&dulint_to_int, + set_precision => \&set_precision, + fuzzy_time => \&fuzzy_time, +); + +# Table definitions {{{3 +# This hash defines every table that can get displayed in every mode. Each +# table specifies columns and column data sources. The column is +# defined by the %columns hash. +# +# Example: foo => { src => 'bar' } means the foo column (look at +# $columns{foo} for its definition) gets its data from the 'bar' element of +# the current data set, whatever that is. +# +# These columns are post-processed after being defined, because they get stuff +# from %columns. After all the config is loaded for columns, there's more +# post-processing too; the subroutines compiled from src get added to +# the hash elements for extract_values to use. +# ########################################################################### + +my %tbl_meta = ( + adaptive_hash_index => { + capt => 'Adaptive Hash Index', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + hash_table_size => { src => 'IB_ib_hash_table_size', trans => [qw(shorten)], }, + used_cells => { src => 'IB_ib_used_cells' }, + bufs_in_node_heap => { src => 'IB_ib_bufs_in_node_heap' }, + hash_searches_s => { src => 'IB_ib_hash_searches_s' }, + non_hash_searches_s => { src => 'IB_ib_non_hash_searches_s' }, + }, + visible => [ qw(cxn hash_table_size used_cells bufs_in_node_heap hash_searches_s non_hash_searches_s) ], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'ib', + group_by => [], + aggregate => 0, + }, + buffer_pool => { + capt => 'Buffer Pool', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + total_mem_alloc => { src => 'IB_bp_total_mem_alloc', trans => [qw(shorten)], }, + awe_mem_alloc => { src => 'IB_bp_awe_mem_alloc', trans => [qw(shorten)], }, + add_pool_alloc => { src => 'IB_bp_add_pool_alloc', trans => [qw(shorten)], }, + buf_pool_size => { src => 'IB_bp_buf_pool_size', trans => [qw(shorten)], }, + buf_free => { src => 'IB_bp_buf_free' }, + buf_pool_hit_rate => { src => 'IB_bp_buf_pool_hit_rate' }, + buf_pool_reads => { src => 'IB_bp_buf_pool_reads' }, + buf_pool_hits => { src => 'IB_bp_buf_pool_hits' }, + dict_mem_alloc => { src => 'IB_bp_dict_mem_alloc' }, + pages_total => { src => 'IB_bp_pages_total' }, + pages_modified => { src => 'IB_bp_pages_modified' }, + reads_pending => { src => 'IB_bp_reads_pending' }, + writes_pending => { src => 'IB_bp_writes_pending' }, + writes_pending_lru => { src => 'IB_bp_writes_pending_lru' }, + writes_pending_flush_list => { src => 'IB_bp_writes_pending_flush_list' }, + writes_pending_single_page => { src => 'IB_bp_writes_pending_single_page' }, + page_creates_sec => { src => 'IB_bp_page_creates_sec' }, + page_reads_sec => { src => 'IB_bp_page_reads_sec' }, + page_writes_sec => { src => 'IB_bp_page_writes_sec' }, + pages_created => { src => 'IB_bp_pages_created' }, + pages_read => { src => 'IB_bp_pages_read' }, + pages_written => { src => 'IB_bp_pages_written' }, + }, + visible => [ qw(cxn buf_pool_size buf_free pages_total pages_modified buf_pool_hit_rate total_mem_alloc add_pool_alloc)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'bp', + group_by => [], + aggregate => 0, + }, + # TODO: a new step in set_to_tbl: join result to itself, grouped? + # TODO: this would also enable pulling Q and T data together. + # TODO: using a SQL-ish language would also allow pivots to be easier -- treat the pivoted data as a view and SELECT from it. + cmd_summary => { + capt => 'Command Summary', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + name => { src => 'name' }, + total => { src => 'total' }, + value => { src => 'value', agg => 'sum'}, + pct => { src => 'value/total', trans => [qw(percent)] }, + last_total => { src => 'last_total' }, + last_value => { src => 'last_value', agg => 'sum'}, + last_pct => { src => 'last_value/last_total', trans => [qw(percent)] }, + }, + visible => [qw(cxn name value pct last_value last_pct)], + filters => [qw()], + sort_cols => '-value', + sort_dir => '1', + innodb => '', + group_by => [qw(name)], + aggregate => 1, + }, + deadlock_locks => { + capt => 'Deadlock Locks', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + mysql_thread_id => { src => 'mysql_thread_id' }, + dl_txn_num => { src => 'dl_txn_num' }, + lock_type => { src => 'lock_type' }, + space_id => { src => 'space_id' }, + page_no => { src => 'page_no' }, + heap_no => { src => 'heap_no' }, + n_bits => { src => 'n_bits' }, + index => { src => 'index' }, + db => { src => 'db' }, + tbl => { src => 'table' }, + lock_mode => { src => 'lock_mode' }, + special => { src => 'special' }, + insert_intention => { src => 'insert_intention' }, + waiting => { src => 'waiting' }, + }, + visible => [ qw(cxn mysql_thread_id waiting lock_mode db tbl index special insert_intention)], + filters => [], + sort_cols => 'cxn mysql_thread_id', + sort_dir => '1', + innodb => 'dl', + group_by => [], + aggregate => 0, + }, + deadlock_transactions => { + capt => 'Deadlock Transactions', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + active_secs => { src => 'active_secs' }, + dl_txn_num => { src => 'dl_txn_num' }, + has_read_view => { src => 'has_read_view' }, + heap_size => { src => 'heap_size' }, + host_and_domain => { src => 'hostname' }, + hostname => { src => $exprs{Host} }, + ip => { src => 'ip' }, + lock_structs => { src => 'lock_structs' }, + lock_wait_time => { src => 'lock_wait_time', trans => [ qw(secs_to_time) ] }, + mysql_thread_id => { src => 'mysql_thread_id' }, + os_thread_id => { src => 'os_thread_id' }, + proc_no => { src => 'proc_no' }, + query_id => { src => 'query_id' }, + query_status => { src => 'query_status' }, + query_text => { src => 'query_text', trans => [ qw(no_ctrl_char) ] }, + row_locks => { src => 'row_locks' }, + tables_in_use => { src => 'tables_in_use' }, + tables_locked => { src => 'tables_locked' }, + thread_decl_inside => { src => 'thread_decl_inside' }, + thread_status => { src => 'thread_status' }, + 'time' => { src => 'active_secs', trans => [ qw(secs_to_time) ] }, + timestring => { src => 'timestring' }, + txn_doesnt_see_ge => { src => 'txn_doesnt_see_ge' }, + txn_id => { src => 'txn_id' }, + txn_sees_lt => { src => 'txn_sees_lt' }, + txn_status => { src => 'txn_status' }, + truncates => { src => 'truncates' }, + undo_log_entries => { src => 'undo_log_entries' }, + user => { src => 'user' }, + victim => { src => 'victim' }, + wait_status => { src => 'lock_wait_status' }, + }, + visible => [ qw(cxn mysql_thread_id timestring user hostname victim time undo_log_entries lock_structs query_text)], + filters => [], + sort_cols => 'cxn mysql_thread_id', + sort_dir => '1', + innodb => 'dl', + group_by => [], + aggregate => 0, + }, + explain => { + capt => 'EXPLAIN Results', + cust => {}, + cols => { + part_id => { src => 'id' }, + select_type => { src => 'select_type' }, + tbl => { src => 'table' }, + partitions => { src => 'partitions' }, + scan_type => { src => 'type' }, + possible_keys => { src => 'possible_keys' }, + index => { src => 'key' }, + key_len => { src => 'key_len' }, + index_ref => { src => 'ref' }, + num_rows => { src => 'rows' }, + special => { src => 'extra' }, + }, + visible => [ qw(select_type tbl partitions scan_type possible_keys index key_len index_ref num_rows special)], + filters => [], + sort_cols => '', + sort_dir => '1', + innodb => '', + group_by => [], + aggregate => 0, + }, + file_io_misc => { + capt => 'File I/O Misc', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + io_bytes_s => { src => 'IB_io_avg_bytes_s' }, + io_flush_type => { src => 'IB_io_flush_type' }, + io_fsyncs_s => { src => 'IB_io_fsyncs_s' }, + io_reads_s => { src => 'IB_io_reads_s' }, + io_writes_s => { src => 'IB_io_writes_s' }, + os_file_reads => { src => 'IB_io_os_file_reads' }, + os_file_writes => { src => 'IB_io_os_file_writes' }, + os_fsyncs => { src => 'IB_io_os_fsyncs' }, + }, + visible => [ qw(cxn os_file_reads os_file_writes os_fsyncs io_reads_s io_writes_s io_bytes_s)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'io', + group_by => [], + aggregate => 0, + }, + fk_error => { + capt => 'Foreign Key Error Info', + cust => {}, + cols => { + timestring => { src => 'IB_fk_timestring' }, + child_db => { src => 'IB_fk_child_db' }, + child_table => { src => 'IB_fk_child_table' }, + child_index => { src => 'IB_fk_child_index' }, + fk_name => { src => 'IB_fk_fk_name' }, + parent_db => { src => 'IB_fk_parent_db' }, + parent_table => { src => 'IB_fk_parent_table' }, + parent_col => { src => 'IB_fk_parent_col' }, + parent_index => { src => 'IB_fk_parent_index' }, + attempted_op => { src => 'IB_fk_attempted_op' }, + }, + visible => [ qw(timestring child_db child_table child_index parent_db parent_table parent_col parent_index fk_name attempted_op)], + filters => [], + sort_cols => '', + sort_dir => '1', + innodb => 'fk', + group_by => [], + aggregate => 0, + }, + index_statistics => { + capt => 'Data from INDEX_STATISTICS', + cust => {}, + cols => { + cxn => { src => 'cxn', minw => 6, maxw => 10 }, + db => { src => 'table_schema' }, + tbl => { src => 'table_name' }, + index => { src => 'index_name' }, + rows_read => { src => 'rows_read', agg => 'sum' }, + }, + visible => [ qw(cxn db tbl index rows_read) ], + filters => [ ], + sort_cols => 'rows_read', + sort_dir => '-1', + innodb => '', + colors => [], + hide_caption => 1, + group_by => [qw(cxn db tbl)], + aggregate => 0, + }, + index_table_statistics => { + capt => 'Data from {TABLE,INDEX}_STATISTICS', + cust => {}, + cols => { + cxn => { src => 'cxn', minw => 6, maxw => 10 }, + db => { src => 'table_schema' }, + tbl => { src => 'table_name' }, + index => { src => 'index_name' }, + rows_read => { src => 'rows_read' }, + rows_read_from_indexes => { src => 'rows_read_from_indexes' }, + rows_changed => { src => 'rows_changed' }, + rows_changed_x_indexes => { src => 'rows_changed_x_indexes' }, + }, + visible => [ qw(cxn db tbl rows_read rows_read_from_indexes rows_changed rows_changed_x_indexes) ], + filters => [ ], + sort_cols => 'rows_read', + sort_dir => '-1', + innodb => '', + colors => [], + hide_caption => 1, + group_by => [qw(cxn db tbl)], + aggregate => 0, + }, + insert_buffers => { + capt => 'Insert Buffers', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + inserts => { src => 'IB_ib_inserts' }, + merged_recs => { src => 'IB_ib_merged_recs' }, + merges => { src => 'IB_ib_merges' }, + size => { src => 'IB_ib_size' }, + free_list_len => { src => 'IB_ib_free_list_len' }, + seg_size => { src => 'IB_ib_seg_size' }, + }, + visible => [ qw(cxn inserts merged_recs merges size free_list_len seg_size)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'ib', + group_by => [], + aggregate => 0, + }, + innodb_blocked_blocker => { + capt => 'InnoDB Blocked/Blocking', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + waiting_thread => { src => 'waiting_thread' }, + waiting_query => { src => 'waiting_query', trans => [qw(distill)] }, + waiting_rows_modified => { src => 'waiting_rows_modified' }, + waiting_age => { src => 'waiting_age', trans => [qw(fuzzy_time)] }, + waiting_wait_secs => { src => 'waiting_wait_secs', trans => [qw(fuzzy_time)] }, + waiting_user => { src => 'waiting_user' }, + waiting_host => { src => 'waiting_host' }, + waiting_db => { src => 'waiting_db' }, + blocking_thread => { src => 'blocking_thread' }, + blocking_query => { src => 'blocking_query', trans => [qw(distill)] }, + blocking_rows_modified => { src => 'blocking_rows_modified' }, + blocking_age => { src => 'blocking_age', trans => [qw(fuzzy_time)] }, + blocking_wait_secs => { src => 'blocking_wait_secs', trans => [qw(fuzzy_time)] }, + blocking_user => { src => 'blocking_user' }, + blocking_host => { src => 'blocking_host' }, + blocking_db => { src => 'blocking_db' }, + blocking_status => { src => 'blocking_status' }, + lock_info => { src => 'lock_info' }, + }, + visible => [ qw(cxn waiting_thread waiting_query waiting_wait_secs + blocking_thread blocking_rows_modified blocking_age blocking_wait_secs + blocking_status blocking_query)], + filters => [], + sort_cols => 'cxn -waiting_wait_secs', + sort_dir => '1', + innodb => 'tx', + colors => [ + ], + group_by => [], + aggregate => 0, + hide_caption => 1, + }, + innodb_locks => { + capt => 'InnoDB Locks', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + db => { src => 'db' }, + index => { src => 'index' }, + insert_intention => { src => 'insert_intention' }, + lock_mode => { src => 'lock_mode' }, + lock_type => { src => 'lock_type' }, + lock_wait_time => { src => 'lock_wait_time', trans => [ qw(secs_to_time) ] }, + mysql_thread_id => { src => 'mysql_thread_id' }, + n_bits => { src => 'n_bits' }, + page_no => { src => 'page_no' }, + space_id => { src => 'space_id' }, + special => { src => 'special' }, + tbl => { src => 'table' }, + 'time' => { src => 'active_secs', hdr => 'Active', trans => [ qw(secs_to_time) ] }, + txn_id => { src => 'txn_id' }, + waiting => { src => 'waiting' }, + }, + visible => [ qw(cxn mysql_thread_id lock_type waiting lock_wait_time time lock_mode db tbl index insert_intention special)], + filters => [], + sort_cols => 'cxn -lock_wait_time', + sort_dir => '1', + innodb => 'tx', + colors => [ + { col => 'lock_wait_time', op => '>', arg => 60, color => 'red' }, + { col => 'lock_wait_time', op => '>', arg => 30, color => 'yellow' }, + { col => 'lock_wait_time', op => '>', arg => 10, color => 'green' }, + ], + group_by => [], + aggregate => 0, + }, + innodb_transactions => { + capt => 'InnoDB Transactions', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + active_secs => { src => 'active_secs' }, + has_read_view => { src => 'has_read_view' }, + heap_size => { src => 'heap_size' }, + hostname => { src => $exprs{Host} }, + ip => { src => 'ip' }, + wait_status => { src => 'lock_wait_status' }, + lock_wait_time => { src => 'lock_wait_time', trans => [ qw(secs_to_time) ] }, + lock_structs => { src => 'lock_structs' }, + mysql_thread_id => { src => 'mysql_thread_id' }, + os_thread_id => { src => 'os_thread_id' }, + proc_no => { src => 'proc_no' }, + query_id => { src => 'query_id' }, + query_status => { src => 'query_status' }, + query_text => { src => 'query_text', trans => [ qw(no_ctrl_char) ] }, + txn_time_remain => { src => $exprs{TxnTimeRemain}, trans => [ qw(secs_to_time) ] }, + row_locks => { src => 'row_locks' }, + tables_in_use => { src => 'tables_in_use' }, + tables_locked => { src => 'tables_locked' }, + thread_decl_inside => { src => 'thread_decl_inside' }, + thread_status => { src => 'thread_status' }, + 'time' => { src => 'active_secs', trans => [ qw(secs_to_time) ], agg => 'sum' }, + txn_doesnt_see_ge => { src => 'txn_doesnt_see_ge' }, + txn_id => { src => 'txn_id' }, + txn_sees_lt => { src => 'txn_sees_lt' }, + txn_status => { src => 'txn_status', minw => 10, maxw => 10 }, + undo_log_entries => { src => 'undo_log_entries' }, + user => { src => 'user', maxw => 10 }, + cnt => { src => 'mysql_thread_id', minw => 0 }, + }, + visible => [ qw(cxn cnt mysql_thread_id user hostname txn_status time undo_log_entries query_text)], + filters => [ qw( hide_self hide_inactive ) ], + sort_cols => '-active_secs txn_status cxn mysql_thread_id', + sort_dir => '1', + innodb => 'tx', + hide_caption => 1, + colors => [ + { col => 'wait_status', op => 'eq', arg => 'LOCK WAIT', color => 'black on_red' }, + { col => 'time', op => '>', arg => 600, color => 'red' }, + { col => 'time', op => '>', arg => 300, color => 'yellow' }, + { col => 'time', op => '>', arg => 60, color => 'green' }, + { col => 'time', op => '>', arg => 30, color => 'cyan' }, + { col => 'txn_status', op => 'eq', arg => 'not started', color => 'white' }, + ], + group_by => [ qw(cxn txn_status) ], + aggregate => 0, + }, + io_threads => { + capt => 'I/O Threads', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + thread => { src => 'thread' }, + thread_purpose => { src => 'purpose' }, + event_set => { src => 'event_set' }, + thread_status => { src => 'state' }, + }, + visible => [ qw(cxn thread thread_purpose thread_status)], + filters => [ qw() ], + sort_cols => 'cxn thread', + sort_dir => '1', + innodb => 'io', + group_by => [], + aggregate => 0, + }, + log_statistics => { + capt => 'Log Statistics', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + last_chkp => { src => 'IB_lg_last_chkp' }, + log_flushed_to => { src => 'IB_lg_log_flushed_to' }, + log_ios_done => { src => 'IB_lg_log_ios_done' }, + log_ios_s => { src => 'IB_lg_log_ios_s' }, + log_seq_no => { src => 'IB_lg_log_seq_no' }, + pending_chkp_writes => { src => 'IB_lg_pending_chkp_writes' }, + pending_log_writes => { src => 'IB_lg_pending_log_writes' }, + }, + visible => [ qw(cxn log_seq_no log_flushed_to last_chkp log_ios_done log_ios_s)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'lg', + group_by => [], + aggregate => 0, + }, + master_status => { + capt => 'Master Status', + cust => {}, + cols => { + cxn => { src => $exprs{chcxn_2_cxn}, hdr => 'CXN' }, + binlog_do_db => { src => 'binlog_do_db' }, + binlog_ignore_db => { src => 'binlog_ignore_db' }, + master_file => { src => 'file' }, + master_pos => { src => 'position' }, + binlog_cache_overflow => { src => '(Binlog_cache_disk_use||0)/(Binlog_cache_use||1)', trans => [ qw(percent) ] }, + executed_gtid_set => { src => '(executed_gtid_set||"N/A")' }, + server_uuid => { src => '(server_uuid||"N/A")' }, + channel_name => { src => $exprs{chcxn_2_ch}, hdr => 'Channel'}, + }, + visible => [ qw(cxn channel_name master_file master_pos binlog_cache_overflow executed_gtid_set server_uuid)], + filters => [ qw(cxn_is_master) ], + sort_cols => 'cxn', + sort_dir => '1', + innodb => '', + group_by => [qw(channel_name)], + aggregate => 0, + }, + pending_io => { + capt => 'Pending I/O', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + p_normal_aio_reads => { src => 'IB_io_pending_normal_aio_reads' }, + p_aio_writes => { src => 'IB_io_pending_aio_writes' }, + p_ibuf_aio_reads => { src => 'IB_io_pending_ibuf_aio_reads' }, + p_sync_ios => { src => 'IB_io_pending_sync_ios' }, + p_buf_pool_flushes => { src => 'IB_io_pending_buffer_pool_flushes' }, + p_log_flushes => { src => 'IB_io_pending_log_flushes' }, + p_log_ios => { src => 'IB_io_pending_log_ios' }, + p_preads => { src => 'IB_io_pending_preads' }, + p_pwrites => { src => 'IB_io_pending_pwrites' }, + }, + visible => [ qw(cxn p_normal_aio_reads p_aio_writes p_ibuf_aio_reads p_sync_ios p_log_flushes p_log_ios)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'io', + group_by => [], + aggregate => 0, + }, + open_tables => { + capt => 'Open Tables', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + db => { src => 'database' }, + tbl => { src => 'table' }, + num_times_open => { src => 'in_use' }, + is_name_locked => { src => 'name_locked' }, + }, + visible => [ qw(cxn db tbl num_times_open is_name_locked)], + filters => [ qw(table_is_open) ], + sort_cols => '-num_times_open cxn db tbl', + sort_dir => '1', + innodb => '', + group_by => [], + aggregate => 0, + }, + page_statistics => { + capt => 'Page Statistics', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + pages_read => { src => 'IB_bp_pages_read' }, + pages_written => { src => 'IB_bp_pages_written' }, + pages_created => { src => 'IB_bp_pages_created' }, + page_reads_sec => { src => 'IB_bp_page_reads_sec' }, + page_writes_sec => { src => 'IB_bp_page_writes_sec' }, + page_creates_sec => { src => 'IB_bp_page_creates_sec' }, + }, + visible => [ qw(cxn pages_read pages_written pages_created page_reads_sec page_writes_sec page_creates_sec)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'bp', + group_by => [], + aggregate => 0, + }, + processlist => { + capt => 'MySQL Process List', + cust => {}, + cols => { + cxn => { src => 'cxn', minw => 6, maxw => 10 }, + mysql_thread_id => { src => 'id', minw => 6, maxw => 0 }, + user => { src => 'user', minw => 5, maxw => 8 }, + hostname => { src => $exprs{Host}, minw => 7, maxw => 15, }, + port => { src => $exprs{Port}, minw => 0, maxw => 0, }, + host_and_port => { src => 'host', minw => 0, maxw => 0 }, + db => { src => 'db', minw => 6, maxw => 12 }, + cmd => { src => 'command', minw => 5, maxw => 0 }, + time => { src => 'time', minw => 5, maxw => 0, trans => [ qw(secs_to_time) ], agg => 'sum' }, + state => { src => 'state', minw => 0, maxw => 0 }, + info => { src => 'info', minw => 0, maxw => 0, trans => [ qw(no_ctrl_char) ] }, + cnt => { src => 'id', minw => 0, maxw => 0 }, + }, + visible => [ qw(cxn cmd cnt mysql_thread_id state user hostname db time info)], + filters => [ qw(hide_self hide_inactive hide_slave_io hide_event hide_connect) ], + sort_cols => '-time cxn hostname mysql_thread_id', + sort_dir => '1', + innodb => '', + hide_caption => 1, + colors => [ + { col => 'state', op => 'eq', arg => 'Locked', color => 'black on_red' }, + { col => 'cmd', op => 'eq', arg => 'Sleep', color => 'white' }, + { col => 'user', op => 'eq', arg => 'system user', color => 'white' }, + { col => 'cmd', op => 'eq', arg => 'Connect', color => 'white' }, + { col => 'cmd', op => 'eq', arg => 'Binlog Dump', color => 'white' }, + { col => 'time', op => '>', arg => 600, color => 'red' }, + { col => 'time', op => '>', arg => 120, color => 'yellow' }, + { col => 'time', op => '>', arg => 60, color => 'green' }, + { col => 'time', op => '>', arg => 30, color => 'cyan' }, + ], + group_by => [qw(cxn cmd)], + aggregate => 0, + }, + + # TODO: some more columns: + # kb_used=hdr='BufUsed' minw='0' num='0' src='percent(1 - ((Key_blocks_unused * key_cache_block_size) / (key_buffer_size||1)))' dec='0' trans='' tbl='q_header' just='-' user='1' maxw='0' label='User-defined' + # retries=hdr='Retries' minw='0' num='0' src='Slave_retried_transactions' dec='0' trans='' tbl='slave_sql_status' just='-' user='1' maxw='0' label='User-defined' + # thd=hdr='Thd' minw='0' num='0' src='Threads_connected' dec='0' trans='' tbl='slave_sql_status' just='-' user='1' maxw='0' label='User-defined' + + q_header => { + capt => 'Q-mode Header', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + questions => { src => 'Questions' }, + qps => { src => 'Questions/Uptime_hires', dec => 1, trans => [qw(shorten)] }, + load => { src => $exprs{ServerLoad}, dec => 1, trans => [qw(shorten)] }, + connections => { src => $exprs{Connection}, dec => 1, trans => [qw(shorten)] }, + slow => { src => 'Slow_queries', dec => 1, trans => [qw(shorten)] }, + q_cache_hit => { src => $exprs{QcacheHitRatio}, dec => 1, trans => [qw(percent)] }, + key_buffer_hit => { src => '1-(Key_reads/(Key_read_requests||1))', dec => 1, trans => [qw(percent)] }, + bps_in => { src => 'Bytes_received/Uptime_hires', dec => 1, trans => [qw(shorten)] }, + bps_out => { src => 'Bytes_sent/Uptime_hires', dec => 1, trans => [qw(shorten)] }, + when => { src => 'when' }, + q_detail => { src => $exprs{QueryDetail} }, + }, + visible => [ qw(cxn when load connections qps slow q_detail q_cache_hit key_buffer_hit bps_in bps_out)], + filters => [], + sort_cols => 'when cxn', + sort_dir => '1', + innodb => '', + hide_caption => 1, + group_by => [], + aggregate => 0, + }, + row_operations => { + capt => 'InnoDB Row Operations', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + num_inserts => { src => 'IB_ro_num_rows_ins' }, + num_updates => { src => 'IB_ro_num_rows_upd' }, + num_reads => { src => 'IB_ro_num_rows_read' }, + num_deletes => { src => 'IB_ro_num_rows_del' }, + num_inserts_sec => { src => 'IB_ro_ins_sec' }, + num_updates_sec => { src => 'IB_ro_upd_sec' }, + num_reads_sec => { src => 'IB_ro_read_sec' }, + num_deletes_sec => { src => 'IB_ro_del_sec' }, + }, + visible => [ qw(cxn num_inserts num_updates num_reads num_deletes num_inserts_sec + num_updates_sec num_reads_sec num_deletes_sec)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'ro', + group_by => [], + aggregate => 0, + }, + row_operation_misc => { + capt => 'Row Operation Misc', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + queries_in_queue => { src => 'IB_ro_queries_in_queue' }, + queries_inside => { src => 'IB_ro_queries_inside' }, + read_views_open => { src => 'IB_ro_read_views_open' }, + main_thread_id => { src => 'IB_ro_main_thread_id' }, + main_thread_proc_no => { src => 'IB_ro_main_thread_proc_no' }, + main_thread_state => { src => 'IB_ro_main_thread_state' }, + num_res_ext => { src => 'IB_ro_n_reserved_extents' }, + }, + visible => [ qw(cxn queries_in_queue queries_inside read_views_open main_thread_state)], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'ro', + group_by => [], + aggregate => 0, + }, + semaphores => { + capt => 'InnoDB Semaphores', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + mutex_os_waits => { src => 'IB_sm_mutex_os_waits' }, + mutex_spin_rounds => { src => 'IB_sm_mutex_spin_rounds' }, + mutex_spin_waits => { src => 'IB_sm_mutex_spin_waits' }, + reservation_count => { src => 'IB_sm_reservation_count' }, + rw_excl_os_waits => { src => 'IB_sm_rw_excl_os_waits' }, + rw_excl_spins => { src => 'IB_sm_rw_excl_spins' }, + rw_shared_os_waits => { src => 'IB_sm_rw_shared_os_waits' }, + rw_shared_spins => { src => 'IB_sm_rw_shared_spins' }, + signal_count => { src => 'IB_sm_signal_count' }, + wait_array_size => { src => 'IB_sm_wait_array_size' }, + }, + visible => [ qw(cxn mutex_os_waits mutex_spin_waits mutex_spin_rounds + rw_excl_os_waits rw_excl_spins rw_shared_os_waits rw_shared_spins + signal_count reservation_count )], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => 'sm', + group_by => [], + aggregate => 0, + }, + slave_io_status => { + capt => 'Slave I/O Status', + cust => {}, + cols => { + cxn => { src => $exprs{chcxn_2_cxn}, hdr => 'CXN' }, + connect_retry => { src => 'connect_retry' }, + master_host => { src => 'master_host', hdr => 'Master'}, + master_uuid => { src => '(master_uuid||"N/A")' }, + master_log_file => { src => 'master_log_file', hdr => 'File' }, + master_port => { src => 'master_port' }, + master_ssl_allowed => { src => 'master_ssl_allowed' }, + master_ssl_ca_file => { src => 'master_ssl_ca_file' }, + master_ssl_ca_path => { src => 'master_ssl_ca_path' }, + master_ssl_cert => { src => 'master_ssl_cert' }, + master_ssl_cipher => { src => 'master_ssl_cipher' }, + master_ssl_key => { src => 'master_ssl_key' }, + master_user => { src => 'master_user' }, + read_master_log_pos => { src => 'read_master_log_pos', hdr => 'Pos' }, + relay_log_size => { src => 'relay_log_space', trans => [qw(shorten)] }, + slave_io_running => { src => 'slave_io_running', hdr => 'On?' }, + slave_io_state => { src => 'slave_io_state', hdr => 'State' }, + channel_name => { src => $exprs{chcxn_2_ch}, hdr => 'Channel'}, + }, + visible => [ qw(cxn channel_name master_host master_uuid slave_io_running master_log_file relay_log_size read_master_log_pos slave_io_state)], + filters => [ qw( cxn_is_slave ) ], + sort_cols => 'slave_io_running channel_name cxn', + colors => [ + { col => 'slave_io_running', op => 'ne', arg => 'Yes', color => 'black on_red' }, + ], + sort_dir => '1', + innodb => '', + group_by => [qw(channel_name)], + aggregate => 0, + }, + slave_sql_status => { + capt => 'Slave SQL Status', + cust => {}, + cols => { + cxn => { src => $exprs{chcxn_2_cxn}, hdr => 'CXN' }, + exec_master_log_pos => { src => 'exec_master_log_pos', hdr => 'Master Pos' }, + last_errno => { src => 'last_errno' }, + last_error => { src => 'last_error' }, + master_host => { src => 'master_host', hdr => 'Master' }, + master_uuid => { src => '(master_uuid||"N/A")' }, + relay_log_file => { src => 'relay_log_file' }, + relay_log_pos => { src => 'relay_log_pos' }, + relay_log_size => { src => 'relay_log_space', trans => [qw(shorten)] }, + relay_master_log_file => { src => 'relay_master_log_file', hdr => 'Master File' }, + replicate_do_db => { src => 'replicate_do_db' }, + replicate_do_table => { src => 'replicate_do_table' }, + replicate_ignore_db => { src => 'replicate_ignore_db' }, + replicate_ignore_table => { src => 'replicate_ignore_table' }, + replicate_wild_do_table => { src => 'replicate_wild_do_table' }, + replicate_wild_ignore_table => { src => 'replicate_wild_ignore_table' }, + skip_counter => { src => 'skip_counter' }, + slave_sql_running => { src => 'slave_sql_running', hdr => 'On?' }, + sort_time => { src => 'int(seconds_behind_master/60)' }, + until_condition => { src => 'until_condition' }, + until_log_file => { src => 'until_log_file' }, + until_log_pos => { src => 'until_log_pos' }, + time_behind_master => { src => 'seconds_behind_master', trans => [ qw(secs_to_time) ] }, + bytes_behind_master => { src => 'master_log_file && master_log_file eq relay_master_log_file ? read_master_log_pos - exec_master_log_pos : 0', trans => [qw(shorten)] }, + slave_catchup_rate => { src => $exprs{SlaveCatchupRate}, trans => [ qw(set_precision) ] }, + slave_open_temp_tables => { src => 'Slave_open_temp_tables' }, + retrieved_gtid_set => { src => '(retrieved_gtid_set||"N/A")' }, + executed_gtid_set => { src => '(executed_gtid_set||"N/A")' }, + channel_name => { src => $exprs{chcxn_2_ch}, hdr => 'Channel'}, + }, + visible => [ qw(cxn channel_name master_host master_uuid slave_sql_running time_behind_master slave_catchup_rate slave_open_temp_tables relay_log_pos last_error retrieved_gtid_set executed_gtid_set)], + filters => [ qw( cxn_is_slave ) ], + sort_cols => 'slave_sql_running -sort_time channel_name cxn', + sort_dir => '1', + innodb => '', + colors => [ + { col => 'slave_sql_running', op => 'ne', arg => 'Yes', color => 'black on_red' }, + { col => 'time_behind_master', op => '>', arg => 600, color => 'red' }, + { col => 'time_behind_master', op => '>', arg => 60, color => 'yellow' }, + { col => 'time_behind_master', op => '==', arg => 0, color => 'white' }, + ], + group_by => [qw(channel_name)], + aggregate => 0, + }, + table_statistics => { + capt => 'Data from TABLE_STATISTICS', + cust => {}, + cols => { + cxn => { src => 'cxn', minw => 6, maxw => 10 }, + db => { src => 'table_schema' }, + tbl => { src => 'table_name' }, + rows_read => { src => 'rows_read' }, + rows_changed => { src => 'rows_changed' }, + rows_changed_x_indexes => { src => 'rows_changed_x_indexes' }, + }, + visible => [ qw(cxn db tbl rows_read rows_changed rows_changed_x_indexes) ], + filters => [ ], + sort_cols => 'rows_read', + sort_dir => '-1', + innodb => '', + colors => [], + hide_caption => 1, + group_by => [], + aggregate => 0, + }, + t_header => { + capt => 'T-Mode Header', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + dirty_bufs => { src => $exprs{DirtyBufs}, trans => [qw(percent)] }, + history_list_len => { src => 'IB_tx_history_list_len' }, + lock_structs => { src => 'IB_tx_num_lock_structs' }, + num_txns => { src => $exprs{NumTxns} }, + max_txn => { src => $exprs{MaxTxnTime}, trans => [qw(secs_to_time)] }, + undo_for => { src => 'IB_tx_purge_undo_for' }, + used_bufs => { src => $exprs{BufPoolFill}, trans => [qw(percent)]}, + versions => { src => $exprs{OldVersions} }, + }, + visible => [ qw(cxn history_list_len versions undo_for dirty_bufs used_bufs num_txns max_txn lock_structs)], + filters => [ ], + sort_cols => 'cxn', + sort_dir => '1', + innodb => '', + colors => [], + hide_caption => 1, + group_by => [], + aggregate => 0, + }, + var_status => { + capt => 'Variables & Status', + cust => {}, + cols => {}, # Generated from current varset + visible => [], # Generated from current varset + filters => [], + sort_cols => '', + sort_dir => 1, + innodb => '', + temp => 1, # Do not persist to config file. + hide_caption => 1, + pivot => 0, + group_by => [], + aggregate => 0, + }, + wait_array => { + capt => 'InnoDB Wait Array', + cust => {}, + cols => { + cxn => { src => 'cxn' }, + thread => { src => 'thread' }, + waited_at_filename => { src => 'waited_at_filename' }, + waited_at_line => { src => 'waited_at_line' }, + 'time' => { src => 'waited_secs', trans => [ qw(secs_to_time) ] }, + request_type => { src => 'request_type' }, + lock_mem_addr => { src => 'lock_mem_addr' }, + lock_cfile_name => { src => 'lock_cfile_name' }, + lock_cline => { src => 'lock_cline' }, + writer_thread => { src => 'writer_thread' }, + writer_lock_mode => { src => 'writer_lock_mode' }, + num_readers => { src => 'num_readers' }, + lock_var => { src => 'lock_var' }, + waiters_flag => { src => 'waiters_flag' }, + last_s_file_name => { src => 'last_s_file_name' }, + last_s_line => { src => 'last_s_line' }, + last_x_file_name => { src => 'last_x_file_name' }, + last_x_line => { src => 'last_x_line' }, + cell_waiting => { src => 'cell_waiting' }, + cell_event_set => { src => 'cell_event_set' }, + }, + visible => [ qw(cxn thread time waited_at_filename waited_at_line request_type num_readers lock_var waiters_flag cell_waiting cell_event_set)], + filters => [], + sort_cols => 'cxn -time', + sort_dir => '1', + innodb => 'sm', + group_by => [], + aggregate => 0, + }, + health_dashboard => { + capt => 'Health Dashboard', + hide_caption => 1, + cust => {}, + cols => { + cxn => { src => 'cxn' }, + uptime => { src => 'Uptime', trans => [qw(fuzzy_time)] }, + qps => { src => 'Questions/Uptime_hires', dec => 1, trans => [qw(shorten)] }, + spark_qps => { src => 'SPARK_qps' }, + run => { src => 'User_threads_running' }, + spark_run => { src => 'SPARK_run' }, + max_query_time => { src => 'Max_query_time', trans => [qw(fuzzy_time)] }, + connections => { src => 'Threads_connected' }, + miss_rate => { src => 'Innodb_buffer_pool_reads/Uptime_hires', trans => [qw(set_precision)] }, + locked_count => { src => 'Locked_count' }, + 'open' => { src => 'Open_tables' }, + slave_running => { src => 'Slave_ok . " " . (Slaves || "")' }, + time_behind_master => { src => 'seconds_behind_master', hdr => 'ReplLag' , trans => [qw(fuzzy_time)] }, + longest_sql => { src => 'Longest_sql', trans => [qw(distill)] }, + }, + visible => [qw( + cxn uptime max_query_time time_behind_master qps connections run + miss_rate locked_count open slave_running longest_sql + )], + filters => [], + sort_cols => 'cxn', + sort_dir => '1', + innodb => '', + colors => [ + { col => 'slave_running', op => '=~', arg => 'No', color => 'black on_red' }, + { col => 'max_query_time', op => '>', arg => 30 * 60, color => 'red' }, + { col => 'max_query_time', op => '>', arg => 600, color => 'yellow' }, + { col => 'time_behind_master', op => '>', arg => 3600, color => 'cyan' }, + ], + group_by => [], + aggregate => 0, + }, +); + +# Initialize %tbl_meta from %columns and do some checks. +foreach my $table_name ( keys %tbl_meta ) { + my $table = $tbl_meta{$table_name}; + my $cols = $table->{cols}; + + foreach my $col_name ( keys %$cols ) { + my $col_def = $table->{cols}->{$col_name}; + die "I can't find a column named '$col_name' for '$table_name'" unless $columns{$col_name}; + $columns{$col_name}->{referenced} = 1; + + foreach my $prop ( keys %col_props ) { + # Each column gets non-existing values set from %columns or defaults from %col_props. + if ( !$col_def->{$prop} ) { + $col_def->{$prop} + = defined($columns{$col_name}->{$prop}) + ? $columns{$col_name}->{$prop} + : $col_props{$prop}; + } + } + + # Ensure transformations and aggregate functions are valid + die "Unknown aggregate function '$col_def->{agg}' " + . "for column '$col_name' in table '$table_name'" + unless exists $agg_funcs{$col_def->{agg}}; + foreach my $trans ( @{$col_def->{trans}} ) { + die "Unknown transformation '$trans' " + . "for column '$col_name' in table '$table_name'" + unless exists $trans_funcs{$trans}; + } + } + + # Ensure each column in visible and group_by exists in cols + foreach my $place ( qw(visible group_by) ) { + foreach my $col_name ( @{$table->{$place}} ) { + if ( !exists $cols->{$col_name} ) { + die "Column '$col_name' is listed in '$place' for '$table_name', but doesn't exist"; + } + } + } + + # Compile sort and color subroutines + $table->{sort_func} = make_sort_func($table); + $table->{color_func} = make_color_func($table); +} + +# This is for code cleanup: +{ + my @unused_cols = grep { !$columns{$_}->{referenced} } sort keys %columns; + if ( @unused_cols ) { + die "The following columns are not used: " + . join(' ', @unused_cols); + } +} + +# ########################################################################### +# Operating modes {{{3 +# ########################################################################### +my %modes = ( + A => { + hdr => 'Dashboard', + cust => {}, + note => 'Shows health/status dashboard', + action_for => { + k => { + action => sub { kill_query('CONNECTION') }, + label => "Kill a query's connection", + }, + x => { + action => sub { kill_query('QUERY') }, + label => "Kill a query", + }, + r => { + action => sub { reverse_sort('health_dashboard'); }, + label => 'Reverse sort order', + }, + s => { + action => sub { choose_sort_cols('health_dashboard'); }, + label => "Choose sort column", + }, + }, + display_sub => \&display_A, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(health_dashboard)], + visible_tables => [qw(health_dashboard)], + }, + B => { + hdr => 'InnoDB Buffers', + cust => {}, + note => 'Shows buffer info from InnoDB', + action_for => { + i => { + action => sub { toggle_config('status_inc') }, + label => 'Toggle incremental status display', + }, + }, + display_sub => \&display_B, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(buffer_pool page_statistics insert_buffers adaptive_hash_index)], + visible_tables => [qw(buffer_pool page_statistics insert_buffers adaptive_hash_index)], + }, + C => { + hdr => 'Command Summary', + cust => {}, + note => 'Shows relative magnitude of variables', + action_for => { + s => { + action => sub { get_config_interactive('cmd_filter') }, + label => 'Choose variable prefix', + }, + }, + display_sub => \&display_C, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(cmd_summary)], + visible_tables => [qw(cmd_summary)], + }, + D => { + hdr => 'InnoDB Deadlocks', + cust => {}, + note => 'View InnoDB deadlock information', + action_for => { + c => { + action => sub { edit_table('deadlock_transactions') }, + label => 'Choose visible columns', + }, + w => { + action => \&create_deadlock, + label => 'Wipe deadlock status info by creating a deadlock', + }, + }, + display_sub => \&display_D, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(deadlock_transactions deadlock_locks)], + visible_tables => [qw(deadlock_transactions deadlock_locks)], + }, + F => { + hdr => 'InnoDB FK Err', + cust => {}, + note => 'View the latest InnoDB foreign key error', + action_for => {}, + display_sub => \&display_F, + connections => [], + server_group => '', + one_connection => 1, + tables => [qw(fk_error)], + visible_tables => [qw(fk_error)], + }, + I => { + hdr => 'InnoDB I/O Info', + cust => {}, + note => 'Shows I/O info (i/o, log...) from InnoDB', + action_for => { + i => { + action => sub { toggle_config('status_inc') }, + label => 'Toggle incremental status display', + }, + }, + display_sub => \&display_I, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(io_threads pending_io file_io_misc log_statistics)], + visible_tables => [qw(io_threads pending_io file_io_misc log_statistics)], + }, + K => { + hdr => 'InnoDB Lock Waits', + cust => {}, + note => 'Shows blocked and blocking transactions', + action_for => { + k => { + action => sub { kill_query('CONNECTION') }, + label => "Kill a query's connection", + }, + }, + display_sub => \&display_K, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(innodb_blocked_blocker)], + visible_tables => [qw(innodb_blocked_blocker)], + }, + L => { + hdr => 'Locks', + cust => {}, + note => 'Shows transaction locks', + action_for => { + a => { + action => sub { send_cmd_to_servers('CREATE TABLE IF NOT EXISTS test.innodb_lock_monitor(a int) ENGINE=InnoDB', 0, '', []); }, + label => 'Start the InnoDB Lock Monitor', + }, + o => { + action => sub { send_cmd_to_servers('DROP TABLE IF EXISTS test.innodb_lock_monitor', 0, '', []); }, + label => 'Stop the InnoDB Lock Monitor', + }, + }, + display_sub => \&display_L, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(innodb_locks)], + visible_tables => [qw(innodb_locks)], + }, + M => { + hdr => 'Replication Status', + cust => {}, + note => 'Shows replication (master and slave) status', + action_for => { + a => { + action => sub { send_cmd_to_servers('START SLAVE', 0, 'START SLAVE SQL_THREAD UNTIL MASTER_LOG_FILE = ?, MASTER_LOG_POS = ?', []); }, + label => 'Start slave(s)', + }, + i => { + action => sub { toggle_config('status_inc') }, + label => 'Toggle incremental status display', + }, + o => { + action => sub { send_cmd_to_servers('STOP SLAVE', 0, '', []); }, + label => 'Stop slave(s)', + }, + b => { + action => sub { purge_master_logs() }, + label => 'Purge unused master logs', + }, + }, + display_sub => \&display_M, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(slave_sql_status slave_io_status master_status)], + visible_tables => [qw(slave_sql_status slave_io_status master_status)], + }, + O => { + hdr => 'Open Tables', + cust => {}, + note => 'Shows open tables in MySQL', + action_for => { + r => { + action => sub { reverse_sort('open_tables'); }, + label => 'Reverse sort order', + }, + s => { + action => sub { choose_sort_cols('open_tables'); }, + label => "Choose sort column", + }, + }, + display_sub => \&display_O, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(open_tables)], + visible_tables => [qw(open_tables)], + }, + U => { + hdr => 'User Statistics', + cust => {}, + note => 'Displays Percona/MariaDB enhancements such as table statistics', + action_for => { + i => { + action => sub { set_visible_table('index_statistics') }, + label => 'Switch to INDEX_STATISTICS', + }, + s => { + action => sub { $clear_screen_sub->(); send_cmd_to_servers('SET @@global.userstat_running := 1 - @@global.userstat_running', 1, undef, []); }, + label => "Change the display's sort column", + }, + t => { + action => sub { set_visible_table('table_statistics') }, + label => 'Switch to TABLE_STATISTICS', + }, + x => { + action => sub { set_visible_table('index_table_statistics') }, + label => 'Switch to {INDEX,TABLE}_STATISTICS', + }, + }, + display_sub => \&display_P, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(table_statistics index_statistics index_table_statistics)], + visible_tables => [qw(index_table_statistics)], + }, + Q => { + hdr => 'Query List', + cust => {}, + note => 'Shows queries from SHOW FULL PROCESSLIST', + action_for => { + a => { + action => sub { toggle_filter('processlist', 'hide_self') }, + label => 'Toggle the innotop process', + }, + c => { + action => sub { edit_table('processlist') }, + label => 'Choose visible columns', + }, + e => { + action => sub { analyze_query('e'); }, + label => "Explain a thread's query", + }, + f => { + action => sub { analyze_query('f'); }, + label => "Show a thread's full query", + }, + h => { + action => sub { toggle_visible_table('Q', 'q_header') }, + label => 'Toggle the header on and off', + }, + i => { + action => sub { toggle_filter('processlist', 'hide_inactive') }, + label => 'Toggle idle processes', + }, + k => { + action => sub { kill_query('CONNECTION') }, + label => "Kill a query's connection", + }, + r => { + action => sub { reverse_sort('processlist'); }, + label => 'Reverse sort order', + }, + s => { + action => sub { choose_sort_cols('processlist'); }, + label => "Change the display's sort column", + }, + t => { + action => sub { toggle_filter('processlist', 'hide_connect') }, + label => 'Toggle slave processes', + }, + x => { + action => sub { kill_query('QUERY') }, + label => "Kill a query", + }, + }, + display_sub => \&display_Q, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(q_header processlist)], + visible_tables => [qw(q_header processlist)], + }, + R => { + hdr => 'InnoDB Row Ops', + cust => {}, + note => 'Shows InnoDB row operation and semaphore info', + action_for => { + i => { + action => sub { toggle_config('status_inc') }, + label => 'Toggle incremental status display', + }, + }, + display_sub => \&display_R, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(row_operations row_operation_misc semaphores wait_array)], + visible_tables => [qw(row_operations row_operation_misc semaphores wait_array)], + }, + S => { + hdr => 'Variables & Status', + cust => {}, + note => 'Shows query load statistics a la vmstat', + action_for => { + '>' => { + action => sub { switch_var_set('S_set', 1) }, + label => 'Switch to next variable set', + }, + '<' => { + action => sub { switch_var_set('S_set', -1) }, + label => 'Switch to prev variable set', + }, + c => { + action => sub { + choose_var_set('S_set'); + start_S_mode(); + }, + label => "Choose which set to display", + }, + e => { + action => \&edit_current_var_set, + label => 'Edit the current set of variables', + }, + i => { + action => sub { $clear_screen_sub->(); toggle_config('status_inc') }, + label => 'Toggle incremental status display', + }, + '-' => { + action => sub { set_display_precision(-1) }, + label => 'Decrease fractional display precision', + }, + '+' => { + action => sub { set_display_precision(1) }, + label => 'Increase fractional display precision', + }, + g => { + action => sub { set_s_mode('g') }, + label => 'Switch to graph (tload) view', + }, + s => { + action => sub { set_s_mode('s') }, + label => 'Switch to standard (vmstat) view', + }, + v => { + action => sub { set_s_mode('v') }, + label => 'Switch to pivoted view', + }, + }, + display_sub => \&display_S, + no_clear_screen => 1, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(var_status)], + visible_tables => [qw(var_status)], + }, + T => { + hdr => 'InnoDB Txns', + cust => {}, + note => 'Shows InnoDB transactions in top-like format', + action_for => { + a => { + action => sub { toggle_filter('innodb_transactions', 'hide_self') }, + label => 'Toggle the innotop process', + }, + c => { + action => sub { edit_table('innodb_transactions') }, + label => 'Choose visible columns', + }, + e => { + action => sub { analyze_query('e'); }, + label => "Explain a thread's query", + }, + f => { + action => sub { analyze_query('f'); }, + label => "Show a thread's full query", + }, + h => { + action => sub { toggle_visible_table('T', 't_header') }, + label => 'Toggle the header on and off', + }, + i => { + action => sub { toggle_filter('innodb_transactions', 'hide_inactive') }, + label => 'Toggle inactive transactions', + }, + k => { + action => sub { kill_query('CONNECTION') }, + label => "Kill a transaction's connection", + }, + r => { + action => sub { reverse_sort('innodb_transactions'); }, + label => 'Reverse sort order', + }, + s => { + action => sub { choose_sort_cols('innodb_transactions'); }, + label => "Change the display's sort column", + }, + x => { + action => sub { kill_query('QUERY') }, + label => "Kill a query", + }, + }, + display_sub => \&display_T, + connections => [], + server_group => '', + one_connection => 0, + tables => [qw(t_header innodb_transactions)], + visible_tables => [qw(t_header innodb_transactions)], + }, +); + +# ########################################################################### +# Global key mappings {{{3 +# Keyed on a single character, which is read from the keyboard. Uppercase +# letters switch modes. Lowercase letters access commands when in a mode. +# These can be overridden by action_for in %modes. +# ########################################################################### +my %action_for = ( + '$' => { + action => \&edit_configuration, + label => 'Edit configuration settings', + }, + '?' => { + action => \&display_help, + label => 'Show help', + }, + '!' => { + action => \&display_license, + label => 'Show license and warranty', + }, + '^' => { + action => \&edit_table, + label => "Edit the displayed table(s)", + }, + '#' => { + action => \&choose_server_groups, + label => 'Select/create server groups', + }, + '@' => { + action => \&choose_servers, + label => 'Select/create server connections', + }, + '/' => { + action => \&add_quick_filter, + label => 'Quickly filter what you see', + }, + '\\' => { + action => \&clear_quick_filters, + label => 'Clear quick-filters', + }, + '%' => { + action => \&choose_filters, + label => 'Choose and edit table filters', + }, + "\t" => { + action => \&next_server_group, + label => 'Switch to the next server group', + key => 'TAB', + }, + '=' => { + action => \&toggle_aggregate, + label => 'Toggle aggregation', + }, + # TODO: can these be auto-generated from %modes? + A => { + action => sub { switch_mode('A') }, + label => '', + }, + B => { + action => sub { switch_mode('B') }, + label => '', + }, + C => { + action => sub { switch_mode('C') }, + label => '', + }, + D => { + action => sub { switch_mode('D') }, + label => '', + }, + F => { + action => sub { switch_mode('F') }, + label => '', + }, + I => { + action => sub { switch_mode('I') }, + label => '', + }, + K => { + action => sub { switch_mode('K') }, + label => '', + }, + L => { + action => sub { switch_mode('L') }, + label => '', + }, + M => { + action => sub { switch_mode('M') }, + label => '', + }, + O => { + action => sub { switch_mode('O') }, + label => '', + }, + Q => { + action => sub { switch_mode('Q') }, + label => '', + }, + R => { + action => sub { switch_mode('R') }, + label => '', + }, + S => { + action => \&start_S_mode, + label => '', + }, + T => { + action => sub { switch_mode('T') }, + label => '', + }, + U => { + action => sub { switch_mode('U') }, + label => '', + }, + d => { + action => sub { get_config_interactive('interval') }, + label => 'Change refresh interval', + }, + n => { action => \&next_server, label => 'Switch to the next connection' }, + p => { action => \&pause, label => 'Pause innotop', }, + q => { action => \&finish, label => 'Quit innotop', }, +); + +# ########################################################################### +# Sleep times after certain statements {{{3 +# ########################################################################### +my %stmt_sleep_time_for = (); + +# ########################################################################### +# Config editor key mappings {{{3 +# ########################################################################### +my %cfg_editor_action = ( + c => { + note => 'Edit columns, etc in the displayed table(s)', + func => \&edit_table, + }, + g => { + note => 'Edit general configuration', + func => \&edit_configuration_variables, + }, + k => { + note => 'Edit row-coloring rules', + func => \&edit_color_rules, + }, + p => { + note => 'Manage plugins', + func => \&edit_plugins, + }, + s => { + note => 'Edit server groups', + func => \&edit_server_groups, + }, + S => { + note => 'Edit SQL statement sleep delays', + func => \&edit_stmt_sleep_times, + }, + t => { + note => 'Choose which table(s) to display in this mode', + func => \&choose_mode_tables, + }, +); + +# ########################################################################### +# Color editor key mappings {{{3 +# ########################################################################### +my %color_editor_action = ( + n => { + note => 'Create a new color rule', + func => sub { + my ( $tbl, $idx ) = @_; + my $meta = $tbl_meta{$tbl}; + + $clear_screen_sub->(); + my $col; + do { + $col = prompt_list( + 'Choose the target column for the rule', + '', + sub { return keys %{$meta->{cols}} }, + { map { $_ => $meta->{cols}->{$_}->{label} } keys %{$meta->{cols}} }); + } while ( !$col ); + ( $col ) = grep { $_ } split(/\W+/, $col); + return $idx unless $col && exists $meta->{cols}->{$col}; + + $clear_screen_sub->(); + my $op; + do { + $op = prompt_list( + 'Choose the comparison operator for the rule', + '', + sub { return keys %comp_ops }, + { map { $_ => $comp_ops{$_} } keys %comp_ops } ); + } until ( $op ); + $op =~ s/\s+//g; + return $idx unless $op && exists $comp_ops{$op}; + + my $arg; + do { + $arg = prompt('Specify an argument for the comparison'); + } until defined $arg; + + my $color; + do { + $color = prompt_list( + 'Choose the color(s) the row should be when the rule matches', + '', + sub { return keys %ansicolors }, + { map { $_ => $_ } keys %ansicolors } ); + } until defined $color; + $color = join(' ', unique(grep { exists $ansicolors{$_} } split(/\W+/, $color))); + return $idx unless $color; + + push @{$tbl_meta{$tbl}->{colors}}, { + col => $col, + op => $op, + arg => $arg, + color => $color + }; + $tbl_meta{$tbl}->{cust}->{colors} = 1; + + return $idx; + }, + }, + d => { + note => 'Remove the selected rule', + func => sub { + my ( $tbl, $idx ) = @_; + my @rules = @{ $tbl_meta{$tbl}->{colors} }; + return 0 unless @rules > 0 && $idx < @rules && $idx >= 0; + splice(@{$tbl_meta{$tbl}->{colors}}, $idx, 1); + $tbl_meta{$tbl}->{cust}->{colors} = 1; + return $idx == @rules ? $#rules : $idx; + }, + }, + j => { + note => 'Move highlight down one', + func => sub { + my ( $tbl, $idx ) = @_; + my $num_rules = scalar @{$tbl_meta{$tbl}->{colors}}; + return ($idx + 1) % $num_rules; + }, + }, + k => { + note => 'Move highlight up one', + func => sub { + my ( $tbl, $idx ) = @_; + my $num_rules = scalar @{$tbl_meta{$tbl}->{colors}}; + return ($idx - 1) % $num_rules; + }, + }, + '+' => { + note => 'Move selected rule up one', + func => sub { + my ( $tbl, $idx ) = @_; + my $meta = $tbl_meta{$tbl}; + my $dest = $idx == 0 ? scalar(@{$meta->{colors}} - 1) : $idx - 1; + my $temp = $meta->{colors}->[$idx]; + $meta->{colors}->[$idx] = $meta->{colors}->[$dest]; + $meta->{colors}->[$dest] = $temp; + $meta->{cust}->{colors} = 1; + return $dest; + }, + }, + '-' => { + note => 'Move selected rule down one', + func => sub { + my ( $tbl, $idx ) = @_; + my $meta = $tbl_meta{$tbl}; + my $dest = $idx == scalar(@{$meta->{colors}} - 1) ? 0 : $idx + 1; + my $temp = $meta->{colors}->[$idx]; + $meta->{colors}->[$idx] = $meta->{colors}->[$dest]; + $meta->{colors}->[$dest] = $temp; + $meta->{cust}->{colors} = 1; + return $dest; + }, + }, +); + +# ########################################################################### +# Plugin editor key mappings {{{3 +# ########################################################################### +my %plugin_editor_action = ( + '*' => { + note => 'Toggle selected plugin active/inactive', + func => sub { + my ( $plugins, $idx ) = @_; + my $plugin = $plugins->[$idx]; + $plugin->{active} = $plugin->{active} ? 0 : 1; + return $idx; + }, + }, + j => { + note => 'Move highlight down one', + func => sub { + my ( $plugins, $idx ) = @_; + return ($idx + 1) % scalar(@$plugins); + }, + }, + k => { + note => 'Move highlight up one', + func => sub { + my ( $plugins, $idx ) = @_; + return $idx == 0 ? @$plugins - 1 : $idx - 1; + }, + }, +); + +# ########################################################################### +# Table editor key mappings {{{3 +# ########################################################################### +my %tbl_editor_action = ( + a => { + note => 'Add a column to the table', + func => sub { + my ( $tbl, $col ) = @_; + my @visible_cols = @{ $tbl_meta{$tbl}->{visible} }; + my %all_cols = %{ $tbl_meta{$tbl}->{cols} }; + delete @all_cols{@visible_cols}; + my $choice = prompt_list( + 'Choose a column', + '', + sub { return keys %all_cols; }, + { map { $_ => $all_cols{$_}->{label} || $all_cols{$_}->{hdr} } keys %all_cols }); + if ( $all_cols{$choice} ) { + push @{$tbl_meta{$tbl}->{visible}}, $choice; + $tbl_meta{$tbl}->{cust}->{visible} = 1; + return $choice; + } + return $col; + }, + }, + n => { + note => 'Create a new column and add it to the table', + func => sub { + my ( $tbl, $col ) = @_; + + $clear_screen_sub->(); + print word_wrap("Choose a name for the column. This name is not displayed, and is used only " + . "for internal reference. It can contain only lowercase letters, numbers, " + . "and underscores."); + print "\n\n"; + do { + $col = prompt("Enter column name"); + $col = '' if $col =~ m/[^a-z0-9_]/; + } while ( !$col ); + + $clear_screen_sub->(); + my $hdr; + do { + $hdr = prompt("Enter column header"); + } while ( !$hdr ); + + $clear_screen_sub->(); + print "Choose a source for the column's data\n\n"; + my ( $src, $sub, $err ); + do { + if ( $err ) { + print "Error: $err\n\n"; + } + $src = prompt("Enter column source"); + if ( $src ) { + ( $sub, $err ) = compile_expr($src); + } + } until ( !$err); + + # TODO: this duplicates %col_props. + $tbl_meta{$tbl}->{cols}->{$col} = { + hdr => $hdr, + src => $src, + just => '-', + num => 0, + label => 'User-defined', + user => 1, + tbl => $tbl, + minw => 0, + maxw => 0, + trans => [], + func => $sub, + dec => 0, + agg => 0, + aggonly => 0, + agghide => 0, + }; + + $tbl_meta{$tbl}->{visible} = [ unique(@{$tbl_meta{$tbl}->{visible}}, $col) ]; + $tbl_meta{$tbl}->{cust}->{visible} = 1; + return $col; + }, + }, + d => { + note => 'Remove selected column', + func => sub { + my ( $tbl, $col ) = @_; + my @visible_cols = @{ $tbl_meta{$tbl}->{visible} }; + my $idx = 0; + return $col unless @visible_cols > 1; + while ( $visible_cols[$idx] ne $col ) { + $idx++; + } + $tbl_meta{$tbl}->{visible} = [ grep { $_ ne $col } @visible_cols ]; + $tbl_meta{$tbl}->{cust}->{visible} = 1; + return $idx == $#visible_cols ? $visible_cols[$idx - 1] : $visible_cols[$idx + 1]; + }, + }, + e => { + note => 'Edit selected column', + func => sub { + # TODO: make this editor hotkey-driven and give readline support. + my ( $tbl, $col ) = @_; + $clear_screen_sub->(); + my $meta = $tbl_meta{$tbl}->{cols}->{$col}; + my @prop = qw(hdr label src just num minw maxw trans agg); # TODO redundant + + my $answer; + do { + # Do what the user asked... + if ( $answer && grep { $_ eq $answer } @prop ) { + # Some properties are arrays, others scalars. + my $ini = ref $col_props{$answer} ? join(' ', @{$meta->{$answer}}) : $meta->{$answer}; + my $val = prompt("New value for $answer", undef, $ini); + $val = [ split(' ', $val) ] if ref($col_props{$answer}); + if ( $answer eq 'trans' ) { + $val = [ unique(grep{ exists $trans_funcs{$_} } @$val) ]; + } + @{$meta}{$answer, 'user', 'tbl' } = ( $val, 1, $tbl ); + } + + my @display_lines = ( + '', + "You are editing column $tbl.$col.\n", + ); + + push @display_lines, create_table2( + \@prop, + { map { $_ => $_ } @prop }, + { map { $_ => ref $meta->{$_} eq 'ARRAY' ? join(' ', @{$meta->{$_}}) + : ref $meta->{$_} ? '[expression code]' + : $meta->{$_} + } @prop + }, + { sep => ' ' }); + draw_screen(\@display_lines, { raw => 1 }); + print "\n\n"; # One to add space, one to clear readline artifacts + $answer = prompt('Edit what? (q to quit)'); + } while ( $answer ne 'q' ); + + return $col; + }, + }, + j => { + note => 'Move highlight down one', + func => sub { + my ( $tbl, $col ) = @_; + my @visible_cols = @{ $tbl_meta{$tbl}->{visible} }; + my $idx = 0; + while ( $visible_cols[$idx] ne $col ) { + $idx++; + } + return $visible_cols[ ($idx + 1) % @visible_cols ]; + }, + }, + k => { + note => 'Move highlight up one', + func => sub { + my ( $tbl, $col ) = @_; + my @visible_cols = @{ $tbl_meta{$tbl}->{visible} }; + my $idx = 0; + while ( $visible_cols[$idx] ne $col ) { + $idx++; + } + return $visible_cols[ $idx - 1 ]; + }, + }, + '+' => { + note => 'Move selected column up one', + func => sub { + my ( $tbl, $col ) = @_; + my $meta = $tbl_meta{$tbl}; + my @visible_cols = @{$meta->{visible}}; + my $idx = 0; + while ( $visible_cols[$idx] ne $col ) { + $idx++; + } + if ( $idx ) { + $visible_cols[$idx] = $visible_cols[$idx - 1]; + $visible_cols[$idx - 1] = $col; + $meta->{visible} = \@visible_cols; + } + else { + shift @{$meta->{visible}}; + push @{$meta->{visible}}, $col; + } + $meta->{cust}->{visible} = 1; + return $col; + }, + }, + '-' => { + note => 'Move selected column down one', + func => sub { + my ( $tbl, $col ) = @_; + my $meta = $tbl_meta{$tbl}; + my @visible_cols = @{$meta->{visible}}; + my $idx = 0; + while ( $visible_cols[$idx] ne $col ) { + $idx++; + } + if ( $idx == $#visible_cols ) { + unshift @{$meta->{visible}}, $col; + pop @{$meta->{visible}}; + } + else { + $visible_cols[$idx] = $visible_cols[$idx + 1]; + $visible_cols[$idx + 1] = $col; + $meta->{visible} = \@visible_cols; + } + $meta->{cust}->{visible} = 1; + return $col; + }, + }, + f => { + note => 'Choose filters', + func => sub { + my ( $tbl, $col ) = @_; + choose_filters($tbl); + return $col; + }, + }, + o => { + note => 'Edit color rules', + func => sub { + my ( $tbl, $col ) = @_; + edit_color_rules($tbl); + return $col; + }, + }, + s => { + note => 'Choose sort columns', + func => sub { + my ( $tbl, $col ) = @_; + choose_sort_cols($tbl); + return $col; + }, + }, + g => { + note => 'Choose group-by (aggregate) columns', + func => sub { + my ( $tbl, $col ) = @_; + choose_group_cols($tbl); + return $col; + }, + }, +); + +# ########################################################################### +# Global variables and environment {{{2 +# ########################################################################### + +my @this_term_size; # w_chars, h_chars, w_pix, h_pix +my @last_term_size; # w_chars, h_chars, w_pix, h_pix +my $char; +my $windows = $OSNAME =~ m/MSWin/; +my $have_color = 0; +my $MAX_ULONG = 4294967295; # 2^32-1 +my $num_regex = qr/^[+-]?(?=\d|\.)\d*(?:\.\d+)?(?:E[+-]?\d+|)$/i; +my $int_regex = qr/^\d+$/; +my $bool_regex = qr/^[01]$/; +my $term = undef; +my $file = undef; # File to watch for InnoDB monitor output +my $file_mtime = undef; # Status of watched file +my $file_data = undef; # Last chunk of text read from file +my $innodb_parser = InnoDBParser->new; + +my $nonfatal_errs = join('|', + 'Access denied for user', + 'Unknown MariaDB server host', + 'Unknown database', + 'Can\'t connect to local MariaDB server through socket', + 'Can\'t connect to MariaDB server on', + 'MySQL server has gone away', + 'Cannot call SHOW INNODB STATUS', + 'Access denied', + 'AutoCommit', + 'Lost connection to MariaDB server', + 'Lost connection to server', + 'Too many connections', +); + +if ( !$opts{n} ) { + require Term::ReadLine; + $term = Term::ReadLine->new('innotop'); +} + +# Stores status, variables, innodb status, master/slave status etc. +# Keyed on connection name. Each entry is a hashref of current and past data sets, +# keyed on clock tick. +my %vars; +my %info_gotten = (); # Which things have been retrieved for the current clock tick. +my %show_variables; # Stores SHOW VARIABLES for each cxn so we don't re-fetch. + +# Stores info on currently displayed queries: cxn, connection ID, query text, +# user, and host. +my @current_queries; + +my $lines_printed = 0; +my $clock = 0; # Incremented with every wake-sleep cycle +my $clearing_deadlocks = 0; + +# If terminal coloring is available, use it. The only function I want from +# the module is the colored() function. +eval { + if ( !$opts{n} ) { + if ( $windows ) { + require Win32::Console::ANSI; + } + require Term::ANSIColor; + import Term::ANSIColor qw(colored); + $have_color = 1; + } +}; +if ( $EVAL_ERROR || $opts{n} ) { + # If there was an error, manufacture my own colored() function that does no + # coloring. + *colored = sub { pop @_; @_; }; +} + +if ( $opts{n} ) { + $clear_screen_sub = sub {}; +} +elsif ( $windows ) { + $clear_screen_sub = sub { $lines_printed = 0; system("cls") }; +} +else { + my $clear = `clear`; + $clear_screen_sub = sub { $lines_printed = 0; print $clear }; +} + +# ########################################################################### +# Config storage. {{{2 +# ########################################################################### +my %config = ( + color => { + val => $have_color, + note => 'Whether to use terminal coloring', + conf => 'ALL', + pat => $bool_regex, + }, + cmd_filter => { + val => 'Com_', + note => 'Prefix for values in C mode', + conf => [qw(C)], + }, + plugin_dir => { + val => "$homepath/.innotop/plugins", + note => 'Directory where plugins can be found', + conf => 'ALL', + }, + show_percent => { + val => 1, + note => 'Show the % symbol after percentages', + conf => 'ALL', + pat => $bool_regex, + }, + skip_innodb => { + val => 0, + note => 'Disable SHOW INNODB STATUS', + conf => 'ALL', + pat => $bool_regex, + }, + S_func => { + val => 's', + note => 'What to display in S mode: graph, status, pivoted status', + conf => [qw(S)], + pat => qr/^[gsv]$/, + }, + cxn_timeout => { + val => 28800, + note => 'Connection timeout for keeping unused connections alive', + conf => 'ALL', + pat => $int_regex, + }, + graph_char => { + val => '*', + note => 'Character for drawing graphs', + conf => [ qw(S) ], + pat => qr/^.$/, + }, + show_cxn_errors_in_tbl => { + val => 1, + note => 'Whether to display connection errors as rows in the table', + conf => 'ALL', + pat => $bool_regex, + }, + hide_hdr => { + val => 0, + note => 'Whether to show column headers', + conf => 'ALL', + pat => $bool_regex, + }, + show_cxn_errors => { + val => 1, + note => 'Whether to print connection errors to STDOUT', + conf => 'ALL', + pat => $bool_regex, + }, + readonly => { + val => 1, + note => 'Whether the config file is read-only', + conf => [ qw() ], + pat => $bool_regex, + }, + global => { + val => 1, + note => 'Whether to show GLOBAL variables and status', + conf => 'ALL', + pat => $bool_regex, + }, + header_highlight => { + val => 'bold', + note => 'How to highlight table column headers', + conf => 'ALL', + pat => qr/^(?:bold|underline)$/, + }, + display_table_captions => { + val => 1, + note => 'Whether to put captions on tables', + conf => 'ALL', + pat => $bool_regex, + }, + charset => { + val => 'ascii', + note => 'What type of characters should be displayed in queries (ascii, unicode, none)', + conf => 'ALL', + pat => qr/^(?:ascii|unicode|none)$/, + }, + auto_wipe_dl => { + val => 0, + note => 'Whether to auto-wipe InnoDB deadlocks', + conf => 'ALL', + pat => $bool_regex, + }, + max_height => { + val => 30, + note => '[Win32] Max window height', + conf => 'ALL', + }, + debug => { + val => 0, + pat => $bool_regex, + note => 'Debug mode (more verbose errors, uses more memory)', + conf => 'ALL', + }, + num_digits => { + val => 2, + pat => $int_regex, + note => 'How many digits to show in fractional numbers and percents', + conf => 'ALL', + }, + debugfile => { + val => "$homepath/.innotop/core_dump", + note => 'A debug file in case you are interested in error output', + }, + show_statusbar => { + val => 1, + pat => $bool_regex, + note => 'Whether to show the status bar in the display', + conf => 'ALL', + }, + mode => { + val => "A", + note => "Which mode to start in", + cmdline => 1, + }, + status_inc => { + val => 0, + note => 'Whether to show raw or incremental values for status variables', + pat => $bool_regex, + }, + interval => { + val => 10, + pat => qr/^(?:(?:\d*?[1-9]\d*(?:\.\d*)?)|(?:\d*\.\d*?[1-9]\d*))$/, + note => "The interval at which the display will be refreshed. Fractional values allowed.", + }, + num_status_sets => { + val => 9, + pat => $int_regex, + note => 'How many sets of STATUS and VARIABLES values to show', + conf => [ qw(S) ], + }, + S_set => { + val => 'general', + pat => qr/^\w+$/, + note => 'Which set of variables to display in S (Variables & Status) mode', + conf => [ qw(S) ], + }, + timeformat => { + val => '%Y-%m-%dT%H:%M:%S', + pat => qr//, + note => 'The strftime() timestamp format to write in -n mode', + }, + spark => { + val => 10, + note => 'How long to make status variable sparklines', + conf => 'ALL', + pat => $int_regex, + }, +); + +# ########################################################################### +# Config file sections {{{2 +# The configuration file is broken up into sections like a .ini file. This +# variable defines those sections and the subroutines responsible for reading +# and writing them. +# ########################################################################### +my %config_file_sections = ( + plugins => { + reader => \&load_config_plugins, + writer => \&save_config_plugins, + }, + group_by => { + reader => \&load_config_group_by, + writer => \&save_config_group_by, + }, + filters => { + reader => \&load_config_filters, + writer => \&save_config_filters, + }, + active_filters => { + reader => \&load_config_active_filters, + writer => \&save_config_active_filters, + }, + visible_tables => { + reader => \&load_config_visible_tables, + writer => \&save_config_visible_tables, + }, + sort_cols => { + reader => \&load_config_sort_cols, + writer => \&save_config_sort_cols, + }, + active_columns => { + reader => \&load_config_active_columns, + writer => \&save_config_active_columns, + }, + tbl_meta => { + reader => \&load_config_tbl_meta, + writer => \&save_config_tbl_meta, + }, + general => { + reader => \&load_config_config, + writer => \&save_config_config, + }, + connections => { + reader => \&load_config_connections, + writer => \&save_config_connections, + }, + active_connections => { + reader => \&load_config_active_connections, + writer => \&save_config_active_connections, + }, + server_groups => { + reader => \&load_config_server_groups, + writer => \&save_config_server_groups, + }, + active_server_groups => { + reader => \&load_config_active_server_groups, + writer => \&save_config_active_server_groups, + }, + max_values_seen => { + reader => \&load_config_mvs, + writer => \&save_config_mvs, + }, + varsets => { + reader => \&load_config_varsets, + writer => \&save_config_varsets, + }, + colors => { + reader => \&load_config_colors, + writer => \&save_config_colors, + }, + stmt_sleep_times => { + reader => \&load_config_stmt_sleep_times, + writer => \&save_config_stmt_sleep_times, + }, +); + +# Config file sections have some dependencies, so they have to be read/written in order. +my @ordered_config_file_sections = qw(general plugins filters active_filters tbl_meta + connections active_connections server_groups active_server_groups max_values_seen + active_columns sort_cols visible_tables varsets colors stmt_sleep_times + group_by); + +# All events for which plugins may register themselves. Entries are arrayrefs. +my %event_listener_for = map { $_ => [] } + qw( + extract_values + set_to_tbl_pre_filter set_to_tbl_pre_sort set_to_tbl_pre_group + set_to_tbl_pre_colorize set_to_tbl_pre_transform set_to_tbl_pre_pivot + set_to_tbl_pre_create set_to_tbl_post_create + draw_screen + ); + +# All variables to which plugins have access. +my %pluggable_vars = ( + action_for => \%action_for, + agg_funcs => \%agg_funcs, + config => \%config, + connections => \%connections, + dbhs => \%dbhs, + filters => \%filters, + modes => \%modes, + server_groups => \%server_groups, + tbl_meta => \%tbl_meta, + trans_funcs => \%trans_funcs, + var_sets => \%var_sets, +); + +# ########################################################################### +# Contains logic to generate prepared statements for a given function for a +# given DB connection. Returns a $sth. +# ########################################################################### +my %stmt_maker_for = ( + INDEX_STATISTICS => sub { + my ( $dbh ) = @_; + # Detect whether there's a Percona Server with INFORMATION_SCHEMA.INDEX_STATISTICS + # and if not, just select nothing. + my $sth; + eval { # This can fail if the table doesn't exist, INFORMATION_SCHEMA doesn't exist, etc. + my $cols = $dbh->selectall_arrayref(q{SHOW /*innotop*/ COLUMNS FROM INFORMATION_SCHEMA.INDEX_STATISTICS}); + if ( @$cols ) { + $sth = $dbh->prepare(q{SELECT /*innotop*/ * FROM INFORMATION_SCHEMA.INDEX_STATISTICS}); + } + }; + $sth ||= $dbh->prepare(q{SELECT /*innotop*/ '' FROM DUAL WHERE 1 = 0}); + return $sth; + }, + INDEX_TABLE_STATISTICS => sub { + my ( $dbh ) = @_; + # Detect whether there's a Percona Server with INFORMATION_SCHEMA.INDEX_STATISTICS + # and if not, just select nothing. + my $sth; + eval { # This can fail if the table doesn't exist, INFORMATION_SCHEMA doesn't exist, etc. + my $cols = $dbh->selectall_arrayref(q{SHOW /*innotop*/ COLUMNS FROM INFORMATION_SCHEMA.INDEX_STATISTICS}); + if ( @$cols ) { + $sth = $dbh->prepare(q{SELECT /*innotop*/ L.TABLE_SCHEMA, L.TABLE_NAME, } + . q{SUM(L.ROWS_READ) AS ROWS_READ, SUM(R.ROWS_READ) AS ROWS_READ_FROM_INDEXES, } + . q{SUM(L.ROWS_CHANGED) AS ROWS_CHANGED, } + . q{SUM(L.ROWS_CHANGED_X_INDEXES) AS ROWS_CHANGED_X_INDEXES } + . q{FROM INFORMATION_SCHEMA.TABLE_STATISTICS AS L LEFT JOIN INFORMATION_SCHEMA.INDEX_STATISTICS AS R } + . q{USING(TABLE_SCHEMA, TABLE_NAME) GROUP BY L.TABLE_SCHEMA, L.TABLE_NAME}); + } + }; + $sth ||= $dbh->prepare(q{SELECT /*innotop*/ '' FROM DUAL WHERE 1 = 0}); + return $sth; + }, + INNODB_BLOCKED_BLOCKER => sub { + my ( $dbh ) = @_; + # Detect whether the server supports the I_S tables and if not, just select nothing. + my $sth; + eval { # This can fail if the table doesn't exist, INFORMATION_SCHEMA doesn't exist, etc. + my $cols = $dbh->selectall_arrayref(q{SHOW /*innotop*/ COLUMNS FROM INFORMATION_SCHEMA.INNODB_LOCK_WAITS}); + if ( @$cols ) { + if ($dbh->{mariadb_serverinfo} =~ /^5.1/) { + $sth = $dbh->prepare(q{ + SELECT /*innotop*/ + r.trx_mysql_thread_id AS waiting_thread, + r.trx_query AS waiting_query, + "n/a" AS waiting_rows_modified, + TIMESTAMPDIFF(SECOND, r.trx_started, NOW()) AS waiting_age, + TIMESTAMPDIFF(SECOND, r.trx_wait_started, NOW()) AS waiting_wait_secs, + rp.user AS waiting_user, + rp.host AS waiting_host, + rp.db AS waiting_db, + b.trx_mysql_thread_id AS blocking_thread, + b.trx_query AS blocking_query, + "n/a" AS blocking_rows_modified, + TIMESTAMPDIFF(SECOND, b.trx_started, NOW()) AS blocking_age, + TIMESTAMPDIFF(SECOND, b.trx_wait_started, NOW()) AS blocking_wait_secs, + bp.user AS blocking_user, + bp.host AS blocking_host, + bp.db AS blocking_db, + CONCAT(bp.command, IF(bp.command = 'Sleep', CONCAT(' ', bp.time), '')) AS blocking_status, + CONCAT(lock_mode, ' ', lock_type, ' ', lock_table, '(', lock_index, ')') AS lock_info + FROM INFORMATION_SCHEMA.INNODB_LOCK_WAITS w + JOIN INFORMATION_SCHEMA.INNODB_TRX b ON b.trx_id = w.blocking_trx_id + JOIN INFORMATION_SCHEMA.INNODB_TRX r ON r.trx_id = w.requesting_trx_id + JOIN INFORMATION_SCHEMA.INNODB_LOCKS l ON l.lock_id = w.requested_lock_id + LEFT JOIN INFORMATION_SCHEMA.PROCESSLIST bp ON bp.id = b.trx_mysql_thread_id + LEFT JOIN INFORMATION_SCHEMA.PROCESSLIST rp ON rp.id = r.trx_mysql_thread_id + }); + } else { + $sth = $dbh->prepare(q{ + SELECT /*innotop*/ + r.trx_mysql_thread_id AS waiting_thread, + r.trx_query AS waiting_query, + r.trx_rows_modified AS waiting_rows_modified, + TIMESTAMPDIFF(SECOND, r.trx_started, NOW()) AS waiting_age, + TIMESTAMPDIFF(SECOND, r.trx_wait_started, NOW()) AS waiting_wait_secs, + rp.user AS waiting_user, + rp.host AS waiting_host, + rp.db AS waiting_db, + b.trx_mysql_thread_id AS blocking_thread, + b.trx_query AS blocking_query, + b.trx_rows_modified AS blocking_rows_modified, + TIMESTAMPDIFF(SECOND, b.trx_started, NOW()) AS blocking_age, + TIMESTAMPDIFF(SECOND, b.trx_wait_started, NOW()) AS blocking_wait_secs, + bp.user AS blocking_user, + bp.host AS blocking_host, + bp.db AS blocking_db, + CONCAT(bp.command, IF(bp.command = 'Sleep', CONCAT(' ', bp.time), '')) AS blocking_status, + CONCAT(lock_mode, ' ', lock_type, ' ', lock_table, '(', lock_index, ')') AS lock_info + FROM INFORMATION_SCHEMA.INNODB_LOCK_WAITS w + JOIN INFORMATION_SCHEMA.INNODB_TRX b ON b.trx_id = w.blocking_trx_id + JOIN INFORMATION_SCHEMA.INNODB_TRX r ON r.trx_id = w.requesting_trx_id + JOIN INFORMATION_SCHEMA.INNODB_LOCKS l ON l.lock_id = w.requested_lock_id + LEFT JOIN INFORMATION_SCHEMA.PROCESSLIST bp ON bp.id = b.trx_mysql_thread_id + LEFT JOIN INFORMATION_SCHEMA.PROCESSLIST rp ON rp.id = r.trx_mysql_thread_id + }); + } + } + }; + $sth ||= $dbh->prepare(q{SELECT /*innotop*/ '' FROM DUAL WHERE 1 = 0}); + return $sth; + }, + INNODB_STATUS => sub { + my ( $dbh ) = @_; + return $dbh->prepare(version_ge( $dbh, '5.0.0' ) + ? 'SHOW /*innotop*/ ENGINE INNODB STATUS' + : 'SHOW /*innotop*/ INNODB STATUS'); + }, + SHOW_VARIABLES => sub { + my ( $dbh ) = @_; + return $dbh->prepare($config{global}->{val} && version_ge( $dbh, '4.0.3' ) + ? 'SHOW /*innotop*/ GLOBAL VARIABLES' + : 'SHOW /*innotop*/ VARIABLES'); + }, + SHOW_STATUS => sub { + my ( $dbh ) = @_; + return $dbh->prepare($config{global}->{val} && version_ge( $dbh, '5.0.2' ) + ? 'SHOW /*innotop*/ GLOBAL STATUS' + : 'SHOW /*innotop*/ STATUS'); + }, + KILL_QUERY => sub { + my ( $dbh ) = @_; + return $dbh->prepare(version_ge( $dbh, '5.0.0' ) + ? 'KILL /*innotop*/ QUERY ?' + : 'KILL /*innotop*/ ?'); + }, + SHOW_MASTER_LOGS => sub { + my ( $dbh ) = @_; + return $dbh->prepare('SHOW /*innotop*/ MASTER LOGS'); + }, + SHOW_MASTER_STATUS => sub { + my ( $dbh ) = @_; + return $dbh->prepare('SHOW /*innotop*/ MASTER STATUS'); + }, + SHOW_SLAVE_STATUS => sub { + my ( $dbh ) = @_; + return $dbh->prepare('SHOW /*innotop*/ SLAVE STATUS'); + }, + GET_CHANNELS => sub { + my ( $dbh ) = @_; + return $dbh->prepare(version_ge( $dbh, '5.7.0' ) + ? 'select CHANNEL_NAME from performance_schema.replication_applier_status where CHANNEL_NAME regexp "^[a-zA-Z].*";' + : 'select "no_channels"'); + }, + KILL_CONNECTION => sub { + my ( $dbh ) = @_; + return $dbh->prepare(version_ge( $dbh, '5.0.0' ) + ? 'KILL /*innotop*/ CONNECTION ?' + : 'KILL /*innotop*/ ?'); + }, + OPEN_TABLES => sub { + my ( $dbh ) = @_; + return version_ge($dbh, '4.0.0') + ? $dbh->prepare('SHOW /*innotop*/ OPEN TABLES') + : undef; + }, + PROCESSLIST => sub { + my ( $dbh ) = @_; + # In newer versions of the server, use INFORMATION_SCHEMA table if it exists, + # and use the TIME_MS column (in Percona Server) if that exists. + my $sth; + eval { # This can fail if the table doesn't exist, INFORMATION_SCHEMA doesn't exist, etc. + my $cols = $dbh->selectall_arrayref(q{SHOW /*innotop*/ COLUMNS FROM INFORMATION_SCHEMA.PROCESSLIST LIKE 'TIME_MS'}); + if ( @$cols ) { # The TIME_MS column exists + $sth = $dbh->prepare(q{SELECT /*innotop*/ ID, USER, HOST, DB, COMMAND, CASE WHEN TIME_MS/1000 > 365*86400 THEN TIME ELSE TIME_MS/1000 END AS TIME, STATE, INFO FROM INFORMATION_SCHEMA.PROCESSLIST}); + } + }; + $sth ||= $dbh->prepare('SHOW /*innotop*/ FULL PROCESSLIST'); + return $sth; + }, + PROCESSLIST_NO_IS => sub { + my ( $dbh ) = @_; + # We do not use INFORMATION_SCHEMA table because it doesn't show slave + # SQL statements. http://bugs.mysql.com/66401 + my $sth = $dbh->prepare('SHOW /*innotop*/ FULL PROCESSLIST'); + return $sth; + }, + TABLE_STATISTICS => sub { + my ( $dbh ) = @_; + # Detect whether there's a Percona Server with INFORMATION_SCHEMA.TABLE_STATISTICS + # and if not, just select nothing. + my $sth; + eval { # This can fail if the table doesn't exist, INFORMATION_SCHEMA doesn't exist, etc. + my $cols = $dbh->selectall_arrayref(q{SHOW /*innotop*/ COLUMNS FROM INFORMATION_SCHEMA.TABLE_STATISTICS}); + if ( @$cols ) { + $sth = $dbh->prepare(q{SELECT /*innotop*/ * FROM INFORMATION_SCHEMA.TABLE_STATISTICS}); + } + }; + $sth ||= $dbh->prepare(q{SELECT /*innotop*/ '' FROM DUAL WHERE 1 = 0}); + return $sth; + }, +); + +# Plugins! +my %plugins = ( +); + +# ########################################################################### +# Run the program {{{1 +# ########################################################################### +sub main { + # This config variable is only useful for MS Windows because its terminal + # can't tell how tall it is. + if ( !$windows ) { + delete $config{max_height}; + } + + # Try to lower my priority. + eval { setpriority(0, 0, getpriority(0, 0) + 10); }; + + # Print stuff to the screen immediately, don't wait for a newline. + $OUTPUT_AUTOFLUSH = 1; + + # Clear the screen and load the configuration. + $clear_screen_sub->(); + load_config(); + + # Override config variables with command-line options + my %cmdline = + map { $_->{c} => $opts{$_->{k}} } + grep { exists $_->{c} && exists $opts{$_->{k}} } + @opt_spec; + + foreach my $name (keys %cmdline) { + next if not defined $cmdline{$name}; + my $val = $cmdline{$name}; + if ( exists($config{$name}) and (!$config{$name}->{pat} or $val =~ m/$config{$name}->{pat}/ )) { + $config{$name}->{val} = $val; + } + } + + post_process_tbl_meta(); + + # Make sure no changes are written to config file in non-interactive mode. + if ( $opts{n} ) { + $config{readonly}->{val} = 1; + } + + eval { + + # Open the file for InnoDB status + if ( @ARGV ) { + my $filename = shift @ARGV; + open $file, "<", $filename + or die "Cannot open '$filename': $OS_ERROR"; + } + + # In certain modes we might have to collect data for two cycles + # before printing anything out, so we need to bump up the count one. + if ( $opts{n} && $opts{count} && $config{status_inc}->{val} + && $config{mode}->{val} =~ m/[S]/ ) + { + $opts{count}++; + } + + while (++$clock) { + + my $mode = $config{mode}->{val} || 'Q'; + if ( !$modes{$mode} ) { + die "Mode '$mode' doesn't exist; try one of these:\n" + . join("\n", map { " $_ $modes{$_}->{hdr}" } sort keys %modes) + . "\n"; + } + + if ( !$opts{n} ) { + @last_term_size = @this_term_size; + @this_term_size = Term::ReadKey::GetTerminalSize(\*STDOUT); + if ( $windows ) { + $this_term_size[0]--; + $this_term_size[1] + = min($this_term_size[1], $config{max_height}->{val}); + } + die("Can't read terminal size") unless @this_term_size; + } + + # If there's no connection to a database server, we need to fix that... + if ( !%connections ) { + print "You have not defined any database connections.\n\n"; + add_new_dsn(); + } + + # See whether there are any connections defined for this mode. If there's only one + # connection total, assume the user wants to just use innotop for a single server + # and don't ask which server to connect to. Also, if we're monitoring from a file, + # we just use the first connection. + if ( !get_connections() ) { + if ( $file || 1 == scalar keys %connections ) { + $modes{$config{mode}->{val}}->{connections} = [ keys %connections ]; + } + else { + choose_connections(); + } + } + + # Term::ReadLine might have re-set $OUTPUT_AUTOFLUSH. + $OUTPUT_AUTOFLUSH = 1; + + # Prune old data + my $sets = $config{num_status_sets}->{val}; + foreach my $store ( values %vars ) { + delete @{$store}{ grep { $_ < $clock - $sets } keys %$store }; + } + %info_gotten = (); + + # Call the subroutine to display this mode. + $modes{$mode}->{display_sub}->(); + + # It may be time to quit now. + if ( $opts{count} && $clock >= $opts{count} ) { + finish(); + } + + # RECON: Try to reconnect failed connections, while the user sees no lag. + foreach my $cxn ( grep { $dbhs{$_}->{failed} } keys %dbhs ) { + eval { connect_to_db($cxn); }; # Ignore errors entirely here. + } + + # Wait for a bit. + if ( $opts{n} ) { + sleep($config{interval}->{val}); + } + else { + ReadMode('cbreak'); + $char = ReadKey($config{interval}->{val}); + ReadMode('normal'); + } + + # Handle whatever action the key indicates. + do_key_action(); + + } + }; + if ( $EVAL_ERROR ) { + core_dump( $EVAL_ERROR ); + } + finish(); +} +main() unless caller(); # make me testable! + +# Subroutines {{{1 +# Mode functions{{{2 +# switch_mode {{{3 +sub switch_mode { + my $mode = shift; + $config{mode}->{val} = $mode; +} + +# Prompting functions {{{2 +# prompt_list {{{3 +# Prompts the user for a value, given a question, initial value, +# a completion function and a hashref of hints. +sub prompt_list { + die "Can't call in non-interactive mode" if $opts{n}; + my ( $question, $init, $completion, $hints ) = @_; + if ( $hints ) { + # Figure out how wide the table will be + my $max_name = max(map { length($_) } keys %$hints ); + $max_name ||= 0; + $max_name += 3; + my @meta_rows = create_table2( + [ sort keys %$hints ], + { map { $_ => $_ } keys %$hints }, + { map { $_ => trunc($hints->{$_}, $this_term_size[0] - $max_name) } keys %$hints }, + { sep => ' ' }); + if (@meta_rows > 10) { + # Try to split and stack the meta rows next to each other + my $split = int(@meta_rows / 2); + @meta_rows = stack_next( + [@meta_rows[0..$split - 1]], + [@meta_rows[$split..$#meta_rows]], + { pad => ' | '}, + ); + } + print join( "\n", + '', + map { ref $_ ? colored(@$_) : $_ } create_caption('Choose from', @meta_rows), ''), + "\n"; + } + $term->Attribs->{completion_function} = $completion; + my $answer = $term->readline("$question: ", $init); + $OUTPUT_AUTOFLUSH = 1; + $answer = '' if !defined($answer); + $answer =~ s/\s+$//; + return $answer; +} + +# prompt {{{3 +# Prints out a prompt and reads from the keyboard, then validates with the +# validation regex until the input is correct. +sub prompt { + die "Can't call in non-interactive mode" if $opts{n}; + my ( $prompt, $regex, $init, $completion ) = @_; + my $response; + my $success = 0; + do { + if ( $completion ) { + $term->Attribs->{completion_function} = $completion; + } + $response = $term->readline("$prompt: ", $init); + if ( $regex && $response !~ m/$regex/ ) { + print "Invalid response.\n\n"; + } + else { + $success = 1; + } + } while ( !$success ); + $OUTPUT_AUTOFLUSH = 1; + $response =~ s/\s+$//; + return $response; +} + +# prompt_noecho {{{3 +# Unfortunately, suppressing echo with Term::ReadLine isn't reliable; the user might not +# have that library, or it might not support that feature. +sub prompt_noecho { + my ( $prompt ) = @_; + print colored("$prompt: ", 'underline'); + my $response; + ReadMode('noecho'); + $response = <STDIN>; + chomp($response); + ReadMode('normal'); + return $response; +} + +# noecho_password {{{3 +# read password for command line parameters with noecho +sub noecho_password { + my $prompt = shift @_; + local $OUTPUT_AUTOFLUSH = 1; + my $response; + eval { + if ( $windows ) { + require Win32::Console::ANSI; + } + require Term::ANSIColor; + import Term::ANSIColor qw(colored); + $response = prompt_noecho($prompt); + print "\n" or die + "Cannot print: $OS_ERROR"; + }; + + if ( $EVAL_ERROR ) { + die "Cannot read response; is Term::ReadKey installed? $EVAL_ERROR"; + } + return $response; +} + +# do_key_action {{{3 +# Depending on whether a key was read, do something. Keys have certain +# actions defined in lookup tables. Each mode may have its own lookup table, +# which trumps the global table -- so keys can be context-sensitive. The key +# may be read and written in a subroutine, so it's a global. +sub do_key_action { + if ( defined $char ) { + my $mode = $config{mode}->{val}; + my $action + = defined($modes{$mode}->{action_for}->{$char}) ? $modes{$mode}->{action_for}->{$char}->{action} + : defined($action_for{$char}) ? $action_for{$char}->{action} + : sub{}; + $action->(); + } +} + +# pause {{{3 +sub pause { + die "Can't call in non-interactive mode" if $opts{n}; + my $msg = shift; + print defined($msg) ? "\n$msg" : "\nPress any key to continue"; + ReadMode('cbreak'); + my $char = ReadKey(0); + ReadMode('normal'); + return $char; +} + +# reverse_sort {{{3 +sub reverse_sort { + my $tbl = shift; + $tbl_meta{$tbl}->{sort_dir} *= -1; +} + +# select_cxn {{{3 +# Selects connection(s). If the mode (or argument list) has only one, returns +# it without prompt. +sub select_cxn { + my ( $prompt, @cxns ) = @_; + if ( !@cxns ) { + @cxns = get_connections(); + } + if ( @cxns == 1 ) { + return $cxns[0]; + } + my $choices = prompt_list( + $prompt, + $cxns[0], + sub{ return @cxns }, + { map { $_ => $connections{$_}->{dsn} } @cxns }); + my @result = unique(grep { my $a = $_; grep { $_ eq $a } @cxns } split(/\s+/, $choices)); + return @result; +} + +# kill_query {{{3 +# Kills a connection, or on new versions, optionally a query but not connection. +sub kill_query { + my ( $q_or_c ) = @_; + + my $info = choose_thread( + sub { 1 }, + 'Select a thread to kill the ' . $q_or_c, + ); + return unless $info; + my $distill = distill($info->{query} || ''); + $distill = " running '$distill'" if $distill; + return unless pause("Kill $info->{id} (" + . ($info->{user} || '') + . '@' + . ($info->{host} || '') + . ")$distill ? ") =~ m/y/i; + + eval { + do_stmt($info->{cxn}, $q_or_c eq 'QUERY' ? 'KILL_QUERY' : 'KILL_CONNECTION', $info->{id} ); + }; + + if ( $EVAL_ERROR ) { + print "\nError: $EVAL_ERROR"; + pause(); + } +} + +# set_display_precision {{{3 +sub set_display_precision { + my $dir = shift; + $config{num_digits}->{val} = min(9, max(0, $config{num_digits}->{val} + $dir)); +} + +sub toggle_visible_table { + my ( $mode, $table ) = @_; + my $visible = $modes{$mode}->{visible_tables}; + if ( grep { $_ eq $table } @$visible ) { + $modes{$mode}->{visible_tables} = [ grep { $_ ne $table } @$visible ]; + } + else { + unshift @$visible, $table; + } + $modes{$mode}->{cust}->{visible_tables} = 1; +} + +# toggle_filter{{{3 +sub toggle_filter { + my ( $tbl, $filter ) = @_; + my $filters = $tbl_meta{$tbl}->{filters}; + if ( grep { $_ eq $filter } @$filters ) { + $tbl_meta{$tbl}->{filters} = [ grep { $_ ne $filter } @$filters ]; + } + else { + push @$filters, $filter; + } + $tbl_meta{$tbl}->{cust}->{filters} = 1; +} + +# toggle_config {{{3 +sub toggle_config { + my ( $key ) = @_; + $config{$key}->{val} ^= 1; +} + +# create_deadlock {{{3 +sub create_deadlock { + $clear_screen_sub->(); + + print "This function will deliberately cause a small deadlock, " + . "clearing deadlock information from the InnoDB monitor.\n\n"; + + my $answer = prompt("Are you sure you want to proceed? Say 'y' if you do"); + return 0 unless $answer eq 'y'; + + my ( $cxn ) = select_cxn('Clear on which server? '); + return unless $cxn && exists($connections{$cxn}); + + clear_deadlock($cxn); +} + +# deadlock_thread {{{3 +sub deadlock_thread { + my ( $id, $tbl, $cxn ) = @_; + + eval { + my $dbh = get_new_db_connection($cxn, 1); + + # disable binary logging for this session + $dbh->do("set SQL_LOG_BIN=0"); + + my @stmts = ( + "set transaction isolation level serializable", + (version_ge($dbh, '4.0.11') ? "start transaction" : 'begin'), + "select * from $tbl where a = $id", + "update $tbl set a = $id where a <> $id", + ); + + foreach my $stmt (@stmts[0..2]) { + $dbh->do($stmt); + } + sleep(1 + $id); + $dbh->do($stmts[-1]); + }; + if ( $EVAL_ERROR ) { + if ( $EVAL_ERROR !~ m/Deadlock found/ ) { + die $EVAL_ERROR; + } + } + exit(0); +} + +# Purges unused binlogs on the master, up to but not including the latest log. +# TODO: guess which connections are slaves of a given master. +sub purge_master_logs { + my @cxns = get_connections(); + + get_master_slave_status(@cxns); + + # Toss out the rows that don't have master/slave status... + my @vars = + grep { $_ && ($_->{file} || $_->{master_host}) } + map { $vars{$_}->{$clock} } @cxns; + @cxns = map { $_->{cxn} } @vars; + + # Figure out which master to purge ons. + my @masters = map { $_->{cxn} } grep { $_->{file} } @vars; + my ( $master ) = select_cxn('Which master?', @masters ); + return unless $master; + my ($master_status) = grep { $_->{cxn} eq $master } @vars; + + # Figure out the result order (not lexical order) of master logs. + my @master_logs = get_master_logs($master); + my $i = 0; + my %master_logs = map { $_->{log_name} => $i++ } @master_logs; + + # Ask which slave(s) are reading from this master. + my @slave_status = grep { $_->{master_host} } @vars; + my @slaves = map { $_->{cxn} } @slave_status; + @slaves = select_cxn("Which slaves are reading from $master?", @slaves); + @slave_status = grep { my $item = $_; grep { $item->{cxn} eq $_ } @slaves } @slave_status; + return unless @slave_status; + + # Find the minimum binary log in use. + my $min_log = min(map { $master_logs{$_->{master_log_file}} } @slave_status); + my $log_name = $master_logs[$min_log]->{log_name}; + + my $stmt = "PURGE MASTER LOGS TO '$log_name'"; + send_cmd_to_servers($stmt, 0, 'PURGE {MASTER | BINARY} LOGS {TO "log_name" | BEFORE "date"}', [$master]); +} + +sub send_cmd_to_servers { + my ( $cmd, $all, $hint, $cxns ) = @_; + if ( $all ) { + @$cxns = get_connections(); + } + elsif ( !@$cxns ) { + @$cxns = select_cxn('Which servers?', @$cxns); + } + if ( $hint ) { + print "\nHint: $hint\n"; + } + $cmd = prompt('Command to send', undef, $cmd); + foreach my $cxn ( @$cxns ) { + eval { + my $sth = do_query($cxn, $cmd); + }; + if ( $EVAL_ERROR ) { + print "Error from $cxn: $EVAL_ERROR\n"; + } + else { + print "Success on $cxn\n"; + } + } + pause(); +} + +# Display functions {{{2 + +sub set_s_mode { + my ( $func ) = @_; + $clear_screen_sub->(); + $config{S_func}->{val} = $func; +} + +# start_S_mode {{{3 +sub start_S_mode { + $clear_screen_sub->(); + switch_mode('S'); +} + +# display_A {{{3 +sub display_A { + my @display_lines; + my @cxns = get_connections(); + get_processlist_stats(@cxns); + get_status_info(@cxns); + get_master_slave_status(@cxns); + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + my @health_dashboard; + my %rows_for = ( + health_dashboard => \@health_dashboard, + ); + + foreach my $cxn ( @cxns ) { + # Get the status variables + my $set = $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock-1} || $set; + my $hash = extract_values($set, $set, $pre, 'health_dashboard'); + # Make QPS and Miss show now, not overall. + if ( exists $vars{$cxn}->{$clock - 1} ) { + my $inc = inc(0, $cxn); + my $hash2 = extract_values($inc, $set, $pre, 'health_dashboard'); + map { $hash->{$_} = $hash2->{$_} } qw(qps miss_rate); + } + push @health_dashboard, $hash; + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_B {{{3 +sub display_B { + my @display_lines; + my @cxns = get_connections(); + get_status_info(@cxns); + get_innodb_status(\@cxns); + + my @buffer_pool; + my @page_statistics; + my @insert_buffers; + my @adaptive_hash_index; + my %rows_for = ( + buffer_pool => \@buffer_pool, + page_statistics => \@page_statistics, + insert_buffers => \@insert_buffers, + adaptive_hash_index => \@adaptive_hash_index, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + foreach my $cxn ( @cxns ) { + my $set = $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock-1} || $set; + + if ( $set->{IB_bp_complete} ) { + if ( $wanted{buffer_pool} ) { + push @buffer_pool, extract_values($set, $set, $pre, 'buffer_pool'); + } + if ( $wanted{page_statistics} ) { + push @page_statistics, extract_values($set, $set, $pre, 'page_statistics'); + } + } + if ( $set->{IB_ib_complete} ) { + if ( $wanted{insert_buffers} ) { + push @insert_buffers, extract_values( + $config{status_inc}->{val} ? inc(0, $cxn) : $set, $set, $pre, + 'insert_buffers'); + } + if ( $wanted{adaptive_hash_index} ) { + push @adaptive_hash_index, extract_values($set, $set, $pre, 'adaptive_hash_index'); + } + } + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_C {{{3 +sub display_C { + my @display_lines; + my @cxns = get_connections(); + get_status_info(@cxns); + + my @cmd_summary; + my %rows_for = ( + cmd_summary => \@cmd_summary, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + # For now, I'm manually pulling these variables out and pivoting. Eventually a SQL-ish + # dialect should let me join a table to a grouped and pivoted table and do this more easily. + # TODO: make it so. + my $prefix = qr/^$config{cmd_filter}->{val}/; # TODO: this is a total hack + my @values; + my ($total, $last_total) = (0, 0); + foreach my $cxn ( @cxns ) { + my $set = $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock-1} || $set; + foreach my $key ( keys %$set ) { + next unless $key =~ m/$prefix/i; + my $val = $set->{$key}; + next unless defined $val && $val =~ m/^\d+$/; + my $last_val = $val - ($pre->{$key} || 0); + $total += $val; + $last_total += $last_val; + push @values, { + name => $key, + value => $val, + last_value => $last_val, + }; + } + } + + # Add aggregation and turn into a real set TODO: total hack + if ( $wanted{cmd_summary} ) { + foreach my $value ( @values ) { + @{$value}{qw(total last_total)} = ($total, $last_total); + push @cmd_summary, extract_values($value, $value, $value, 'cmd_summary'); + } + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_D {{{3 +sub display_D { + my @display_lines; + my @cxns = get_connections(); + get_status_info(@cxns); + get_innodb_status(\@cxns); + + my @deadlock_transactions; + my @deadlock_locks; + my %rows_for = ( + deadlock_transactions => \@deadlock_transactions, + deadlock_locks => \@deadlock_locks, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + foreach my $cxn ( @cxns ) { + my $innodb_status = $vars{$cxn}->{$clock}; + my $prev_status = $vars{$cxn}->{$clock-1} || $innodb_status; + + if ( $innodb_status->{IB_dl_timestring} ) { + + my $victim = $innodb_status->{IB_dl_rolled_back} || 0; + + if ( %wanted ) { + foreach my $txn_id ( keys %{$innodb_status->{IB_dl_txns}} ) { + my $txn = $innodb_status->{IB_dl_txns}->{$txn_id}; + my $pre = $prev_status->{IB_dl_txns}->{$txn_id} || $txn; + + if ( $wanted{deadlock_transactions} ) { + my $hash = extract_values($txn->{tx}, $txn->{tx}, $pre->{tx}, 'deadlock_transactions'); + $hash->{cxn} = $cxn; + $hash->{dl_txn_num} = $txn_id; + $hash->{victim} = $txn_id == $victim ? 'Yes' : 'No'; + $hash->{timestring} = $innodb_status->{IB_dl_timestring}; + $hash->{truncates} = $innodb_status->{IB_dl_complete} ? 'No' : 'Yes'; + push @deadlock_transactions, $hash; + } + + if ( $wanted{deadlock_locks} ) { + foreach my $lock ( @{$txn->{locks}} ) { + my $hash = extract_values($lock, $lock, $lock, 'deadlock_locks'); + $hash->{dl_txn_num} = $txn_id; + $hash->{cxn} = $cxn; + $hash->{mysql_thread_id} = $txn->{tx}->{mysql_thread_id}; + push @deadlock_locks, $hash; + } + } + + } + } + } + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_F {{{3 +sub display_F { + my @display_lines; + my ( $cxn ) = get_connections(); + get_status_info($cxn); + get_innodb_status([$cxn]); + my $innodb_status = $vars{$cxn}->{$clock}; + + if ( $innodb_status->{IB_fk_timestring} ) { + + push @display_lines, 'Reason: ' . ($innodb_status->{IB_fk_reason} || 'unknown'); + + # Display FK errors caused by invalid DML. + if ( $innodb_status->{IB_fk_txn} ) { + my $txn = $innodb_status->{IB_fk_txn}; + push @display_lines, + '', + "User $txn->{user} from $txn->{hostname}, thread $txn->{mysql_thread_id} was executing:", + '', no_ctrl_char($txn->{query_text}); + } + + my @fk_table = create_table2( + $tbl_meta{fk_error}->{visible}, + meta_to_hdr('fk_error'), + extract_values($innodb_status, $innodb_status, $innodb_status, 'fk_error'), + { just => '-', sep => ' '}); + push @display_lines, '', @fk_table; + + } + else { + push @display_lines, '', 'No foreign key error data.'; + } + draw_screen(\@display_lines, { raw => 1 } ); +} + +# display_I {{{3 +sub display_I { + my @display_lines; + my @cxns = get_connections(); + get_status_info(@cxns); + get_innodb_status(\@cxns); + + my @io_threads; + my @pending_io; + my @file_io_misc; + my @log_statistics; + my %rows_for = ( + io_threads => \@io_threads, + pending_io => \@pending_io, + file_io_misc => \@file_io_misc, + log_statistics => \@log_statistics, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + foreach my $cxn ( @cxns ) { + my $set = $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock-1} || $set; + + if ( $set->{IB_io_complete} ) { + if ( $wanted{io_threads} ) { + my $cur_threads = $set->{IB_io_threads}; + my $pre_threads = $pre->{IB_io_threads} || $cur_threads; + foreach my $key ( sort keys %$cur_threads ) { + my $cur_thd = $cur_threads->{$key}; + my $pre_thd = $pre_threads->{$key} || $cur_thd; + my $hash = extract_values($cur_thd, $cur_thd, $pre_thd, 'io_threads'); + $hash->{cxn} = $cxn; + push @io_threads, $hash; + } + } + if ( $wanted{pending_io} ) { + push @pending_io, extract_values($set, $set, $pre, 'pending_io'); + } + if ( $wanted{file_io_misc} ) { + push @file_io_misc, extract_values( + $config{status_inc}->{val} ? inc(0, $cxn) : $set, + $set, $pre, 'file_io_misc'); + } + } + if ( $set->{IB_lg_complete} && $wanted{log_statistics} ) { + push @log_statistics, extract_values($set, $set, $pre, 'log_statistics'); + } + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_K {{{3 +sub display_K { + my @display_lines; + my @cxns = get_connections(); + + my %rows_for = ( + innodb_blocked_blocker => [], + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + # Get info on locks + if ( $wanted{innodb_blocked_blocker} ) { + my @rows = get_innodb_blocked_blocker(@cxns); + push @{$rows_for{innodb_blocked_blocker}}, map { extract_values($_, $_, $_, 'innodb_blocked_blocker') } @rows; + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + # Save queries in global variable for analysis. The rows in %rows_for have been + # filtered, etc as a side effect of set_to_tbl(), so they are the same as the rows + # that get pushed to the screen. + @current_queries = map { + my %hash; + @hash{ qw(cxn id user host db query time) } + = @{$_}{ qw(cxn blocking_thread blocking_user blocking_host blocking_db blocking_query blocking_age) }; + # time is in fuzzy-time format; convert into something ascii-sortable. + $hash{time} = sprintf('%012s', fuzzy_to_secs($hash{time})); + $hash{host} =~ s/:.*$// if $hash{host}; + \%hash; + } @{$rows_for{innodb_blocked_blocker}}; + + draw_screen(\@display_lines); +} + +# display_L {{{3 +sub display_L { + my @display_lines; + my @cxns = get_connections(); + get_status_info(@cxns); + get_innodb_status(\@cxns); + + my @innodb_locks; + my %rows_for = ( + innodb_locks => \@innodb_locks, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + # Get info on locks + foreach my $cxn ( @cxns ) { + my $set = $vars{$cxn}->{$clock} or next; + my $pre = $vars{$cxn}->{$clock-1} || $set; + + if ( $wanted{innodb_locks} && defined $set->{IB_tx_transactions} && @{$set->{IB_tx_transactions}} ) { + + my $cur_txns = $set->{IB_tx_transactions}; + my $pre_txns = $pre->{IB_tx_transactions} || $cur_txns; + my %cur_txns = map { $_->{mysql_thread_id} => $_ } grep { defined $_->{mysql_thread_id} } @$cur_txns; + my %pre_txns = map { $_->{mysql_thread_id} => $_ } grep { defined $_->{mysql_thread_id} } @$pre_txns; + foreach my $txn ( @$cur_txns ) { + foreach my $lock ( @{$txn->{locks}} ) { + my %hash = map { $_ => $txn->{$_} } qw(txn_id mysql_thread_id lock_wait_time active_secs); + map { $hash{$_} = $lock->{$_} } qw(lock_type space_id page_no n_bits index db table txn_id lock_mode special insert_intention waiting); + $hash{cxn} = $cxn; + push @innodb_locks, extract_values(\%hash, \%hash, \%hash, 'innodb_locks'); + } + } + } + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_M {{{3 +sub display_M { + my @display_lines; + my @cxns = get_connections(); + get_master_slave_status(@cxns); + get_status_info(@cxns); + + my @slave_sql_status; + my @slave_io_status; + my @master_status; + my %rows_for = ( + slave_sql_status => \@slave_sql_status, + slave_io_status => \@slave_io_status, + master_status => \@master_status, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + foreach my $cxn ( @cxns ) { + my $linecount=0; + my $sth = do_stmt($cxn, 'GET_CHANNELS'); + my ( $channel ); + $sth->execute(); + $sth->bind_columns( \$channel ); + while ( $sth->fetch() ) { + $linecount=$linecount+1; + if ( length $channel < 1 ) { + $channel = 'no_channels'; + } + my $chcxn = $channel . '=' . $cxn; + get_slave_status($cxn,$channel); + my $set = $config{status_inc}->{val} ? inc(0, $chcxn) : $vars{$chcxn}->{$clock}; + my $pre = $vars{$chcxn}->{$clock - 1} || $set; + if ( $wanted{slave_sql_status} ) { + push @slave_sql_status, extract_values($set, $set, $pre, 'slave_sql_status'); + } + if ( $wanted{slave_io_status} ) { + push @slave_io_status, extract_values($set, $set, $pre, 'slave_io_status'); + } + } + if ( $linecount < 1 ) { + $channel = 'no_channels'; + my $chcxn = $channel . '=' . $cxn; + get_slave_status($cxn,$channel); + my $set = $config{status_inc}->{val} ? inc(0, $chcxn) : $vars{$chcxn}->{$clock}; + my $pre = $vars{$chcxn}->{$clock - 1} || $set; + if ( $wanted{slave_sql_status} ) { + push @slave_sql_status, extract_values($set, $set, $pre, 'slave_sql_status'); + } + if ( $wanted{slave_io_status} ) { + push @slave_io_status, extract_values($set, $set, $pre, 'slave_io_status'); + } + } + my $set = $config{status_inc}->{val} ? inc(0, $cxn) : $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock - 1} || $set; + if ( $wanted{master_status} ) { + push @master_status, extract_values($set, $set, $pre, 'master_status'); + } + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_O {{{3 +sub display_O { + my @display_lines = (''); + my @cxns = get_connections(); + my @open_tables = get_open_tables(@cxns); + my @tables = map { extract_values($_, $_, $_, 'open_tables') } @open_tables; + push @display_lines, set_to_tbl(\@tables, 'open_tables'), get_cxn_errors(@cxns); + draw_screen(\@display_lines); +} + +# display_P {{{3 +sub display_P { + my @display_lines; + + my @table_statistics; + my @index_statistics; + my @index_table_statistics; + my %rows_for = ( + table_statistics => \@table_statistics, + index_statistics => \@index_statistics, + index_table_statistics => \@index_table_statistics, + ); + + my @visible = $opts{n} ? 'index_table_statistics' : get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + # Get the data + my @cxns = get_connections(); + + if ( $wanted{table_statistics} ) { + my @rows = get_table_statistics(@cxns); + push @table_statistics, map { extract_values($_, $_, $_, 'table_statistics') } @rows; + } + elsif ( $wanted{index_statistics} ) { + my @rows = get_index_statistics(@cxns); + push @index_statistics, map { extract_values($_, $_, $_, 'index_statistics') } @rows; + } + elsif ( $wanted{index_table_statistics} ) { + my @rows = get_index_table_statistics(@cxns); + push @index_table_statistics, map { extract_values($_, $_, $_, 'index_table_statistics') } @rows; + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + next unless $wanted{$tbl}; + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_Q {{{3 +sub display_Q { + my @display_lines; + + my @q_header; + my @processlist; + my %rows_for = ( + q_header => \@q_header, + processlist => \@processlist, + ); + + my @visible = $opts{n} ? 'processlist' : get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + # Get the data + my @cxns = get_connections(); + my @full_processlist = get_full_processlist(@cxns); + + # Create header + if ( $wanted{q_header} ) { + get_status_info(@cxns); + foreach my $cxn ( @cxns ) { + my $set = $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock-1} || $set; + my $hash = extract_values($set, $set, $pre, 'q_header'); + $hash->{cxn} = $cxn; + $hash->{when} = 'Total'; + push @q_header, $hash; + + if ( exists $vars{$cxn}->{$clock - 1} ) { + my $inc = inc(0, $cxn); + my $hash = extract_values($inc, $set, $pre, 'q_header'); + $hash->{cxn} = $cxn; + $hash->{when} = 'Now'; + push @q_header, $hash; + } + } + } + + if ( $wanted{processlist} ) { + # TODO: save prev values + push @processlist, map { extract_values($_, $_, $_, 'processlist') } @full_processlist; + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + next unless $wanted{$tbl}; + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + # Save queries in global variable for analysis. The rows in %rows_for have been + # filtered, etc as a side effect of set_to_tbl(), so they are the same as the rows + # that get pushed to the screen. + @current_queries = map { + my %hash; + @hash{ qw(cxn id db query time user host) } + = @{$_}{ qw(cxn mysql_thread_id db info time user hostname) }; + # time is in seconds-to-time format; convert into something + # ascii-sortable. + $hash{time} = sprintf('%012s', $hash{time} =~ m/^([^.]*)/); + \%hash; + } @{$rows_for{processlist}}; + + draw_screen(\@display_lines); +} + +# display_R {{{3 +sub display_R { + my @display_lines; + my @cxns = get_connections(); + get_status_info(@cxns); + get_innodb_status(\@cxns); + + my @row_operations; + my @row_operation_misc; + my @semaphores; + my @wait_array; + my %rows_for = ( + row_operations => \@row_operations, + row_operation_misc => \@row_operation_misc, + semaphores => \@semaphores, + wait_array => \@wait_array, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + my $incvar = $config{status_inc}->{val}; + + foreach my $cxn ( @cxns ) { + my $set = $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock-1} || $set; + my $inc; # Only assigned to if wanted + + if ( $set->{IB_ro_complete} ) { + if ( $wanted{row_operations} ) { + $inc ||= $incvar ? inc(0, $cxn) : $set; + push @row_operations, extract_values($inc, $set, $pre, 'row_operations'); + } + if ( $wanted{row_operation_misc} ) { + push @row_operation_misc, extract_values($set, $set, $pre, 'row_operation_misc'), + } + } + + if ( $set->{IB_sm_complete} && $wanted{semaphores} ) { + $inc ||= $incvar ? inc(0, $cxn) : $set; + push @semaphores, extract_values($inc, $set, $pre, 'semaphores'); + } + + if ( $set->{IB_sm_wait_array_size} && $wanted{wait_array} ) { + foreach my $wait ( @{$set->{IB_sm_waits}} ) { + my $hash = extract_values($wait, $wait, $wait, 'wait_array'); + $hash->{cxn} = $cxn; + push @wait_array, $hash; + } + } + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + draw_screen(\@display_lines); +} + +# display_T {{{3 +sub display_T { + my @display_lines; + + my @t_header; + my @innodb_transactions; + my %rows_for = ( + t_header => \@t_header, + innodb_transactions => \@innodb_transactions, + ); + + my @visible = $opts{n} ? 'innodb_transactions' : get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + + my @cxns = get_connections(); + + get_status_info(@cxns); + + # If the header is to be shown, buffer pool data is required. + get_innodb_status( \@cxns, [ $wanted{t_header} ? qw(bp) : () ] ); + + foreach my $cxn ( get_connections() ) { + my $set = $vars{$cxn}->{$clock}; + my $pre = $vars{$cxn}->{$clock-1} || $set; + + next unless $set->{IB_tx_transactions}; + + if ( $wanted{t_header} ) { + my $hash = extract_values($set, $set, $pre, 't_header'); + push @t_header, $hash; + } + + if ( $wanted{innodb_transactions} ) { + my $cur_txns = $set->{IB_tx_transactions}; + my $pre_txns = $pre->{IB_tx_transactions} || $cur_txns; + my %cur_txns = map { $_->{mysql_thread_id} => $_ } grep { defined $_->{mysql_thread_id} } @$cur_txns; + my %pre_txns = map { $_->{mysql_thread_id} => $_ } grep { defined $_->{mysql_thread_id} } @$pre_txns; + foreach my $thd_id ( sort keys %cur_txns ) { + my $cur_txn = $cur_txns{$thd_id}; + my $pre_txn = $pre_txns{$thd_id} || $cur_txn; + my $hash = extract_values($cur_txn, $cur_txn, $pre_txn, 'innodb_transactions'); + $hash->{cxn} = $cxn; + push @innodb_transactions, $hash; + } + } + + } + + my $first_table = 0; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + + # Save queries in global variable for analysis. The rows in %rows_for have been + # filtered, etc as a side effect of set_to_tbl(), so they are the same as the rows + # that get pushed to the screen. + @current_queries = map { + my %hash; + @hash{ qw(cxn id db query time user host) } + = @{$_}{ qw(cxn mysql_thread_id db query_text active_secs user hostname) }; + \%hash; + } @{$rows_for{innodb_transactions}}; + + draw_screen(\@display_lines); +} + +# display_S {{{3 +sub display_S { + my $fmt = get_var_set('S_set'); + my $func = $config{S_func}->{val}; + my $inc = $func eq 'g' || $config{status_inc}->{val}; + + # The table's meta-data is generated from the compiled var_set. + my ( $cols, $visible ); + if ( $tbl_meta{var_status}->{fmt} && $fmt eq $tbl_meta{var_status}->{fmt} ) { + ( $cols, $visible ) = @{$tbl_meta{var_status}}{qw(cols visible)}; + } + else { + ( $cols, $visible ) = compile_select_stmt($fmt); + + # Apply missing values to columns. Always apply averages across all connections. + map { + $_->{agg} = 'avg'; + $_->{label} = $_->{hdr}; + } values %$cols; + + $tbl_meta{var_status}->{cols} = $cols; + $tbl_meta{var_status}->{visible} = $visible; + $tbl_meta{var_status}->{fmt} = $fmt; + map { $tbl_meta{var_status}->{cols}->{$_}->{just} = ''} @$visible; + } + + my @var_status; + my %rows_for = ( + var_status => \@var_status, + ); + + my @visible = get_visible_tables(); + my %wanted = map { $_ => 1 } @visible; + my @cxns = get_connections(); + + get_status_info(@cxns); + get_innodb_status(\@cxns); + + # Set up whether to pivot and how many sets to extract. + $tbl_meta{var_status}->{pivot} = $func eq 'v'; + + my $num_sets + = $func eq 'v' + ? $config{num_status_sets}->{val} + : 0; + foreach my $set ( 0 .. $num_sets ) { + my @rows; + foreach my $cxn ( @cxns ) { + my $vars = $inc ? inc($set, $cxn) : $vars{$cxn}->{$clock - $set}; + my $cur = $vars{$cxn}->{$clock-$set}; + my $pre = $vars{$cxn}->{$clock-$set-1} || $cur; + next unless $vars && %$vars; + my $hash = extract_values($vars, $cur, $pre, 'var_status'); + push @rows, $hash; + } + @rows = apply_group_by('var_status', [], @rows); + push @var_status, @rows; + } + + # Recompile the sort func. TODO: avoid recompiling at every refresh. + # Figure out whether the data is all numeric and decide on a sort type. + # my $cmp + # = scalar( + # grep { !defined $_ || $_ !~ m/^\d+$/ } + # map { my $col = $_; map { $_->{$col} } @var_status } + # $tbl_meta{var_status}->{sort_cols} =~ m/(\w+)/g) + # ? 'cmp' + # : '<=>'; + $tbl_meta{var_status}->{sort_func} = make_sort_func($tbl_meta{var_status}); + + # ################################################################ + # Now there is specific display code based on $config{S_func} + # ################################################################ + if ( $func =~ m/s|g/ ) { + my $min_width = 4; + + # Clear the screen if the display width changed. + if ( @last_term_size && $this_term_size[0] != $last_term_size[0] ) { + $lines_printed = 0; + $clear_screen_sub->(); + } + + if ( $func eq 's' ) { + # Decide how wide columns should be. + my $num_cols = scalar(@$visible); + my $width = $opts{n} ? 0 : max($min_width, int(($this_term_size[0] - $num_cols + 1) / $num_cols)); + my $g_format = $opts{n} ? ( "%s\t" x $num_cols ) : ( "%-${width}s " x $num_cols ); + + # Print headers every now and then. Headers can get really long, so compact them. + my @hdr = @$visible; + if ( $opts{n} ) { + if ( $lines_printed == 0 ) { + print join("\t", @hdr), "\n"; + $lines_printed++; + } + } + elsif ( $lines_printed == 0 || $lines_printed > $this_term_size[1] - 2 ) { + @hdr = map { donut(crunch($_, $width), $width) } @hdr; + print join(' ', map { sprintf( "%${width}s", donut($_, $width)) } @hdr) . "\n"; + $lines_printed = 1; + } + + # Design a column format for the values. + my $format + = $opts{n} + ? join("\t", map { '%s' } @$visible) . "\n" + : join(' ', map { "%${width}s" } @hdr) . "\n"; + + foreach my $row ( @var_status ) { + printf($format, map { defined $_ ? $_ : '' } @{$row}{ @$visible }); + $lines_printed++; + } + } + else { # 'g' mode + # Design a column format for the values. + my $num_cols = scalar(@$visible); + my $width = $opts{n} ? 0 : int(($this_term_size[0] - $num_cols + 1) / $num_cols); + my $format = $opts{n} ? ( "%s\t" x $num_cols ) : ( "%-${width}s " x $num_cols ); + $format =~ s/\s$/\n/; + + # Print headers every now and then. + if ( $opts{n} ) { + if ( $lines_printed == 0 ) { + print join("\t", @$visible), "\n"; + print join("\t", map { shorten($mvs{$_}) } @$visible), "\n"; + } + } + elsif ( $lines_printed == 0 || $lines_printed > $this_term_size[1] - 2 ) { + printf($format, map { donut(crunch($_, $width), $width) } @$visible); + printf($format, map { shorten($mvs{$_} || 0) } @$visible); + $lines_printed = 2; + } + + # Update the max ever seen, and scale by the max ever seen. + my $set = $var_status[0]; + foreach my $col ( @$visible ) { + $set->{$col} = 1 unless defined $set->{$col} && $set->{$col} =~ m/$num_regex/; + $set->{$col} = ($set->{$col} || 1) / ($set->{Uptime_hires} || 1); + $mvs{$col} = max($mvs{$col} || 1, $set->{$col}); + $set->{$col} /= $mvs{$col}; + } + printf($format, map { ( $config{graph_char}->{val} x int( $width * $set->{$_} )) || '.' } @$visible ); + $lines_printed++; + + } + } + else { # 'v' + my $first_table = 0; + my @display_lines; + foreach my $tbl ( @visible ) { + push @display_lines, '', set_to_tbl($rows_for{$tbl}, $tbl); + push @display_lines, get_cxn_errors(@cxns) + if ( $config{debug}->{val} || !$first_table++ ); + } + $clear_screen_sub->(); + draw_screen( \@display_lines ); + } +} + +# display_explain {{{3 +sub display_explain { + my $info = shift; + my $cxn = $info->{cxn}; + my $db = $info->{db}; + + my ( $mods, $query ) = rewrite_for_explain($info->{query}); + + my @display_lines; + + if ( $query ) { + + my $part = version_ge($dbhs{$cxn}->{dbh}, '5.1.5') ? 'PARTITIONS' : ''; + $query = "EXPLAIN $part\n" . $query; + + eval { + if ( $db ) { + do_query($cxn, "use $db"); + } + my $sth = do_query($cxn, $query); + + my $res; + while ( $res = $sth->fetchrow_hashref() ) { + map { $res->{$_} ||= '' } ( 'partitions', keys %$res); + my @this_table = create_caption("Sub-Part $res->{id}", + create_table2( + $tbl_meta{explain}->{visible}, + meta_to_hdr('explain'), + extract_values($res, $res, $res, 'explain'))); + @display_lines = stack_next(\@display_lines, \@this_table, { pad => ' ', vsep => 2 }); + } + }; + + if ( $EVAL_ERROR ) { + push @display_lines, + '', + "The query could not be explained. Only SELECT queries can be " + . "explained; innotop tries to rewrite certain REPLACE and INSERT queries " + . "into SELECT, but this doesn't always succeed."; + } + + } + else { + push @display_lines, '', 'The query could not be explained.'; + } + + if ( $mods ) { + push @display_lines, '', '[This query has been re-written to be explainable]'; + } + + unshift @display_lines, no_ctrl_char($query); + draw_screen(\@display_lines, { raw => 1 } ); +} + +# rewrite_for_explain {{{3 +sub rewrite_for_explain { + my $query = shift; + + my $mods = 0; + my $orig = $query; + $mods += $query =~ s/^\s*(?:replace|insert).*?select/select/is; + $mods += $query =~ s/^ + \s*create\s+(?:temporary\s+)?table + \s+(?:\S+\s+)as\s+select/select/xis; + $mods += $query =~ s/\s+on\s+duplicate\s+key\s+update.*$//is; + return ( $mods, $query ); +} + +# show_optimized_query {{{3 +sub show_optimized_query { + my $info = shift; + my $cxn = $info->{cxn}; + my $db = $info->{db}; + my $meta = $dbhs{$cxn}; + + my @display_lines; + + my ( $mods, $query ) = rewrite_for_explain($info->{query}); + + if ( $mods ) { + push @display_lines, '[This query has been re-written to be explainable]'; + } + + if ( $query ) { + push @display_lines, no_ctrl_char($info->{query}); + + eval { + if ( $db ) { + do_query($cxn, "use $db"); + } + do_query( $cxn, 'EXPLAIN EXTENDED ' . $query ) or die "Can't explain query"; + my $sth = do_query($cxn, 'SHOW WARNINGS'); + my $res = $sth->fetchall_arrayref({}); + + if ( $res ) { + foreach my $result ( @$res ) { + push @display_lines, 'Note:', no_ctrl_char($result->{message}); + } + } + else { + push @display_lines, '', 'The query optimization could not be generated.'; + } + }; + + if ( $EVAL_ERROR ) { + push @display_lines, '', "The optimization could not be generated: $EVAL_ERROR"; + } + + } + else { + push @display_lines, '', 'The query optimization could not be generated.'; + } + + draw_screen(\@display_lines, { raw => 1 } ); +} + +# display_help {{{3 +sub display_help { + my $mode = $config{mode}->{val}; + + # Get globally mapped keys, then overwrite them with mode-specific ones. + my %keys = map { + $_ => $action_for{$_}->{label} + } keys %action_for; + foreach my $key ( keys %{$modes{$mode}->{action_for}} ) { + $keys{$key} = $modes{$mode}->{action_for}->{$key}->{label}; + } + delete $keys{'?'}; + + # Split them into three kinds of keys: MODE keys, action keys, and + # magic (special character) keys. + my @modes = sort grep { m/[A-Z]/ } keys %keys; + my @actions = sort grep { m/[a-z]/ } keys %keys; + my @magic = sort grep { m/[^A-Z]/i } keys %keys; + + my @display_lines = ( '', 'Switch to a different mode:' ); + + # Mode keys + my @all_modes = map { "$_ $modes{$_}->{hdr}" } @modes; + my @col1 = splice(@all_modes, 0, ceil(@all_modes/3)); + my @col2 = splice(@all_modes, 0, ceil(@all_modes/2)); + my $max1 = max(map {length($_)} @col1); + my $max2 = max(map {length($_)} @col2); + while ( @col1 ) { + push @display_lines, sprintf(" %-${max1}s %-${max2}s %s", + (shift @col1 || ''), + (shift @col2 || ''), + (shift @all_modes || '')); + } + + # Action keys + my @all_actions = map { "$_ $keys{$_}" } @actions; + @col1 = splice(@all_actions, 0, ceil(@all_actions/2)); + $max1 = max(map {length($_)} @col1); + push @display_lines, '', 'Actions:'; + while ( @col1 ) { + push @display_lines, sprintf(" %-${max1}s %s", + (shift @col1 || ''), + (shift @all_actions || '')); + } + + # Magic keys + my @all_magic = map { + my $k = $action_for{$_} ? ($action_for{$_}->{key} || $_) : $_; + sprintf('%4s %s', $k, $keys{$_}); + } @magic; + @col1 = splice(@all_magic, 0, ceil(@all_magic/2)); + $max1 = max(map {length($_)} @col1); + push @display_lines, '', 'Other:'; + while ( @col1 ) { + push @display_lines, sprintf("%-${max1}s%s", + (shift @col1 || ''), + (shift @all_magic || '')); + } + + $clear_screen_sub->(); + draw_screen(\@display_lines, { show_all => 1 } ); + pause(); + $clear_screen_sub->(); +} + +# show_full_query {{{3 +sub show_full_query { + my $info = shift; + my @display_lines = no_ctrl_char($info->{query}); + draw_screen(\@display_lines, { raw => 1 }); +} + +# Formatting functions {{{2 + +# create_table2 {{{3 +# Makes a two-column table, labels on left, data on right. +# Takes refs of @cols, %labels and %data, %user_prefs +sub create_table2 { + my ( $cols, $labels, $data, $user_prefs ) = @_; + my @rows; + + if ( @$cols && %$data ) { + + # Override defaults + my $p = { + just => '', + sep => ':', + just1 => '-', + }; + if ( $user_prefs ) { + map { $p->{$_} = $user_prefs->{$_} } keys %$user_prefs; + } + + # Fix undef values + map { $data->{$_} = '' unless defined $data->{$_} } @$cols; + + # Format the table + my $max_l = max(map{ length($labels->{$_}) } @$cols); + my $max_v = max(map{ length($data->{$_}) } @$cols); + my $format = "%$p->{just}${max_l}s$p->{sep} %$p->{just1}${max_v}s"; + foreach my $col ( @$cols ) { + push @rows, sprintf($format, $labels->{$col}, $data->{$col}); + } + } + return @rows; +} + +# stack_next {{{3 +# Stacks one display section next to the other. Accepts left-hand arrayref, +# right-hand arrayref, and options hashref. Tries to stack as high as +# possible, so +# aaaaaa +# bbb +# can stack ccc next to the bbb. +# NOTE: this DOES modify its arguments, even though it returns a new array. +sub stack_next { + my ( $left, $right, $user_prefs ) = @_; + my @result; + + my $p = { + pad => ' ', + vsep => 0, + }; + if ( $user_prefs ) { + map { $p->{$_} = $user_prefs->{$_} } keys %$user_prefs; + } + + # Find out how wide the LHS can be and still let the RHS fit next to it. + my $pad = $p->{pad}; + my $max_r = max( map { length($_) } @$right) || 0; + my $max_l = $this_term_size[0] - $max_r - length($pad); + + # Find the minimum row on the LHS that the RHS will fit next to. + my $i = scalar(@$left) - 1; + while ( $i >= 0 && length($left->[$i]) <= $max_l ) { + $i--; + } + $i++; + my $offset = $i; + + if ( $i < scalar(@$left) ) { + # Find the max width of the section of the LHS against which the RHS + # will sit. + my $max_i_in_common = min($i + scalar(@$right) - 1, scalar(@$left) - 1); + my $max_width = max( map { length($_) } @{$left}[$i..$max_i_in_common]); + + # Append the RHS onto the LHS until one runs out. + while ( $i < @$left && $i - $offset < @$right ) { + my $format = "%-${max_width}s$pad%${max_r}s"; + $left->[$i] = sprintf($format, $left->[$i], $right->[$i - $offset]); + $i++; + } + while ( $i - $offset < @$right ) { + # There is more RHS to push on the end of the array + push @$left, + sprintf("%${max_width}s$pad%${max_r}s", ' ', $right->[$i - $offset]); + $i++; + } + push @result, @$left; + } + else { + # There is no room to put them side by side. Add them below, with + # a blank line above them if specified. + push @result, @$left; + push @result, (' ' x $this_term_size[0]) if $p->{vsep} && @$left; + push @result, @$right; + } + return @result; +} + +# create_caption {{{3 +sub create_caption { + my ( $caption, @rows ) = @_; + if ( @rows ) { + + # Calculate the width of what will be displayed, so it can be centered + # in that space. When the thing is wider than the display, center the + # caption in the display. + my $width = min($this_term_size[0], max(map { length(ref($_) ? $_->[0] : $_) } @rows)); + + my $cap_len = length($caption); + + # It may be narrow enough to pad the sides with underscores and save a + # line on the screen. + if ( $cap_len <= $width - 6 ) { + my $left = int(($width - 2 - $cap_len) / 2); + unshift @rows, + ("_" x $left) . " $caption " . ("_" x ($width - $left - $cap_len - 2)); + } + + # The caption is too wide to add underscores on each side. + else { + + # Color is supported, so we can use terminal underlining. + if ( $config{color}->{val} ) { + my $left = int(($width - $cap_len) / 2); + unshift @rows, [ + (" " x $left) . $caption . (" " x ($width - $left - $cap_len)), + 'underline', + ]; + } + + # Color is not supported, so we have to add a line underneath to separate the + # caption from whatever it's captioning. + else { + my $left = int(($width - $cap_len) / 2); + unshift @rows, ('-' x $width); + unshift @rows, (" " x $left) . $caption . (" " x ($width - $left - $cap_len)); + } + + # The caption is wider than the thing it labels, so we have to pad the + # thing it labels to a consistent width. + if ( $cap_len > $width ) { + @rows = map { + ref($_) + ? [ sprintf('%-' . $cap_len . 's', $_->[0]), $_->[1] ] + : sprintf('%-' . $cap_len . 's', $_); + } @rows; + } + + } + } + return @rows; +} + +# create_table {{{3 +# Input: an arrayref of columns, hashref of col info, and an arrayref of hashes +# Example: [ 'a', 'b' ] +# { a => spec, b => spec } +# [ { a => 1, b => 2}, { a => 3, b => 4 } ] +# The 'spec' is a hashref of hdr => label, just => ('-' or ''). It also supports min and max-widths +# vi the minw and maxw params. +# Output: an array of strings, one per row. +# Example: +# Column One Column Two +# ---------- ---------- +# 1 2 +# 3 4 +sub create_table { + my ( $cols, $info, $data, $prefs ) = @_; + $prefs ||= {}; + $prefs->{no_hdr} ||= ($opts{n} && $clock != 1); + + # Truncate rows that will surely be off screen even if this is the only table. + if ( !$opts{n} && !$prefs->{raw} && !$prefs->{show_all} && $this_term_size[1] < @$data-1 ) { + $data = [ @$data[0..$this_term_size[1] - 1] ]; + } + + my @rows = (); + + if ( @$cols && %$info ) { + + # Fix undef values, collapse whitespace. + foreach my $row ( @$data ) { + map { $row->{$_} = collapse_ws($row->{$_}) } @$cols; + } + + my $col_sep = $opts{n} ? "\t" : ' '; + + # Find each column's max width. + my %width_for; + if ( !$opts{n} ) { + %width_for = map { + my $col_name = $_; + if ( $info->{$_}->{dec} ) { + # Align along the decimal point + my $max_rodp = max(0, map { $_->{$col_name} =~ m/([^\s\d-].*)$/ ? length($1) : 0 } @$data); + foreach my $row ( @$data ) { + my $col = $row->{$col_name}; + my ( $l, $r ) = $col =~ m/^([\s\d]*)(.*)$/; + $row->{$col_name} = sprintf("%s%-${max_rodp}s", $l, $r); + } + } + my $max_width = max( length($info->{$_}->{hdr}), map { length($_->{$col_name}) } @$data); + if ( $info->{$col_name}->{maxw} ) { + $max_width = min( $max_width, $info->{$col_name}->{maxw} ); + } + if ( $info->{$col_name}->{minw} ) { + $max_width = max( $max_width, $info->{$col_name}->{minw} ); + } + $col_name => $max_width; + } @$cols; + } + + # The table header. + if ( !$config{hide_hdr}->{val} && !$prefs->{no_hdr} ) { + push @rows, $opts{n} + ? join( $col_sep, @$cols ) + : join( $col_sep, map { sprintf( "%-$width_for{$_}s", trunc($info->{$_}->{hdr}, $width_for{$_}) ) } @$cols ); + if ( $config{color}->{val} && $config{header_highlight}->{val} ) { + push @rows, [ pop @rows, $config{header_highlight}->{val} ]; + } + elsif ( !$opts{n} ) { + push @rows, join( $col_sep, map { "-" x $width_for{$_} } @$cols ); + } + } + + # The table data. + if ( $opts{n} ) { + foreach my $item ( @$data ) { + push @rows, join($col_sep, map { $item->{$_} } @$cols ); + } + } + else { + my $format = join( $col_sep, + map { "%$info->{$_}->{just}$width_for{$_}s" } @$cols ); + foreach my $item ( @$data ) { + my $row = sprintf($format, map { trunc($item->{$_}, $width_for{$_}) } @$cols ); + if ( $config{color}->{val} && $item->{_color} ) { + push @rows, [ $row, $item->{_color} ]; + } + else { + push @rows, $row; + } + } + } + } + + return @rows; +} + +# Aggregates a table. If $group_by is an arrayref of columns, the grouping key +# is the specified columns; otherwise it's just the empty string (e.g. +# everything is grouped as one group). +sub apply_group_by { + my ( $tbl, $group_by, @rows ) = @_; + my $meta = $tbl_meta{$tbl}; + my %is_group = map { $_ => 1 } @$group_by; + my @non_grp = grep { !$is_group{$_} } keys %{$meta->{cols}}; + + my %temp_table; + foreach my $row ( @rows ) { + my $group_key + = @$group_by + ? '{' . join('}{', map { defined $_ ? $_ : '' } @{$row}{@$group_by}) . '}' + : ''; + $temp_table{$group_key} ||= []; + push @{$temp_table{$group_key}}, $row; + } + + # Crush the rows together... + my @new_rows; + foreach my $key ( sort keys %temp_table ) { + my $group = $temp_table{$key}; + my %new_row; + @new_row{@$group_by} = @{$group->[0]}{@$group_by}; + foreach my $col ( @non_grp ) { + my $agg = $meta->{cols}->{$col}->{agg} || 'first'; + $new_row{$col} = $agg_funcs{$agg}->( map { $_->{$col} } @$group ); + } + push @new_rows, \%new_row; + } + return @new_rows; +} + +# set_to_tbl {{{3 +# Unifies all the work of filtering, sorting etc. Alters the input. +# TODO: pull all the little pieces out into subroutines and stick events in each of them. +sub set_to_tbl { + my ( $rows, $tbl ) = @_; + my $meta = $tbl_meta{$tbl} or die "No such table $tbl in tbl_meta"; + + # don't show / hide cxn if there's only one connection being displayed + my (@visible, @group_by); + my $num_cxn = scalar get_connections(); + if ($num_cxn <= 1) { + map { push @visible, $_ if $_ !~ /^cxn$/ } @{$meta->{visible}}; + $meta->{visible} = \@visible; + map { push @group_by, $_ if $_ !~ /^cxn$/ } @{$meta->{group_by}}; + $meta->{group_by} = \@group_by; + } + # if cxn is not visible and there is now more than one connection, + # make cxn visible again. assume it's not in group_by if it's not + # visible + else { + my $has_cxn = 0; + foreach my $column (@{$meta->{visible}}) { + if ($column eq "cxn") { + $has_cxn = 1; + last; + } + } + if (not $has_cxn) { + map { push @visible, $_ if $_ !~ /^cxn$/ } @{$meta->{visible}}; + $meta->{visible} = \@visible; + map { push @group_by, $_ if $_ !~ /^cxn$/ } @{$meta->{group_by}}; + $meta->{group_by} = \@group_by; + } + } + + if ( !$meta->{pivot} ) { + + # Hook in event listeners + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_filter}} ) { + $listener->set_to_tbl_pre_filter($rows, $tbl); + } + + # Apply filters. Note that if the table is pivoted, filtering and sorting + # are applied later. + foreach my $filter ( @{$meta->{filters}} ) { + eval { + @$rows = grep { $filters{$filter}->{func}->($_) } @$rows; + }; + if ( $EVAL_ERROR && $config{debug}->{val} ) { + die $EVAL_ERROR; + } + } + + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_sort}} ) { + $listener->set_to_tbl_pre_sort($rows, $tbl); + } + + # Sort. Note that if the table is pivoted, sorting might have the wrong + # columns and it could crash. This will only be an issue if it's possible + # to toggle pivoting on and off, which it's not at the moment. + if ( @$rows && $meta->{sort_func} && !$meta->{aggregate} ) { + if ( $meta->{sort_dir} > 0 ) { + @$rows = $meta->{sort_func}->( @$rows ); + } + else { + @$rows = reverse $meta->{sort_func}->( @$rows ); + } + } + + } + + # Stop altering arguments now. + my @rows = @$rows; + + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_group}} ) { + $listener->set_to_tbl_pre_group(\@rows, $tbl); + } + + # Apply group-by. + if ( $meta->{aggregate} ) { + @rows = apply_group_by($tbl, $meta->{group_by}, @rows); + + # Sort. Note that if the table is pivoted, sorting might have the wrong + # columns and it could crash. This will only be an issue if it's possible + # to toggle pivoting on and off, which it's not at the moment. + if ( @rows && $meta->{sort_func} ) { + if ( $meta->{sort_dir} > 0 ) { + @rows = $meta->{sort_func}->( @rows ); + } + else { + @rows = reverse $meta->{sort_func}->( @rows ); + } + } + + } + + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_colorize}} ) { + $listener->set_to_tbl_pre_colorize(\@rows, $tbl); + } + + if ( !$meta->{pivot} ) { + # Colorize. Adds a _color column to rows. + if ( @rows && $meta->{color_func} ) { + eval { + foreach my $row ( @rows ) { + $row->{_color} = $meta->{color_func}->($row); + } + }; + if ( $EVAL_ERROR ) { + pause($EVAL_ERROR); + } + } + } + + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_transform}} ) { + $listener->set_to_tbl_pre_transform(\@rows, $tbl); + } + + # Apply_transformations. + if ( @rows ) { + my $cols = $meta->{cols}; + foreach my $col ( keys %{$rows->[0]} ) { + # Don't auto-vivify $tbl_meta{tbl}-{cols}->{_color}->{trans} + next if $col eq '_color'; + foreach my $trans ( @{$cols->{$col}->{trans}} ) { + map { $_->{$col} = $trans_funcs{$trans}->($_->{$col}) } @rows; + } + } + } + + my ($fmt_cols, $fmt_meta); + + # Pivot. + if ( $meta->{pivot} ) { + + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_pivot}} ) { + $listener->set_to_tbl_pre_pivot(\@rows, $tbl); + } + + my @vars = @{$meta->{visible}}; + my @tmp = map { { name => $_ } } @vars; + my @cols = 'name'; + foreach my $i ( 0..@$rows-1 ) { + my $col = "set_$i"; + push @cols, $col; + foreach my $j ( 0..@vars-1 ) { + $tmp[$j]->{$col} = $rows[$i]->{$vars[$j]}; + } + } + $fmt_meta = { map { $_ => { hdr => $_, just => '-' } } @cols }; + $fmt_cols = \@cols; + @rows = @tmp; + + # Hook in event listeners + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_filter}} ) { + $listener->set_to_tbl_pre_filter($rows, $tbl); + } + + # Apply filters. + foreach my $filter ( @{$meta->{filters}} ) { + eval { + @rows = grep { $filters{$filter}->{func}->($_) } @rows; + }; + if ( $EVAL_ERROR && $config{debug}->{val} ) { + die $EVAL_ERROR; + } + } + + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_sort}} ) { + $listener->set_to_tbl_pre_sort($rows, $tbl); + } + + # Sort. + if ( @rows && $meta->{sort_func} ) { + if ( $meta->{sort_dir} > 0 ) { + @rows = $meta->{sort_func}->( @rows ); + } + else { + @rows = reverse $meta->{sort_func}->( @rows ); + } + } + + } + else { + # If the table isn't pivoted, just show all columns that are supposed to + # be shown; but eliminate aggonly columns if the table isn't aggregated. + my $aggregated = $meta->{aggregate}; + $fmt_cols = [ grep { $aggregated || !$meta->{cols}->{$_}->{aggonly} } @{$meta->{visible}} ]; + $fmt_meta = { map { $_ => $meta->{cols}->{$_} } @$fmt_cols }; + + # If the table is aggregated, re-order the group_by columns to the left of + # the display, and suppress 'agghide' columns. + if ( $aggregated ) { + my %is_group = map { $_ => 1 } @{$meta->{group_by}}; + $fmt_cols = [ @{$meta->{group_by}}, grep { !$is_group{$_} } @$fmt_cols ]; + $fmt_cols = [ grep { !$meta->{cols}->{$_}->{agghide} } @$fmt_cols ]; + } + } + + foreach my $listener ( @{$event_listener_for{set_to_tbl_pre_create}} ) { + $listener->set_to_tbl_pre_create(\@rows, $tbl); + } + + @rows = create_table( $fmt_cols, $fmt_meta, \@rows); + if ( !$meta->{hide_caption} && !$opts{n} && $config{display_table_captions}->{val} ) { + @rows = create_caption($meta->{capt}, @rows) + } + + foreach my $listener ( @{$event_listener_for{set_to_tbl_post_create}} ) { + $listener->set_to_tbl_post_create(\@rows, $tbl); + } + + return @rows; +} + +# meta_to_hdr {{{3 +sub meta_to_hdr { + my $tbl = shift; + my $meta = $tbl_meta{$tbl}; + my %labels = map { $_ => $meta->{cols}->{$_}->{hdr} } @{$meta->{visible}}; + return \%labels; +} + +# commify {{{3 +# From perlfaq5: add commas. +sub commify { + my ( $num ) = @_; + $num = 0 unless defined $num; + $num =~ s/(^[-+]?\d+?(?=(?>(?:\d{3})+)(?!\d))|\G\d{3}(?=\d))/$1,/g; + return $num; +} + +# set_precision {{{3 +# Trim to desired precision. +sub set_precision { + my ( $num, $precision ) = @_; + $num = 0 unless defined $num; + $precision = $config{num_digits}->{val} if !defined $precision; + sprintf("%.${precision}f", $num); +} + +# percent {{{3 +# Convert to percent +sub percent { + my ( $num ) = @_; + $num = 0 unless defined $num; + my $digits = $config{num_digits}->{val}; + return sprintf("%.${digits}f", $num * 100) + . ($config{show_percent}->{val} ? '%' : ''); +} + +# sparkify {{{3 +# Find the range (min to max) and divide it up. Each value then gets put into +# a bucket and represented by one of these characters: _.-=^ +sub sparkify { + my @vals = @_; + my @chars = qw(_ . - = ^); + my $min = min(@vals); + my $max = max(@vals); + my $range = ($max - $min) / 4; + return "_" x scalar(@vals) if !$min || !$max || $max == $min || !$range; + my $result = ""; + foreach my $v ( @vals ) { + $result .= $chars[ int(($v - $min) / $range) ]; + } + return $result; +} + +# shorten {{{3 +sub shorten { + my ( $num, $opts ) = @_; + + return $num if !defined($num) || $opts{n} || $num !~ m/$num_regex/; + + $opts ||= {}; + my $pad = defined $opts->{pad} ? $opts->{pad} : ''; + my $num_digits = defined $opts->{num_digits} + ? $opts->{num_digits} + : $config{num_digits}->{val}; + my $force = defined $opts->{force}; + + my $n = 0; + while ( $num >= 1_024 ) { + $num /= 1_024; + ++$n; + } + return $num =~ m/\./ || $n || $force + ? sprintf("%.${num_digits}f%s", $num, ($pad,'k','M','G','T')[$n]) + : $num; +} + +# Utility functions {{{2 +# unique {{{3 +sub unique { + my %seen; + return grep { !$seen{$_}++ } @_; +} + +# make_color_func {{{3 +sub make_color_func { + my ( $tbl ) = @_; + my @criteria; + foreach my $spec ( @{$tbl->{colors}} ) { + next unless exists $comp_ops{$spec->{op}}; + my $val = $spec->{op} =~ m/^(?:eq|ne|le|ge|lt|gt)$/ ? "'$spec->{arg}'" + : $spec->{op} =~ m/^(?:=~|!~)$/ ? "m/" . quotemeta($spec->{arg}) . "/" + : $spec->{arg}; + push @criteria, + "( defined \$set->{$spec->{col}} && \$set->{$spec->{col}} $spec->{op} $val ) { return '$spec->{color}'; }"; + } + return undef unless @criteria; + my $sub = eval 'sub { my ( $set ) = @_; if ' . join(" elsif ", @criteria) . '}'; + die if $EVAL_ERROR; + return $sub; +} + +# make_sort_func {{{3 +# Gets a list of sort columns from the table, like "+cxn -time" and returns a +# subroutine that will sort that way. +sub make_sort_func { + my ( $tbl ) = @_; + my @criteria; + + # Pivoted tables can be sorted by 'name' and set_x columns; others must be + # sorted by existing columns. TODO: this will crash if you toggle between + # pivoted and nonpivoted. I have several other 'crash' notes about this if + # this ever becomes possible. + + if ( $tbl->{pivot} ) { + # Sort type is not really possible on pivoted columns, because a 'column' + # contains data from an entire non-pivoted row, so there could be a mix of + # numeric and non-numeric data. Thus everything has to be 'cmp' type. + foreach my $col ( split(/\s+/, $tbl->{sort_cols} ) ) { + next unless $col; + my ( $dir, $name ) = $col =~ m/([+-])?(\w+)$/; + next unless $name && $name =~ m/^(?:name|set_\d+)$/; + $dir ||= '+'; + my $op = 'cmp'; + my $df = "''"; + push @criteria, + $dir eq '+' + ? "(\$a->{$name} || $df) $op (\$b->{$name} || $df)" + : "(\$b->{$name} || $df) $op (\$a->{$name} || $df)"; + } + } + else { + foreach my $col ( split(/\s+/, $tbl->{sort_cols} ) ) { + next unless $col; + my ( $dir, $name ) = $col =~ m/([+-])?(\w+)$/; + next unless $name && $tbl->{cols}->{$name}; + $dir ||= '+'; + my $op = $tbl->{cols}->{$name}->{num} ? "<=>" : "cmp"; + my $df = $tbl->{cols}->{$name}->{num} ? "0" : "''"; + push @criteria, + $dir eq '+' + ? "(\$a->{$name} || $df) $op (\$b->{$name} || $df)" + : "(\$b->{$name} || $df) $op (\$a->{$name} || $df)"; + } + } + return sub { return @_ } unless @criteria; + my $sub = eval 'sub { sort {' . join("||", @criteria) . '} @_; }'; + die if $EVAL_ERROR; + return $sub; +} + +# trunc {{{3 +# Shortens text to specified length. +sub trunc { + my ( $text, $len ) = @_; + if ( length($text) <= $len ) { + return $text; + } + return substr($text, 0, $len); +} + +# donut {{{3 +# Takes out the middle of text to shorten it. +sub donut { + my ( $text, $len ) = @_; + return $text if length($text) <= $len; + my $max = length($text) - $len; + my $min = $max - 1; + + # Try to remove a single "word" from somewhere in the center + if ( $text =~ s/_[^_]{$min,$max}_/_/ ) { + return $text; + } + + # Prefer removing the end of a "word" + if ( $text =~ s/([^_]+)[^_]{$max}_/$1_/ ) { + return $text; + } + + $text = substr($text, 0, int($len/2)) + . "_" + . substr($text, int($len/2) + $max + 1); + return $text; +} + +# crunch {{{3 +# Removes vowels and compacts repeated letters to shorten text. +sub crunch { + my ( $text, $len ) = @_; + return $text if $len && length($text) <= $len; + $text =~ s/^IB_\w\w_//; + $text =~ s/(?<![_ ])[aeiou]//g; + $text =~ s/(.)\1+/$1/g; + return $text; +} + +# collapse_ws {{{3 +# Collapses all whitespace to a single space. +sub collapse_ws { + my ( $text ) = @_; + return '' unless defined $text; + $text =~ s/\s+/ /g; + return $text; +} + +# Strips out non-printable characters within fields, which freak terminals out. +sub no_ctrl_char { + my ( $text ) = @_; + return '' unless defined $text; + my $charset = $config{charset}->{val}; + if ( $charset && $charset eq 'unicode' ) { + $text =~ s/ + ("(?:(?!(?<!\\)").)*" # Double-quoted string + |'(?:(?!(?<!\\)').)*') # Or single-quoted string + /$1 =~ m#\p{IsC}# ? "[BINARY]" : $1/egx; + } + elsif ( $charset && $charset eq 'none' ) { + $text =~ s/ + ("(?:(?!(?<!\\)").)*" + |'(?:(?!(?<!\\)').)*') + /[TEXT]/gx; + } + else { # The default is 'ascii' + $text =~ s/ + ("(?:(?!(?<!\\)").)*" + |'(?:(?!(?<!\\)').)*') + /$1 =~ m#[^\040-\176]# ? "[BINARY]" : $1/egx; + } + return $text; +} + +# word_wrap {{{3 +# Wraps text at word boundaries so it fits the screen. +sub word_wrap { + my ( $text, $width) = @_; + $width ||= $this_term_size[0]; + $text =~ s/(.{0,$width})(?:\s+|$)/$1\n/g; + $text =~ s/ +$//mg; + return $text; +} + +# draw_screen {{{3 +# Prints lines to the screen. The first argument is an arrayref. Each +# element of the array is either a string or an arrayref. If it's a string it +# just gets printed. If it's an arrayref, the first element is the string to +# print, and the second is args to colored(). +sub draw_screen { + my ( $display_lines, $prefs ) = @_; + if ( !$opts{n} && $config{show_statusbar}->{val} ) { + unshift @$display_lines, create_statusbar(); + } + + foreach my $listener ( @{$event_listener_for{draw_screen}} ) { + $listener->draw_screen($display_lines); + } + + $clear_screen_sub->() + if $prefs->{clear} || !$modes{$config{mode}->{val}}->{no_clear_screen}; + if ( $opts{n} || $prefs->{raw} ) { + my $num_lines = 0; + my $ts = $opts{t} ? POSIX::strftime($config{timeformat}->{val}, localtime) : ''; + if ( $opts{t} ) { + if ( $opts{t} == 1 ) { + print "\n$ts\n\n"; + $ts = ""; # Prevent it from being written on every line. + $num_lines++; + } + else { + $ts .= " "; + } + } + print join("\n", + map { + $num_lines++; + ref $_ + ? colored($ts . $_->[0], $_->[1]) + : $ts . $_; + } + grep { !$opts{n} || $_ } # Suppress empty lines + @$display_lines); + if ( $opts{n} && $num_lines ) { + print "\n"; + } + } + else { + my $max_lines = $prefs->{show_all} + ? scalar(@$display_lines)- 1 + : min(scalar(@$display_lines), $this_term_size[1]); + print join("\n", + map { + ref $_ + ? colored(substr($_->[0], 0, $this_term_size[0]), $_->[1]) + : substr($_, 0, $this_term_size[0]); + } @$display_lines[0..$max_lines - 1]); + } +} + +# fuzzy_time {{{3 +sub fuzzy_time { + my ( $secs ) = @_; + return '' unless $secs; + return sprintf('%.2f', $secs) if $secs =~ m/^.\./; + $secs =~ s/\..*$//; + return $secs < 180 ? "${secs}s" + : $secs < 3600 ? sprintf("%dm", $secs / 60) + : $secs < 3600 * 3 ? sprintf("%dh%dm", $secs / 3600, ($secs % 3600) / 60) + : $secs < 86400 ? sprintf("%dh", $secs / 3600) + : $secs < 86400* 3 ? sprintf("%dd%dh", $secs / 86400, ($secs % 86400) / 3600) + : sprintf("%dd", $secs / 86400); +} + +sub fuzzy_to_secs { + my ($t) = @_; + return 0 unless $t; + my ($num, $suffix) = $t =~ m/(\d+)([a-z])?$/; + return $num unless $suffix; + return $suffix eq 's' ? $num # Seconds + : $suffix eq 'm' ? $num * 60 # Minutes + : $suffix eq 'h' ? $num * 3600 # Hours + : $num * 86400; # Days +} + +# distill {{{3 +sub distill { + my ( $query ) = @_; + return "" unless $query; + my $orig_query = $query; + + $query =~ m/\A\s*call\s+(\S+)\(/i && return "CALL $1"; + $query =~ m/\A\s*use\s+/ && return "USE"; + $query =~ m/\A\s*UNLOCK TABLES/i && return "UNLOCK"; + $query =~ m/\A\s*xa\s+(\S+)/i && return "XA_$1"; + + # Strip out comments + my $olc_re = qr/(?:--|#)[^'"\r\n]*(?=[\r\n]|\Z)/; # One-line comments + my $mlc_re = qr#/\*[^!].*?\*/#sm; # But not /*!version */ + my $vlc_re = qr#/\*.*?[0-9+].*?\*/#sm; # For SHOW + /*!version */ + my $vlc_rf = qr#^(SHOW).*?/\*![0-9+].*?\*/#sm; # Variation for SHOW + $query =~ s/$olc_re//go; + $query =~ s/$mlc_re//go; + if ( $query =~ m/$vlc_rf/i ) { # contains show + version + $query =~ s/$vlc_re//go; + } + + # Handle SHOW queries + if ( $query =~ m/\A\s*SHOW\s+/i ) { + $query = uc $query; + $query =~ s/\s+(?:GLOBAL|SESSION|FULL|STORAGE|ENGINE)\b/ /g; + $query =~ s/\s+COUNT[^)]+\)//g; + $query =~ s/\s+(?:FOR|FROM|LIKE|WHERE|LIMIT|IN)\b.+//ms; + $query =~ s/\A(SHOW(?:\s+\S+){1,2}).*\Z/$1/s; + $query =~ s/\s+/ /g; + } + + # Find verbs and tables. + my ($verbs, $table); + + # Handle DDL operations (dds) + my $tbl_ident = qr/(?:`[^`]+`|\w+)(?:\.(?:`[^`]+`|\w+))?/; + my $tbl_regex = qr{ + \b(?:FROM|JOIN|(?<!KEY\s)UPDATE|INTO(?!\s+TABLE)|INTO\s+TABLE) # Words that precede table names + \b\s* + \(? # Optional paren around tables + ($tbl_ident + (?: (?:\s+ (?:AS\s+)? \w+)?, \s*$tbl_ident )* + ) + }xio; + my ( $ddl ) = $query =~ m/^\s*((?:CREATE|ALTER|TRUNCATE|DROP|RENAME|CHECK|REPAIR))\b/i; + if ( $ddl ) { + my ( $obj ) = $query =~ m/$ddl.+(DATABASE|EVENT|FUNCTION|INDEX|LOGFILE GROUP|PROCEDURE|SERVER|TABLE|TABLESPACE|TRIGGER|VIEW)\b/i; + if ( $obj ) { + $query =~ s/$obj\s*if\s*(?:not\s*)?exists/$obj/i; + $verbs = uc($ddl . ($obj ? " $obj" : '')); + ($table) = $query =~ m/(?:TABLE|DATABASE)\s+($tbl_ident)(\s+.*)?/i; + } + } + else { + my $verbs_pat = qr{^SHOW|^FLUSH|^COMMIT|^ROLLBACK|^BEGIN|SELECT|INSERT + |UPDATE|DELETE|REPLACE|^SET|UNION|^START|^LOCK|^LOAD}xi; + my @verbs = $query =~ m/\b($verbs_pat)\b/gio; + @verbs = do { + my $last = ''; + grep { my $pass = $_ ne $last; $last = $_; $pass } map { uc } @verbs; + }; + + if ( ($verbs[0] || '') eq 'SELECT' && @verbs > 1 ) { + # False-positive verbs after SELECT + my $union = grep { $_ eq 'UNION' } @verbs; + @verbs = $union ? qw(SELECT UNION) : qw(SELECT); + } + + my %seen; + $verbs = join(q{ }, grep { !$seen{$_}++ } @verbs); + } + + if ( $verbs && $verbs =~ m/^SHOW/ ) { + my %alias_for = qw( + SCHEMA DATABASE + KEYS INDEX + INDEXES INDEX + ); + map { $verbs =~ s/$_/$alias_for{$_}/ } keys %alias_for; + $query = $verbs; + } + else { + my @tables; + $query =~ s/ (?:LOW_PRIORITY|IGNORE|STRAIGHT_JOIN)//ig; + if ( $query =~ /^\s*LOCK\s+TABLES/i ) { + $query =~ s/^(\s*LOCK\s+TABLES\s+)//i; + $query =~ s/\s+(?:READ|WRITE|LOCAL)+\s*//gi; + $query = "FROM $query"; + } + $query =~ s/\\["']//g; # quoted strings + $query =~ s/".*?"/?/sg; # quoted strings + $query =~ s/'.*?'/?/sg; # quoted strings + + foreach my $tbls ( $query =~ m/$tbl_regex/gio ) { + next if $tbls =~ m/\ASELECT\b/i; + foreach my $tbl ( split(',', $tbls) ) { + $tbl =~ s/\s*($tbl_ident)(\s+.*)?/$1/gio; + if ( $tbl !~ m/[a-zA-Z]/ ) { + # Skip suspicious table name + next; + } + push @tables, $tbl; + } + } + + # If we have a bunch of tables like db1.tbl1 db1.tbl2, convert to + # db1.tbl1 -.tbl2 etc. Also remove repeated tables, and strip `quotes`. + $query = $verbs; + my $prev = ''; + foreach my $t ( @tables, $table ) { + next unless $t; + $t =~ s/`//g; + next if $t eq $prev; + my ($prefix, undef) = split(/\./, $prev); + $prev = $t; + if ( $prefix ) { + $t =~ s/^$prefix\./-./; + } + $query .= " " . $t; + } + } + + # die $orig_query if $query eq 'LOCK lock'; + return $query; +} + +# secs_to_time {{{3 +sub secs_to_time { + my ( $secs, $fmt ) = @_; + $secs ||= 0; + + # If the inbound value has a decimal point, then format the seconds with milliseconds. + my $hires = $secs =~ m/\./ ? '%06.3f' : '%02d'; + + if ( !$secs ) { + return sprintf("00:$hires", $secs); + } + + # Decide what format to use, if not given + $fmt ||= $secs >= 86_400 ? 'd' + : $secs >= 3_600 ? 'h' + : 'm'; + + return + $fmt eq 'd' ? sprintf( + "%d+%02d:%02d:$hires", + int($secs / 86_400), + int(($secs % 86_400) / 3_600), + int(($secs % 3_600) / 60), + $secs % 60 + ($secs - int($secs))) + : $fmt eq 'h' ? sprintf( + "%02d:%02d:$hires", + int(($secs % 86_400) / 3_600), + int(($secs % 3_600) / 60), + $secs % 60 + ($secs - int($secs))) + : sprintf( + "%02d:$hires", + int(($secs % 3_600) / 60), + $secs % 60 + ($secs - int($secs))); +} + +# dulint_to_int {{{3 +# Takes a number that InnoDB formats as two ulint integers, like transaction IDs +# and such, and turns it into a single integer +sub dulint_to_int { + my $num = shift; + return 0 unless $num; + my ( $high, $low ) = $num =~ m/^(\d+) (\d+)$/; + return $low unless $high; + return $low + ( $high * $MAX_ULONG ); +} + +# create_statusbar {{{3 +sub create_statusbar { + my $mode = $config{mode}->{val}; + my @cxns = sort { $a cmp $b } get_connections(); + + my $modeline = ( $config{readonly}->{val} ? '[RO] ' : '' ) + . $modes{$mode}->{hdr} . " (? for help)"; + my $mode_width = length($modeline); + my $remaining_width = $this_term_size[0] - $mode_width - 1; + my $result; + + # The thingie in top-right that says what we're monitoring. + my $cxn = ''; + + if ( 1 == @cxns && $dbhs{$cxns[0]} && $dbhs{$cxns[0]}->{dbh} ) { + $cxn = $dbhs{$cxns[0]}->{dbh}->{mariadb_serverinfo} || ''; + } + else { + if ( $modes{$mode}->{server_group} ) { + $cxn = "Servers: " . $modes{$mode}->{server_group}; + my $err_count = grep { $dbhs{$_} && $dbhs{$_}->{failed} } @cxns; + if ( $err_count ) { + $cxn .= "(" . ( scalar(@cxns) - $err_count ) . "/" . scalar(@cxns) . ")"; + } + } + else { + $cxn = join(' ', map { ($dbhs{$_}->{failed} ? '!' : '') . $_ } + grep { $dbhs{$_} } @cxns); + } + } + + if ( 1 == @cxns ) { + get_driver_status(@cxns); + my $vars = $vars{$cxns[0]}->{$clock}; + my $inc = inc(0, $cxns[0]); + + # Format server uptime human-readably, calculate QPS... + my $uptime = fuzzy_time( $vars->{Uptime_hires} ); + my $qps = ($inc->{Questions}||0) / ($inc->{Uptime_hires}||1); + my $ibinfo = ''; + + if ( exists $vars->{IB_last_secs} ) { + $ibinfo .= "InnoDB $vars->{IB_last_secs}s "; + if ( $vars->{IB_got_all} ) { + if ( ($mode eq 'T' || $mode eq 'W') + && $vars->{IB_tx_is_truncated} ) { + $ibinfo .= ':^|'; + } + else { + $ibinfo .= ':-)'; + } + } + else { + $ibinfo .= ':-('; + } + } + $result = sprintf( + "%-${mode_width}s %${remaining_width}s", + $modeline, + join(', ', grep { $_ } ( + $cxns[0], + $uptime, + $ibinfo, + shorten($qps) . " QPS", + ($vars->{Threads} || 0) . "/" . ($vars->{Threads_running} || 0) . "/" . ($vars->{Threads_cached} || 0) . " con/run/cac thds", + $cxn))); + } + else { + $result = sprintf( + "%-${mode_width}s %${remaining_width}s", + $modeline, + $cxn); + } + + return $config{color}->{val} ? [ $result, 'bold reverse' ] : $result; +} + +# Database connections {{{3 +sub add_new_dsn { + my ( $name, $dsn, $dl_table, $have_user, $user, $have_pass, $pass, $savepass ) = @_; + + if ( defined $name ) { + $name =~ s/[\s:;]//g; + } + + if ( !$name ) { + print word_wrap("Choose a name for the connection. It cannot contain " + . "whitespace, colons or semicolons."), "\n\n"; + do { + $name = prompt("Enter a name"); + $name =~ s/[\s:;]//g; + } until ( $name ); + } + + if ( !$dsn ) { + do { + $clear_screen_sub->(); + print "Typical DSN strings look like\n DBI:MariaDB:;host=hostname;port=port\n" + . "The db and port are optional and can usually be omitted.\n" + . "If you specify 'mariadb_read_default_group=mysql' many options can be read\n" + . "from your mysql options files (~/.my.cnf, /etc/my.cnf).\n\n"; + $dsn = prompt("Enter a DSN string", undef, "DBI:MariaDB:;mariadb_read_default_group=mysql;host=$name"); + } until ( $dsn ); + } + if ( !$dl_table ) { + $clear_screen_sub->(); + my $dl_table = prompt("Optional: enter a table (must not exist) to use when resetting InnoDB deadlock information", + undef, 'test.innotop_dl'); + } + + $connections{$name} = { + dsn => $dsn, + dl_table => $dl_table, + have_user => $have_user, + user => $user, + have_pass => $have_pass, + pass => $pass, + savepass => $savepass + }; +} + +sub add_new_server_group { + my ( $name ) = @_; + + if ( defined $name ) { + $name =~ s/[\s:;]//g; + } + + if ( !$name ) { + print word_wrap("Choose a name for the group. It cannot contain " + . "whitespace, colons or semicolons."), "\n\n"; + do { + $name = prompt("Enter a name"); + $name =~ s/[\s:;]//g; + } until ( $name ); + } + + my @cxns; + do { + $clear_screen_sub->(); + @cxns = select_cxn("Choose servers for $name", keys %connections); + } until ( @cxns ); + + $server_groups{$name} = \@cxns; + return $name; +} + +sub get_var_set { + my ( $name ) = @_; + while ( !$name || !exists($var_sets{$config{$name}->{val}}) ) { + $name = choose_var_set($name); + } + return $var_sets{$config{$name}->{val}}->{text}; +} + +sub add_new_var_set { + my ( $name ) = @_; + + if ( defined $name ) { + $name =~ s/\W//g; + } + + if ( !$name ) { + do { + $name = prompt("Enter a name"); + $name =~ s/\W//g; + } until ( $name ); + } + + my $variables; + do { + $clear_screen_sub->(); + $variables = prompt("Enter variables for $name", undef ); + } until ( $variables ); + + $var_sets{$name} = { text => $variables, user => 1 }; +} + +sub next_server { + my $mode = $config{mode}->{val}; + my @cxns = sort keys %connections; + my ($cur) = get_connections($mode); + $cur ||= $cxns[0]; + my $pos = grep { $_ lt $cur } @cxns; + my $newpos = ($pos + 1) % @cxns; + $modes{$mode}->{server_group} = ''; + $modes{$mode}->{connections} = [ $cxns[$newpos] ]; + $clear_screen_sub->(); +} + +sub next_server_group { + my $mode = shift || $config{mode}->{val}; + my @grps = sort keys %server_groups; + my $curr = $modes{$mode}->{server_group}; + + return unless @grps; + + if ( $curr ) { + # Find the current group's position. + my $pos = 0; + while ( $curr ne $grps[$pos] ) { + $pos++; + } + $modes{$mode}->{server_group} = $grps[ ($pos + 1) % @grps ]; + } + else { + $modes{$mode}->{server_group} = $grps[0]; + } +} + +# Get a list of connection names used in this mode. +sub get_connections { + if ( $file ) { + return qw(file); + } + my $mode = shift || $config{mode}->{val}; + my @connections = $modes{$mode}->{server_group} + ? @{$server_groups{$modes{$mode}->{server_group}}} + : @{$modes{$mode}->{connections}}; + if ( $modes{$mode}->{one_connection} ) { + @connections = @connections ? $connections[0] : (); + } + # If the connections are the same as a server group, we set the mode's + # group to that group. + if ( ! $modes{$mode}->{server_group} ) { + my $maybe_group = join(',', sort @connections); + foreach my $g ( keys %server_groups ) { + my $group_conns = join(',', sort @{$server_groups{$g}}); + if ( $maybe_group eq $group_conns ) { + $modes{$mode}->{server_group} = $g; + last; + } + } + } + return unique(@connections); +} + +# Get a list of tables used in this mode. If innotop is running non-interactively, just use the first. +sub get_visible_tables { + my $mode = shift || $config{mode}->{val}; + my @tbls = @{$modes{$mode}->{visible_tables}}; + if ( $opts{n} ) { + return $tbls[0]; + } + else { + return @tbls; + } +} + +# Choose from among available connections or server groups. +# If the mode has a server set in use, prefers that instead. +sub choose_connections { + $clear_screen_sub->(); + my $mode = $config{mode}->{val}; + my $meta = { map { $_ => $connections{$_}->{dsn} } keys %connections }; + foreach my $group ( keys %server_groups ) { + $meta->{"#$group"} = join(' ', @{$server_groups{$group}}); + } + + my $choices = prompt_list("Choose connections or a group for $mode mode", + undef, sub { return keys %$meta }, $meta); + + my @choices = unique(grep { $_ } $choices =~ m/(\S+)/g); + if ( @choices ) { + if ( $choices[0] =~ s/^#// && exists $server_groups{$choices[0]} ) { + $modes{$mode}->{server_group} = $choices[0]; + } + else { + $modes{$mode}->{connections} = [ grep { exists $connections{$_} } @choices ]; + } + } +} + +# Accepts a DB connection name and the name of a prepared query (e.g. status, kill). +# Also a list of params for the prepared query. This allows not storing prepared +# statements globally. Returns a $sth that's been executed. +# ERROR-HANDLING SEMANTICS: if the statement throws an error, propagate, but if the +# connection has gone away or can't connect, DO NOT. Just return undef. +sub do_stmt { + my ( $cxn, $stmt_name, @args ) = @_; + + return undef if $file; + + # Test if the cxn should not even be tried + return undef if $dbhs{$cxn} + && $dbhs{$cxn}->{failed} + && ( !$dbhs{$cxn}->{dbh} || !$dbhs{$cxn}->{dbh}->{Active} || $dbhs{$cxn}->{mode} eq $config{mode}->{val} ); + + my $sth; + my $retries = 1; + my $success = 0; + TRY: + while ( $retries-- >= 0 && !$success ) { + + eval { + my $dbh = connect_to_db($cxn); + + # If the prepared query doesn't exist, make it. + if ( !exists $dbhs{$cxn}->{stmts}->{$stmt_name} ) { + $dbhs{$cxn}->{stmts}->{$stmt_name} = $stmt_maker_for{$stmt_name}->($dbh); + } + + $sth = $dbhs{$cxn}->{stmts}->{$stmt_name}; + if ( $sth ) { + $sth->execute(@args); + } + $success = 1; + }; + if ( $EVAL_ERROR ) { + if ( $EVAL_ERROR =~ m/$nonfatal_errs/ ) { + handle_cxn_error($cxn, $EVAL_ERROR); + } + else { + die "$cxn $stmt_name: $EVAL_ERROR"; + } + if ( $retries < 0 ) { + $sth = undef; + } + } + } + + if ( $sth && $sth->{NUM_OF_FIELDS} ) { + sleep($stmt_sleep_time_for{$stmt_name}) if $stmt_sleep_time_for{$stmt_name}; + return $sth; + } +} + +# Marks a connection as failed. When we sleep between redraws, we try to +# reopen. +sub handle_cxn_error { + my ( $cxn, $err ) = @_; + my $meta = $dbhs{$cxn}; + $meta->{failed} = 1; + + # This is used so errors that have to do with permissions needed by the current + # mode will get displayed as long as we're in this mode, but get ignored if the + # mode changes. + $meta->{mode} = $config{mode}->{val}; + + # Strip garbage from the error text if possible. + $err =~ s/\s+/ /g; + if ( $err =~ m/failed: (.*?) at \S*innotop line/ ) { + $err = $1; + } + + $meta->{last_err} = $err; + if ( $config{show_cxn_errors}->{val} ) { + print STDERR "DB error: $cxn $err" if $config{debug}->{val}; + } +} + +# Accepts a DB connection name and a (string) query. Returns a $sth that's been +# executed. +sub do_query { + my ( $cxn, $query ) = @_; + + return undef if $file; + + # Test if the cxn should not even be tried + return undef if $dbhs{$cxn} + && $dbhs{$cxn}->{failed} + && ( !$dbhs{$cxn}->{dbh} || !$dbhs{$cxn}->{dbh}->{Active} || $dbhs{$cxn}->{mode} eq $config{mode}->{val} ); + + my $sth; + my $retries = 1; + my $success = 0; + TRY: + while ( $retries-- >= 0 && !$success ) { + + eval { + my $dbh = connect_to_db($cxn); + + $sth = $dbh->prepare($query); + $sth->execute(); + $success = 1; + }; + if ( $EVAL_ERROR ) { + if ( $EVAL_ERROR =~ m/$nonfatal_errs/ ) { + handle_cxn_error($cxn, $EVAL_ERROR); + } + else { + die $EVAL_ERROR; + } + if ( $retries < 0 ) { + $sth = undef; + } + } + } + + return $sth; +} + +sub get_uptime { + my ( $cxn ) = @_; + $dbhs{$cxn}->{start_time} ||= time(); + # Avoid dividing by zero + return (time() - $dbhs{$cxn}->{start_time}) || .001; +} + +sub connect_to_db { + my ( $cxn ) = @_; + + $dbhs{$cxn} ||= { + stmts => {}, # bucket for prepared statements. + start_time => 0, + dbh => undef, + }; + my $href = $dbhs{$cxn}; + + if ( !$href->{dbh} || ref($href->{dbh}) !~ m/DBI/ || !$href->{dbh}->ping ) { + my $dbh = get_new_db_connection($cxn); + @{$href}{qw(dbh failed start_time stmts)} = ($dbh, 0, 0, {}); + + # Derive and store the server's start time in hi-res + my $uptime = $dbh->selectrow_hashref("show status like 'Uptime'")->{value}; + $href->{start_time} = time() - $uptime; + + # Set timeouts so an unused connection stays alive. + # For example, a connection might be used in Q mode but idle in T mode. + if ( version_ge($dbh, '4.0.3')) { + my $timeout = $config{cxn_timeout}->{val}; + $dbh->do("set session wait_timeout=$timeout, interactive_timeout=$timeout"); + } + } + return $href->{dbh}; +} + +# Compares versions like 5.0.27 and 4.1.15-standard-log +sub version_ge { + my ( $dbh, $target ) = @_; + my $version = sprintf('%03d%03d%03d', $dbh->{mariadb_serverinfo} =~ m/^(\d+).(\d+).(\d+)/g); + return $version ge sprintf('%03d%03d%03d', $target =~ m/(\d+)/g); +} + +# Extracts status values that can be gleaned from the DBD driver without doing a whole query. +sub get_driver_status { + my @cxns = @_; + if ( !$info_gotten{driver_status}++ ) { + foreach my $cxn ( @cxns ) { + next unless $dbhs{$cxn} && $dbhs{$cxn}->{dbh} && $dbhs{$cxn}->{dbh}->{Active}; + $vars{$cxn}->{$clock} ||= {}; + my $vars = $vars{$cxn}->{$clock}; + my %res = map { $_ =~ s/ +/_/g; $_ } $dbhs{$cxn}->{dbh}->{mariadb_stat} =~ m/(\w[^:]+): ([\d\.]+)/g; + map { $vars->{$_} ||= $res{$_} } keys %res; + $vars->{Uptime_hires} ||= get_uptime($cxn); + $vars->{cxn} = $cxn; + } + } +} + +sub get_new_db_connection { + my ( $connection, $destroy ) = @_; + if ( $file ) { + die "You can't connect to a MariaDB server while monitoring a file. This is probably a bug."; + } + + my $dsn = $connections{$connection} + or die "No connection named '$connection' is defined in your configuration"; + + # don't ask for a username if mariadb_read_default_group=client is in the DSN + if ( !defined $dsn->{have_user} and $dsn->{dsn} !~ /mariadb_read_default_group=client/ ) { + my $answer = prompt("Do you want to specify a username for $connection?", undef, 'n'); + $dsn->{have_user} = $answer && $answer =~ m/1|y/i; + } + + # don't ask for a password if mariadb_read_default_group=client is in the DSN + if ( !defined $dsn->{have_pass} and $dsn->{dsn} !~ /mariadb_read_default_group=client/ ) { + my $answer = prompt("Do you want to specify a password for $connection?", undef, 'n'); + $dsn->{have_pass} = $answer && $answer =~ m/1|y/i; + } + + if ( !$dsn->{user} && $dsn->{have_user} ) { + my $user = $ENV{USERNAME} || $ENV{USER} || getlogin() || getpwuid($REAL_USER_ID) || undef; + $dsn->{user} = prompt("Enter username for $connection", undef, $user); + } + + if ( !defined $dsn->{user} ) { + $dsn->{user} = ''; + } + + if ( !$dsn->{pass} && !$dsn->{savepass} && $dsn->{have_pass} ) { + $dsn->{pass} = prompt_noecho("Enter password for '$dsn->{user}' on $connection"); + print "\n"; + if ( !defined($dsn->{savepass}) ) { + my $answer = prompt("Save password in plain text in the config file?", undef, 'y'); + $dsn->{savepass} = $answer && $answer =~ m/1|y/i; + } + } + + my $dbh = DBI->connect( + $dsn->{dsn}, $dsn->{user}, $dsn->{pass}, + { RaiseError => 1, PrintError => 0, AutoCommit => 1 }); + $dbh->{InactiveDestroy} = 1 unless $destroy; # Can't be set in $db_options + $dbh->{FetchHashKeyName} = 'NAME_lc'; # Lowercases all column names for fetchrow_hashref + return $dbh; +} + +sub get_cxn_errors { + my @cxns = @_; + return () unless $config{show_cxn_errors_in_tbl}->{val}; + return + map { [ $_ . ': ' . $dbhs{$_}->{last_err}, 'red' ] } + grep { $dbhs{$_} && $dbhs{$_}->{failed} && $dbhs{$_}->{mode} eq $config{mode}->{val} } + @cxns; +} + +# Setup and tear-down functions {{{2 + +# Takes a string and turns it into a hashref you can apply to %tbl_meta tables. The string +# can be in the form 'foo, bar, foo/bar, foo as bar' much like a SQL SELECT statement. +sub compile_select_stmt { + my ($str) = @_; + my @exps = $str =~ m/\s*([^,]+(?i:\s+as\s+[^,\s]+)?)\s*(?=,|$)/g; + my %cols; + my @visible; + foreach my $exp ( @exps ) { + my ( $text, $colname ); + if ( $exp =~ m/as\s+(\w+)\s*/ ) { + $colname = $1; + $exp =~ s/as\s+(\w+)\s*//; + $text = $exp; + } + else { + $text = $colname = $exp; + } + my ($func, $err) = compile_expr($text); + $cols{$colname} = { + src => $text, + hdr => $colname, + num => 0, + func => $func, + }; + push @visible, $colname; + } + return (\%cols, \@visible); +} + +# compile_filter {{{3 +sub compile_filter { + my ( $text ) = @_; + my ( $sub, $err ); + eval "\$sub = sub { my \$set = shift; $text }"; + if ( $EVAL_ERROR ) { + $EVAL_ERROR =~ s/at \(eval.*$//; + $sub = sub { return $EVAL_ERROR }; + $err = $EVAL_ERROR; + } + return ( $sub, $err ); +} + +# compile_expr {{{3 +sub compile_expr { + my ( $expr ) = @_; + # Leave built-in functions alone so they get called as Perl functions, unless + # they are the only word in $expr, in which case treat them as hash keys. + if ( $expr =~ m/\W/ ) { + $expr =~ s/(?<!\{|\$)\b([A-Za-z]\w{2,})\b/is_func($1) ? $1 : "\$set->{$1}"/eg; + } + else { + $expr = "\$set->{$expr}"; + } + my ( $sub, $err ); + my $quoted = quotemeta($expr); + eval qq{ + \$sub = sub { + my (\$set, \$cur, \$pre) = \@_; + my \$val = eval { $expr }; + if ( \$EVAL_ERROR && \$config{debug}->{val} ) { + \$EVAL_ERROR =~ s/ at \\(eval.*//s; + die "\$EVAL_ERROR in expression $quoted"; + } + return \$val; + } + }; + if ( $EVAL_ERROR ) { + if ( $config{debug}->{val} ) { + die $EVAL_ERROR; + } + $EVAL_ERROR =~ s/ at \(eval.*$//; + $sub = sub { return $EVAL_ERROR }; + $err = $EVAL_ERROR; + } + return ( $sub, $err ); +} + +# finish {{{3 +# This is a subroutine because it's called from a key to quit the program. +sub finish { + save_config(); + foreach my $cxn ( values %dbhs ) { + eval { + foreach my $sth ( values %{$cxn->{stmts}} ) { + $sth->finish; + } + $cxn->{dbh}->disconnect; + }; + # Ignore eval errors, we just don't care + } + ReadMode('normal') unless $opts{n}; + print "\n"; + exit(0); +} + +# core_dump {{{3 +sub core_dump { + my $msg = shift; + if ($config{debugfile}->{val} && $config{debug}->{val}) { + eval { + open my $file, '>>', $config{debugfile}->{val}; + if ( %vars ) { + print $file "Current variables:\n" . Dumper(\%vars); + } + close $file; + }; + } + print $msg; +} + +# migrate_config {{{3 +sub migrate_config { + + my ($old_filename, $new_filename) = @_; + + # don't proceed if old file doesn't exist + if ( ! -f $old_filename ) { + die "Error migrating '$old_filename': file doesn't exist.\n"; + } + # don't migrate files if new file exists + elsif ( -f $new_filename ) { + die "Error migrating '$old_filename' to '$new_filename': new file already exists.\n"; + } + # if migrating from one file to another in the same directory, just rename them + if (dirname($old_filename) eq dirname($new_filename)) { + rename($old_filename, $new_filename) + or die "Can't rename '$old_filename' to '$new_filename': $OS_ERROR"; + } + # otherwise, move the existing conf file to a temp file, make the necessary directory structure, + # and move the temp conf file to its new home + else { + my $tmp = File::Temp->new( TEMPLATE => 'innotopXXXXX', DIR => $homepath, SUFFIX => '.conf'); + my $tmp_filename = $tmp->filename; + my $dirname = dirname($new_filename); + rename($old_filename, $tmp_filename) + or die "Can't rename '$old_filename' to '$tmp_filename': $OS_ERROR"; + mkdir($dirname) or die "Can't create directory '$dirname': $OS_ERROR"; + mkdir("$dirname/plugins") or die "Can't create directory '$dirname/plugins': $OS_ERROR"; + rename($tmp_filename, $new_filename) + or die "Can't rename '$tmp_filename' to '$new_filename': $OS_ERROR"; + } +} + +# load_config {{{3 +sub load_config { + + my ($old_filename, $answer); + + if ( $opts{u} or $opts{p} or $opts{h} or $opts{P} or $opts{S} ) { + my @params = $dsn_parser->get_cxn_params(\%opts); # dsn=$params[0] + add_new_dsn($opts{h} || 'localhost', $params[0], 'test.innotop_dl', + $opts{u} ? 1 : 0, $opts{u}, $opts{p} ? 1 : 0, $opts{p}); + } + if ($opts{c}) { + $conf_file = $opts{c}; + } + # If we don't have a new config file but we do have an old one, + # innotop got upgraded and this is an old config file. Convert it, but + # don't overwrite something existing. + elsif ( ! -f $default_home_conf && ( -f "$homepath/.innotop" or -f "$homepath/.innotop/innotop.ini" ) ) { + $conf_file = $default_home_conf; + if ( -f "$homepath/.innotop") { + $old_filename = "$homepath/.innotop"; + } + elsif ( -f "$homepath/.innotop/innotop.ini" ) { + $old_filename = "$homepath/.innotop/innotop.ini"; + } + $answer = pause("Innotop's default config location has moved to '$conf_file'. Move old config file '$old_filename' there now? y/n"); + if ( lc $answer eq 'y' ) { + migrate_config($old_filename, $conf_file); + } + else { + print "\nInnotop will now exit so you can fix the config file.\n"; + exit(0); + } + } + elsif ( -f $default_home_conf ) { + $conf_file = $default_home_conf; + } + elsif ( -f $default_central_conf and not $opts{s} ) { + $conf_file = $default_central_conf; + } + else { + # If no config file was loaded, set readonly to 0 if the user wants to + # write a config + $config{readonly}->{val} = 0 if $opts{w}; + # If no connections have been defined, connect to a MySQL database + # on localhost using mariadb_read_default_group=client + if (!%connections) { + add_new_dsn('localhost', + 'DBI:MariaDB:;host=localhost;mariadb_read_default_group=client', + 'test.innotop_dl'); + } + } + + if ( -f "$conf_file" ) { + open my $file, "<", $conf_file or die("Can't open '$conf_file': $OS_ERROR"); + + # Check config file version. Just ignore if either innotop or the file has + # garbage in the version number. + if ( defined(my $line = <$file>) && $VERSION =~ m/\d/ ) { + chomp $line; + if ( my ($maj, $min, $rev) = $line =~ m/^version=(\d+)\.(\d+)(?:\.(\d+))?$/ ) { + $rev ||= 0; + my $cfg_ver = sprintf('%03d-%03d-%03d', $maj, $min, $rev); + ( $maj, $min, $rev ) = $VERSION =~ m/^(\d+)\.(\d+)(?:\.(\d+))?$/; + $rev ||= 0; + my $innotop_ver = sprintf('%03d-%03d-%03d', $maj, $min, $rev); + + if ( $cfg_ver gt $innotop_ver ) { + pause("The config file is for a newer version of innotop and may not be read correctly."); + } + else { + my @ver_history = @config_versions; + while ( my ($start, $end) = splice(@ver_history, 0, 2) ) { + # If the config file is between the endpoints and innotop is greater than + # the endpoint, innotop has a newer config file format than the file. + if ( $cfg_ver ge $start && $cfg_ver lt $end && $innotop_ver ge $end ) { + my $msg = "innotop's config file format has changed. Overwrite $conf_file? y or n"; + if ( pause($msg) eq 'n' ) { + $config{readonly}->{val} = 1; + print "\ninnotop will not save any configuration changes you make."; + pause(); + print "\n"; + } + close $file; + return; + } + } + } + } + } + + while ( my $line = <$file> ) { + chomp $line; + next unless $line =~ m/^\[([a-z_]+)\]$/; + if ( exists $config_file_sections{$1} ) { + $config_file_sections{$1}->{reader}->($file); + } + else { + warn "Unknown config file section '$1'"; + } + } + close $file or die("Can't close $conf_file: $OS_ERROR"); + } + +} + +# Do some post-processing on %tbl_meta: compile src properties into func etc. +sub post_process_tbl_meta { + foreach my $table ( values %tbl_meta ) { + foreach my $col_name ( keys %{$table->{cols}} ) { + my $col_def = $table->{cols}->{$col_name}; + my ( $sub, $err ) = compile_expr($col_def->{src}); + $col_def->{func} = $sub; + } + } +} + +# load_config_plugins {{{3 +sub load_config_plugins { + my ( $file ) = @_; + + # First, find a list of all plugins that exist on disk, and get information about them. + my $dir = $config{plugin_dir}->{val}; + foreach my $p_file ( <$dir/*.pm> ) { + my ($package, $desc); + eval { + open my $p_in, "<", $p_file or die $OS_ERROR; + while ( my $line = <$p_in> ) { + chomp $line; + if ( $line =~ m/^package\s+(.*?);/ ) { + $package = $1; + } + elsif ( $line =~ m/^# description: (.*)/ ) { + $desc = $1; + } + last if $package && $desc; + } + close $p_in; + }; + if ( $package ) { + $plugins{$package} = { + file => $p_file, + desc => $desc, + class => $package, + active => 0, + }; + if ( $config{debug}->{val} && $EVAL_ERROR ) { + die $EVAL_ERROR; + } + } + } + + # Now read which ones the user has activated. Each line simply represents an active plugin. + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + next unless $line && $plugins{$line}; + + my $obj; + eval { + require $plugins{$line}->{file}; + $obj = $line->new(%pluggable_vars); + foreach my $event ( $obj->register_for_events() ) { + my $queue = $event_listener_for{$event}; + if ( $queue ) { + push @$queue, $obj; + } + } + }; + if ( $config{debug}->{val} && $EVAL_ERROR ) { + die $EVAL_ERROR; + } + if ( $obj ) { + $plugins{$line}->{active} = 1; + $plugins{$line}->{object} = $obj; + } + } +} + +# save_config_plugins {{{3 +sub save_config_plugins { + my $file = shift; + foreach my $class ( sort keys %plugins ) { + next unless $plugins{$class}->{active}; + print $file "$class\n"; + } +} + +# load_config_active_server_groups {{{3 +sub load_config_active_server_groups { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $mode, $group ) = $line =~ m/^(.*?)=(.*)$/; + next unless $mode && $group + && exists $modes{$mode} && exists $server_groups{$group}; + $modes{$mode}->{server_group} = $group; + } +} + +# save_config_active_server_groups {{{3 +sub save_config_active_server_groups { + my $file = shift; + foreach my $mode ( sort keys %modes ) { + print $file "$mode=$modes{$mode}->{server_group}\n"; + } +} + +# load_config_server_groups {{{3 +sub load_config_server_groups { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $name, $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $name && $rest; + my @vars = unique(grep { $_ && exists $connections{$_} } split(/\s+/, $rest)); + next unless @vars; + $server_groups{$name} = \@vars; + } +} + +# save_config_server_groups {{{3 +sub save_config_server_groups { + my $file = shift; + foreach my $set ( sort keys %server_groups ) { + print $file "$set=", join(' ', @{$server_groups{$set}}), "\n"; + } +} + +# load_config_varsets {{{3 +sub load_config_varsets { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $name, $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $name && $rest; + $var_sets{$name} = { + text => $rest, + user => 1, + }; + } +} + +# save_config_varsets {{{3 +sub save_config_varsets { + my $file = shift; + foreach my $varset ( sort keys %var_sets ) { + next unless $var_sets{$varset}->{user}; + print $file "$varset=$var_sets{$varset}->{text}\n"; + } +} + +# load_config_group_by {{{3 +sub load_config_group_by { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $tbl , $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $tbl && exists $tbl_meta{$tbl}; + my @parts = unique(grep { exists($tbl_meta{$tbl}->{cols}->{$_}) } split(/\s+/, $rest)); + $tbl_meta{$tbl}->{group_by} = [ @parts ]; + $tbl_meta{$tbl}->{cust}->{group_by} = 1; + } +} + +# save_config_group_by {{{3 +sub save_config_group_by { + my $file = shift; + foreach my $tbl ( sort keys %tbl_meta ) { + next if $tbl_meta{$tbl}->{temp}; + next unless $tbl_meta{$tbl}->{cust}->{group_by}; + my $aref = $tbl_meta{$tbl}->{group_by}; + print $file "$tbl=", join(' ', @$aref), "\n"; + } +} + +# load_config_filters {{{3 +sub load_config_filters { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $key, $rest ) = $line =~ m/^(.+?)=(.*)$/; + next unless $key && $rest; + + my %parts = $rest =~ m/(\w+)='((?:(?!(?<!\\)').)*)'/g; # Properties are single-quoted + next unless $parts{text} && $parts{tbls}; + + foreach my $prop ( keys %parts ) { + # Un-escape escaping + $parts{$prop} =~ s/\\\\/\\/g; + $parts{$prop} =~ s/\\'/'/g; + } + + my ( $sub, $err ) = compile_filter($parts{text}); + my @tbls = unique(split(/\s+/, $parts{tbls})); + @tbls = grep { exists $tbl_meta{$_} } @tbls; + $filters{$key} = { + func => $sub, + text => $parts{text}, + user => 1, + name => $key, + note => 'User-defined filter', + tbls => \@tbls, + } + } +} + +# save_config_filters {{{3 +sub save_config_filters { + my $file = shift; + foreach my $key ( sort keys %filters ) { + next if !$filters{$key}->{user} || $filters{$key}->{quick}; + my $text = $filters{$key}->{text}; + $text =~ s/([\\'])/\\$1/g; + my $tbls = join(" ", @{$filters{$key}->{tbls}}); + print $file "$key=text='$text' tbls='$tbls'\n"; + } +} + +# load_config_visible_tables {{{3 +sub load_config_visible_tables { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $mode, $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $mode && exists $modes{$mode}; + $modes{$mode}->{visible_tables} = + [ unique(grep { $_ && exists $tbl_meta{$_} } split(/\s+/, $rest)) ]; + $modes{$mode}->{cust}->{visible_tables} = 1; + } +} + +# save_config_visible_tables {{{3 +sub save_config_visible_tables { + my $file = shift; + foreach my $mode ( sort keys %modes ) { + next unless $modes{$mode}->{cust}->{visible_tables}; + my $tables = $modes{$mode}->{visible_tables}; + print $file "$mode=", join(' ', @$tables), "\n"; + } +} + +# load_config_sort_cols {{{3 +sub load_config_sort_cols { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $key , $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $key && exists $tbl_meta{$key}; + $tbl_meta{$key}->{sort_cols} = $rest; + $tbl_meta{$key}->{cust}->{sort_cols} = 1; + $tbl_meta{$key}->{sort_func} = make_sort_func($tbl_meta{$key}); + } +} + +# save_config_sort_cols {{{3 +sub save_config_sort_cols { + my $file = shift; + foreach my $tbl ( sort keys %tbl_meta ) { + next unless $tbl_meta{$tbl}->{cust}->{sort_cols}; + my $col = $tbl_meta{$tbl}->{sort_cols}; + print $file "$tbl=$col\n"; + } +} + +# load_config_active_filters {{{3 +sub load_config_active_filters { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $tbl , $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $tbl && exists $tbl_meta{$tbl}; + my @parts = unique(grep { exists($filters{$_}) } split(/\s+/, $rest)); + @parts = grep { grep { $tbl eq $_ } @{$filters{$_}->{tbls}} } @parts; + $tbl_meta{$tbl}->{filters} = [ @parts ]; + $tbl_meta{$tbl}->{cust}->{filters} = 1; + } +} + +# save_config_active_filters {{{3 +sub save_config_active_filters { + my $file = shift; + foreach my $tbl ( sort keys %tbl_meta ) { + next if $tbl_meta{$tbl}->{temp}; + next unless $tbl_meta{$tbl}->{cust}->{filters}; + my $aref = $tbl_meta{$tbl}->{filters}; + print $file "$tbl=", join(' ', @$aref), "\n"; + } +} + +# load_config_active_columns {{{3 +sub load_config_active_columns { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $key , $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $key && exists $tbl_meta{$key}; + my @parts = grep { exists($tbl_meta{$key}->{cols}->{$_}) } unique split(/ /, $rest); + $tbl_meta{$key}->{visible} = [ @parts ]; + $tbl_meta{$key}->{cust}->{visible} = 1; + } +} + +# save_config_active_columns {{{3 +sub save_config_active_columns { + my $file = shift; + foreach my $tbl ( sort keys %tbl_meta ) { + next unless $tbl_meta{$tbl}->{cust}->{visible}; + my $aref = $tbl_meta{$tbl}->{visible}; + print $file "$tbl=", join(' ', @$aref), "\n"; + } +} + +# save_config_tbl_meta {{{3 +sub save_config_tbl_meta { + my $file = shift; + foreach my $tbl ( sort keys %tbl_meta ) { + foreach my $col ( keys %{$tbl_meta{$tbl}->{cols}} ) { + my $meta = $tbl_meta{$tbl}->{cols}->{$col}; + next unless $meta->{user}; + print $file "$col=", join( + " ", + map { + # Some properties (trans) are arrays, others scalars + my $val = ref($meta->{$_}) ? join(',', @{$meta->{$_}}) : $meta->{$_}; + $val =~ s/([\\'])/\\$1/g; # Escape backslashes and single quotes + "$_='$val'"; # Enclose in single quotes + } + grep { $_ ne 'func' } + keys %$meta + ), "\n"; + } + } +} + +# save_config_config {{{3 +sub save_config_config { + my $file = shift; + foreach my $key ( sort keys %config ) { + eval { + if ( $key ne 'password' || $config{savepass}->{val} ) { + print $file "# $config{$key}->{note}\n" + or die "Cannot print to file: $OS_ERROR"; + my $val = $config{$key}->{val}; + $val = '' unless defined($val); + if ( ref( $val ) eq 'ARRAY' ) { + print $file "$key=" + . join( " ", @$val ) . "\n" + or die "Cannot print to file: $OS_ERROR"; + } + elsif ( ref( $val ) eq 'HASH' ) { + print $file "$key=" + . join( " ", + map { "$_:$val->{$_}" } keys %$val + ) . "\n"; + } + else { + print $file "$key=$val\n"; + } + } + }; + if ( $EVAL_ERROR ) { print "$EVAL_ERROR in $key"; }; + } + +} + +# load_config_config {{{3 +sub load_config_config { + my ( $file ) = @_; + + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $name, $val ) = $line =~ m/^(.+?)=(.*)$/; + next unless defined $name && defined $val; + + # Validate the incoming values... + if ( $name && exists( $config{$name} ) ) { + if ( !$config{$name}->{pat} || $val =~ m/$config{$name}->{pat}/ ) { + $config{$name}->{val} = $val; + $config{$name}->{read} = 1; + } + } + } +} + +# load_config_tbl_meta {{{3 +sub load_config_tbl_meta { + my ( $file ) = @_; + + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + # Each tbl_meta section has all the properties defined in %col_props. + my ( $col , $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $col; + my %parts = $rest =~ m/(\w+)='((?:(?!(?<!\\)').)*)'/g; # Properties are single-quoted + + # Each section read from the config file has one extra property: which table it + # goes in. + my $tbl = $parts{tbl} or die "There's no table for tbl_meta $col"; + my $meta = $tbl_meta{$tbl} or die "There's no table in tbl_meta named $tbl"; + + # The section is user-defined by definition (if that makes sense). + $parts{user} = 1; + + # The column may already exist in the table, in which case this is just a + # customization. + $meta->{cols}->{$col} ||= {}; + + foreach my $prop ( keys %col_props ) { + if ( !defined($parts{$prop}) ) { + # Make it default to whatever's in col_props. + $parts{$prop} = $col_props{$prop}; + } + + # Un-escape escaping + $parts{$prop} =~ s/\\\\/\\/g; + $parts{$prop} =~ s/\\'/'/g; + + if ( ref $col_props{$prop} ) { + if ( $prop eq 'trans' ) { + $meta->{cols}->{$col}->{trans} + = [ unique(grep { exists $trans_funcs{$_} } split(',', $parts{$prop})) ]; + } + else { + $meta->{cols}->{$col}->{$prop} = [ split(',', $parts{$prop}) ]; + } + } + else { + $meta->{cols}->{$col}->{$prop} = $parts{$prop}; + } + } + + } +} + +# save_config {{{3 +sub save_config { + print "\n"; + return if $config{readonly}->{val}; + # return if no config file was loaded and -w wasn't specified + if (not $conf_file) { + if (not $opts{w}) { + return; + } + else { + # if no config was loaded but -w was specified, + # write to $default_home_conf + $conf_file = $default_home_conf; + } + } + elsif ($conf_file and $opts{w}) { + print "Loaded config file on start-up, so ignoring -w (see --help)\n" + } + + my $dirname = dirname($conf_file); + + # if directories don't exist, create them. This could cause errors + # or warnings if a central config doesn't have readonly=1, but being + # flexible requires giving the user enough rope to hang themselves with. + if ( ! -d $dirname ) { + mkdir $dirname + or die "Can't create directory '$dirname': $OS_ERROR"; + } + if ( ! -d "$dirname/plugins" ) { + mkdir "$dirname/plugins" + or warn "Can't create directory '$dirname/plugins': $OS_ERROR\n"; + } + + # Save to a temp file first, so a crash doesn't destroy the main config file + my $tmpfile = File::Temp->new( TEMPLATE => 'innotopXXXXX', DIR => $dirname, SUFFIX => '.conf.tmp'); + open my $file, "+>", $tmpfile + or die("Can't write to $tmpfile: $OS_ERROR"); + print $file "version=$VERSION\n"; + + foreach my $section ( @ordered_config_file_sections ) { + die "No such config file section $section" unless $config_file_sections{$section}; + print $file "\n[$section]\n\n"; + $config_file_sections{$section}->{writer}->($file); + print $file "\n[/$section]\n"; + } + + # Now clobber the main config file with the temp. + close $file or die("Can't close $tmpfile: $OS_ERROR"); + rename($tmpfile, $conf_file) or die("Can't rename $tmpfile to $conf_file: $OS_ERROR"); +} + +# load_config_connections {{{3 +sub load_config_connections { + return if $opts{u} or $opts{p} or $opts{h} or $opts{P}; # don't load connections if DSN or user/pass options used + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $key , $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $key; + my %parts = $rest =~ m/(\S+?)=(\S*)/g; + my %conn = map { $_ => $parts{$_} || '' } @conn_parts; + $connections{$key} = \%conn; + } +} + +# save_config_connections {{{3 +sub save_config_connections { + my $file = shift; + foreach my $conn ( sort keys %connections ) { + my $href = $connections{$conn}; + my @keys = $href->{savepass} ? @conn_parts : grep { $_ ne 'pass' } @conn_parts; + print $file "$conn=", join(' ', map { "$_=$href->{$_}" } grep { defined $href->{$_} } @keys), "\n"; + } +} + +sub load_config_colors { + my ( $file ) = @_; + my %rule_set_for; + + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $tbl, $rule ) = $line =~ m/^(.*?)=(.*)$/; + next unless $tbl && $rule; + next unless exists $tbl_meta{$tbl}; + my %parts = $rule =~ m/(\w+)='((?:(?!(?<!\\)').)*)'/g; # Properties are single-quoted + next unless $parts{col} && exists $tbl_meta{$tbl}->{cols}->{$parts{col}}; + next unless $parts{op} && exists $comp_ops{$parts{op}}; + next unless defined $parts{arg}; + next unless defined $parts{color}; + my @colors = unique(grep { exists $ansicolors{$_} } split(/\W+/, $parts{color})); + next unless @colors; + + # Finally! Enough validation... + $rule_set_for{$tbl} ||= []; + push @{$rule_set_for{$tbl}}, \%parts; + } + + foreach my $tbl ( keys %rule_set_for ) { + $tbl_meta{$tbl}->{colors} = $rule_set_for{$tbl}; + $tbl_meta{$tbl}->{color_func} = make_color_func($tbl_meta{$tbl}); + $tbl_meta{$tbl}->{cust}->{colors} = 1; + } +} + +# save_config_colors {{{3 +sub save_config_colors { + my $file = shift; + foreach my $tbl ( sort keys %tbl_meta ) { + my $meta = $tbl_meta{$tbl}; + next unless $meta->{cust}->{colors}; + foreach my $rule ( @{$meta->{colors}} ) { + print $file "$tbl=", join( + ' ', + map { + my $val = $rule->{$_}; + $val =~ s/([\\'])/\\$1/g; # Escape backslashes and single quotes + "$_='$val'"; # Enclose in single quotes + } + qw(col op arg color) + ), "\n"; + } + } +} + +# load_config_active_connections {{{3 +sub load_config_active_connections { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $key , $rest ) = $line =~ m/^(.*?)=(.*)$/; + next unless $key && exists $modes{$key}; + my @parts = grep { exists $connections{$_} } split(/ /, $rest); + $modes{$key}->{connections} = [ @parts ] if exists $modes{$key}; + } +} + +# save_config_active_connections {{{3 +sub save_config_active_connections { + my $file = shift; + foreach my $mode ( sort keys %modes ) { + my @connections = get_connections($mode); + print $file "$mode=", join(' ', @connections), "\n"; + } +} + +# load_config_stmt_sleep_times {{{3 +sub load_config_stmt_sleep_times { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $key , $val ) = split('=', $line); + next unless $key && defined $val && $val =~ m/$num_regex/; + $stmt_sleep_time_for{$key} = $val; + } +} + +# save_config_stmt_sleep_times {{{3 +sub save_config_stmt_sleep_times { + my $file = shift; + foreach my $key ( sort keys %stmt_sleep_time_for ) { + print $file "$key=$stmt_sleep_time_for{$key}\n"; + } +} + +# load_config_mvs {{{3 +sub load_config_mvs { + my ( $file ) = @_; + while ( my $line = <$file> ) { + chomp $line; + next if $line =~ m/^#/; + last if $line =~ m/^\[/; + + my ( $key , $val ) = split('=', $line); + next unless $key && defined $val && $val =~ m/$num_regex/; + $mvs{$key} = $val; + } +} + +# save_config_mvs {{{3 +sub save_config_mvs { + my $file = shift; + foreach my $key ( sort keys %mvs ) { + print $file "$key=$mvs{$key}\n"; + } +} + +# edit_configuration {{{3 +sub edit_configuration { + my $key = ''; + while ( $key ne 'q' ) { + $clear_screen_sub->(); + my @display_lines = ''; + + if ( $key && $cfg_editor_action{$key} ) { + $cfg_editor_action{$key}->{func}->(); + } + + # Show help + push @display_lines, create_caption('What configuration do you want to edit?', + create_table2( + [ sort keys %cfg_editor_action ], + { map { $_ => $_ } keys %cfg_editor_action }, + { map { $_ => $cfg_editor_action{$_}->{note} } keys %cfg_editor_action }, + { sep => ' ' })); + + draw_screen(\@display_lines); + $key = pause(''); + } +} + +# edit_configuration_variables {{{3 +sub edit_configuration_variables { + $clear_screen_sub->(); + my $mode = $config{mode}->{val}; + + my %config_choices + = map { $_ => $config{$_}->{note} || '' } + # Only config values that are marked as applying to this mode. + grep { + my $key = $_; + $config{$key}->{conf} && + ( $config{$key}->{conf} eq 'ALL' + || grep { $mode eq $_ } @{$config{$key}->{conf}} ) + } keys %config; + + my $key = prompt_list( + "Enter the name of the variable you wish to configure", + '', + sub{ return keys %config_choices }, + \%config_choices); + + if ( exists($config_choices{$key}) ) { + get_config_interactive($key); + } +} + +# edit_color_rules {{{3 +sub edit_color_rules { + my ( $tbl ) = @_; + $clear_screen_sub->(); + $tbl ||= choose_visible_table(); + if ( $tbl && exists($tbl_meta{$tbl}) ) { + my $meta = $tbl_meta{$tbl}; + my @cols = ('', qw(col op arg color)); + my $info = { map { $_ => { hdr => $_, just => '-', } } @cols }; + $info->{label}->{maxw} = 30; + my $key; + my $selected_rule; + + # This loop builds a tabular view of the rules. + do { + + # Show help + if ( $key && $key eq '?' ) { + my @display_lines = ''; + push @display_lines, create_caption('Editor key mappings', + create_table2( + [ sort keys %color_editor_action ], + { map { $_ => $_ } keys %color_editor_action }, + { map { $_ => $color_editor_action{$_}->{note} } keys %color_editor_action }, + { sep => ' ' })); + draw_screen(\@display_lines); + pause(); + $key = ''; + } + else { + + # Do the action specified + $selected_rule ||= 0; + if ( $key && $color_editor_action{$key} ) { + $selected_rule = $color_editor_action{$key}->{func}->($tbl, $selected_rule); + $selected_rule ||= 0; + } + + # Build the table of rules. If the terminal has color, the selected rule + # will be highlighted; otherwise a > at the left will indicate. + my $data = $meta->{colors} || []; + foreach my $i ( 0..@$data - 1 ) { + $data->[$i]->{''} = $i == $selected_rule ? '>' : ''; + } + my @display_lines = create_table(\@cols, $info, $data); + + # Highlight selected entry + for my $i ( 0 .. $#display_lines ) { + if ( $display_lines[$i] =~ m/^>/ ) { + $display_lines[$i] = [ $display_lines[$i], 'reverse' ]; + } + } + + # Draw the screen and wait for a command. + unshift @display_lines, '', + "Editing color rules for $meta->{capt}. Press ? for help, q to " + . "quit.", ''; + draw_screen(\@display_lines); + print "\n\n", word_wrap('Rules are applied in order from top to ' + . 'bottom. The first matching rule wins and prevents the ' + . 'rest of the rules from being applied.'); + $key = pause(''); + } + } while ( $key ne 'q' ); + $meta->{color_func} = make_color_func($meta); + } +} + +# add_quick_filter {{{3 +sub add_quick_filter { + my $tbl = choose_visible_table(); + if ( $tbl && exists($tbl_meta{$tbl}) ) { + print "\n"; + my $response = prompt_list( + "Enter column name and filter text", + '', + sub { return keys %{$tbl_meta{$tbl}->{cols}} }, + () + ); + my ( $col, $text ) = split(/\s+/, $response, 2); + + # You can't filter on a nonexistent column. But if you filter on a pivoted + # table, the columns are different, so on a pivoted table, allow filtering + # on the 'name' column. + # NOTE: if a table is pivoted and un-pivoted, this will likely cause crashes. + # Currently not an issue since there's no way to toggle pivot/nopivot. + return unless $col && $text && + (exists($tbl_meta{$tbl}->{cols}->{$col}) + || ($tbl_meta{$tbl}->{pivot} && $col eq 'name')); + + my ( $sub, $err ) = compile_filter( "defined \$set->{$col} && \$set->{$col} =~ m/$text/" ); + return if !$sub || $err; + my $name = "quick_$tbl.$col"; + $filters{$name} = { + func => $sub, + text => $text, + user => 1, + quick => 1, + name => $name, + note => 'Quick-filter', + tbls => [$tbl], + }; + push @{$tbl_meta{$tbl}->{filters}}, $name; + } +} + +# clear_quick_filters {{{3 +sub clear_quick_filters { + my $tbl = choose_visible_table( + # Only tables that have quick-filters + sub { + my ( $tbl ) = @_; + return scalar grep { $filters{$_}->{quick} } @{ $tbl_meta{$tbl}->{filters} }; + } + ); + if ( $tbl && exists($tbl_meta{$tbl}) ) { + my @current = @{$tbl_meta{$tbl}->{filters}}; + @current = grep { !$filters{$_}->{quick} } @current; + $tbl_meta{$tbl}->{filters} = \@current; + } +} + +sub edit_plugins { + $clear_screen_sub->(); + + my @cols = ('', qw(class desc active)); + my $info = { map { $_ => { hdr => $_, just => '-', } } @cols }; + my @rows = map { $plugins{$_} } sort keys %plugins; + my $key; + my $selected; + + # This loop builds a tabular view of the plugins. + do { + + # Show help + if ( $key && $key eq '?' ) { + my @display_lines = ''; + push @display_lines, create_caption('Editor key mappings', + create_table2( + [ sort keys %plugin_editor_action ], + { map { $_ => $_ } keys %plugin_editor_action }, + { map { $_ => $plugin_editor_action{$_}->{note} } keys %plugin_editor_action }, + { sep => ' ' })); + draw_screen(\@display_lines); + pause(); + $key = ''; + } + + # Do the action specified + else { + $selected ||= 0; + if ( $key && $plugin_editor_action{$key} ) { + $selected = $plugin_editor_action{$key}->{func}->(\@rows, $selected); + $selected ||= 0; + } + + # Build the table of plugins. + foreach my $row ( 0.. $#rows ) { + $rows[$row]->{''} = $row eq $selected ? '>' : ' '; + } + my @display_lines = create_table(\@cols, $info, \@rows); + + # Highlight selected entry + for my $i ( 0 .. $#display_lines ) { + if ( $display_lines[$i] =~ m/^>/ ) { + $display_lines[$i] = [ $display_lines[$i], 'reverse' ]; + } + } + + # Draw the screen and wait for a command. + unshift @display_lines, '', + "Plugin Management. Press ? for help, q to quit.", ''; + draw_screen(\@display_lines); + $key = pause(''); + } + } while ( $key ne 'q' ); +} + +# edit_table {{{3 +sub edit_table { + $clear_screen_sub->(); + my ( $tbl ) = @_; + $tbl ||= choose_visible_table(); + if ( $tbl && exists($tbl_meta{$tbl}) ) { + my $meta = $tbl_meta{$tbl}; + my @cols = ('', qw(name hdr label src)); + my $info = { map { $_ => { hdr => $_, just => '-', } } @cols }; + $info->{label}->{maxw} = 30; + my $key; + my $selected_column; + + # This loop builds a tabular view of the tbl_meta's structure, showing each column + # in the entry as a row. + do { + + # Show help + if ( $key && $key eq '?' ) { + my @display_lines = ''; + push @display_lines, create_caption('Editor key mappings', + create_table2( + [ sort keys %tbl_editor_action ], + { map { $_ => $_ } keys %tbl_editor_action }, + { map { $_ => $tbl_editor_action{$_}->{note} } keys %tbl_editor_action }, + { sep => ' ' })); + draw_screen(\@display_lines); + pause(); + $key = ''; + } + else { + + # Do the action specified + $selected_column ||= $meta->{visible}->[0]; + if ( $key && $tbl_editor_action{$key} ) { + $selected_column = $tbl_editor_action{$key}->{func}->($tbl, $selected_column); + $selected_column ||= $meta->{visible}->[0]; + } + + # Build the pivoted view of the table's meta-data. If the terminal has color, + # The selected row will be highlighted; otherwise a > at the left will indicate. + my $data = []; + foreach my $row ( @{$meta->{visible}} ) { + my %hash; + @hash{ @cols } = @{$meta->{cols}->{$row}}{@cols}; + $hash{src} = '' if ref $hash{src}; + $hash{name} = $row; + $hash{''} = $row eq $selected_column ? '>' : ' '; + push @$data, \%hash; + } + my @display_lines = create_table(\@cols, $info, $data); + + # Highlight selected entry + for my $i ( 0 .. $#display_lines ) { + if ( $display_lines[$i] =~ m/^>/ ) { + $display_lines[$i] = [ $display_lines[$i], 'reverse' ]; + } + } + + # Draw the screen and wait for a command. + unshift @display_lines, '', + "Editing table definition for $meta->{capt}. Press ? for help, q to quit.", ''; + draw_screen(\@display_lines, { clear => 1 }); + $key = pause(''); + } + } while ( $key ne 'q' ); + } +} + +# choose_mode_tables {{{3 +# Choose which table(s), and in what order, to display in a given mode. +sub choose_mode_tables { + my $mode = $config{mode}->{val}; + my @tbls = @{$modes{$mode}->{visible_tables}}; + my $new = prompt_list( + "Choose tables to display", + join(' ', @tbls), + sub { return @{$modes{$mode}->{tables}} }, + { map { $_ => $tbl_meta{$_}->{capt} } @{$modes{$mode}->{tables}} } + ); + $modes{$mode}->{visible_tables} = + [ unique(grep { $_ && exists $tbl_meta{$_} } split(/\s+/, $new)) ]; + $modes{$mode}->{cust}->{visible_tables} = 1; +} + +# set_visible_table {{{3 +sub set_visible_table { + my ( $tbl ) = @_; + my $mode = $config{mode}->{val}; + my @tbls = grep { $_ eq $tbl } @{$modes{$mode}->{tables}}; + if ( @tbls == 1 ) { + $modes{$mode}->{visible_tables} = [ $tbl ]; + $modes{$mode}->{cust}->{visible_tables} = 1; + } +} + +# choose_visible_table {{{3 +sub choose_visible_table { + my ( $grep_cond ) = @_; + my $mode = $config{mode}->{val}; + my @tbls + = grep { $grep_cond ? $grep_cond->($_) : 1 } + @{$modes{$mode}->{visible_tables}}; + my $tbl = $tbls[0]; + if ( @tbls > 1 ) { + $tbl = prompt_list( + "Choose a table", + '', + sub { return @tbls }, + { map { $_ => $tbl_meta{$_}->{capt} } @tbls } + ); + } + return $tbl; +} + +sub toggle_aggregate { + my ( $tbl ) = @_; + $tbl ||= choose_visible_table(); + return unless $tbl && exists $tbl_meta{$tbl}; + my $meta = $tbl_meta{$tbl}; + $meta->{aggregate} ^= 1; +} + +sub choose_filters { + my ( $tbl ) = @_; + $tbl ||= choose_visible_table(); + return unless $tbl && exists $tbl_meta{$tbl}; + my $meta = $tbl_meta{$tbl}; + $clear_screen_sub->(); + + print "Choose filters for $meta->{capt}:\n"; + + my $ini = join(' ', @{$meta->{filters}}); + my $val = prompt_list( + 'Choose filters', + $ini, + sub { return keys %filters }, + { + map { $_ => $filters{$_}->{note} } + grep { grep { $tbl eq $_ } @{$filters{$_}->{tbls}} } + keys %filters + } + ); + + my @choices = unique($val =~ m/(\S+)/g); + foreach my $new ( grep { !exists($filters{$_}) } @choices ) { + my $answer = prompt("There is no filter called '$new'. Create it?", undef, 'y'); + if ( $answer eq 'y' ) { + create_new_filter($new, $tbl); + } + } + @choices = grep { exists $filters{$_} } @choices; + @choices = grep { grep { $tbl eq $_ } @{$filters{$_}->{tbls}} } @choices; + $meta->{filters} = [ @choices ]; + $meta->{cust}->{filters} = 1; +} + +sub choose_group_cols { + my ( $tbl ) = @_; + $tbl ||= choose_visible_table(); + return unless $tbl && exists $tbl_meta{$tbl}; + $clear_screen_sub->(); + my $meta = $tbl_meta{$tbl}; + my $curr = join(', ', @{$meta->{group_by}}); + my $val = prompt_list( + 'Group-by columns', + $curr, + sub { return keys %{$meta->{cols}} }, + { map { $_ => $meta->{cols}->{$_}->{label} } keys %{$meta->{cols}} }); + if ( $curr ne $val ) { + $meta->{group_by} = [ grep { exists $meta->{cols}->{$_} } $val =~ m/(\w+)/g ]; + $meta->{cust}->{group_by} = 1; + } +} + +sub choose_sort_cols { + my ( $tbl ) = @_; + $tbl ||= choose_visible_table(); + return unless $tbl && exists $tbl_meta{$tbl}; + $clear_screen_sub->(); + my $meta = $tbl_meta{$tbl}; + + my ( $cols, $hints ); + if ( $meta->{pivot} ) { + $cols = sub { qw(name set_0) }; + $hints = { name => 'name', set_0 => 'set_0' }; + } + else { + $cols = sub { return keys %{$meta->{cols}} }; + $hints = { map { $_ => $meta->{cols}->{$_}->{label} } keys %{$meta->{cols}} }; + } + + my $val = prompt_list( + 'Sort columns (reverse sort with -col)', + $meta->{sort_cols}, + $cols, + $hints ); + if ( $meta->{sort_cols} ne $val ) { + $meta->{sort_cols} = $val; + $meta->{cust}->{sort_cols} = 1; + $tbl_meta{$tbl}->{sort_func} = make_sort_func($tbl_meta{$tbl}); + } +} + +# create_new_filter {{{3 +sub create_new_filter { + my ( $filter, $tbl ) = @_; + $clear_screen_sub->(); + + if ( !$filter || $filter =~ m/\W/ ) { + print word_wrap("Choose a name for the filter. This name is not displayed, and is only used " + . "for internal reference. It can only contain lowercase letters, numbers, and underscores."); + print "\n\n"; + do { + $filter = prompt("Enter filter name"); + } while ( !$filter || $filter =~ m/\W/ ); + } + + my $completion = sub { keys %{$tbl_meta{$tbl}->{cols}} }; + my ( $err, $sub, $body ); + do { + $clear_screen_sub->(); + print word_wrap("A filter is a Perl subroutine that accepts a hashref of columns " + . "called \$set, and returns a true value if the filter accepts the row. Example:\n" + . " \$set->{active_secs} > 5\n" + . "will only allow rows if their active_secs column is greater than 5."); + print "\n\n"; + if ( $err ) { + print "There's an error in your filter expression: $err\n\n"; + } + $body = prompt("Enter subroutine body", undef, undef, $completion); + ( $sub, $err ) = compile_filter($body); + } while ( $err ); + + $filters{$filter} = { + func => $sub, + text => $body, + user => 1, + name => $filter, + note => 'User-defined filter', + tbls => [$tbl], + }; +} + +# get_config_interactive {{{3 +sub get_config_interactive { + my $key = shift; + $clear_screen_sub->(); + + # Print help first. + print "Enter a new value for '$key' ($config{$key}->{note}).\n"; + + my $current = ref($config{$key}->{val}) ? join(" ", @{$config{$key}->{val}}) : $config{$key}->{val}; + + my $new_value = prompt('Enter a value', $config{$key}->{pat}, $current); + $config{$key}->{val} = $new_value; +} + +sub edit_current_var_set { + my $mode = $config{mode}->{val}; + my $name = $config{"${mode}_set"}->{val}; + my $variables = $var_sets{$name}->{text}; + + my $new = $variables; + do { + $clear_screen_sub->(); + $new = prompt("Enter variables for $name", undef, $variables); + } until ( $new ); + + if ( $new ne $variables ) { + @{$var_sets{$name}}{qw(text user)} = ( $new, 1); + } +} + + +sub choose_var_set { + my ( $key ) = @_; + $clear_screen_sub->(); + + my $new_value = prompt_list( + 'Choose a set of values to display, or enter the name of a new one', + $config{$key}->{val}, + sub { return keys %var_sets }, + { map { $_ => $var_sets{$_}->{text} } keys %var_sets }); + + if ( !exists $var_sets{$new_value} ) { + add_new_var_set($new_value); + } + + $config{$key}->{val} = $new_value if exists $var_sets{$new_value}; +} + +sub switch_var_set { + my ( $cfg_var, $dir ) = @_; + my @var_sets = sort keys %var_sets; + my $cur = $config{$cfg_var}->{val}; + my $pos = grep { $_ lt $cur } @var_sets; + my $newpos = ($pos + $dir) % @var_sets; + $config{$cfg_var}->{val} = $var_sets[$newpos]; + $clear_screen_sub->(); +} + +# Online configuration and prompting functions {{{2 + +# edit_stmt_sleep_times {{{3 +sub edit_stmt_sleep_times { + $clear_screen_sub->(); + my $stmt = prompt_list('Specify a statement', '', sub { return sort keys %stmt_maker_for }); + return unless $stmt && exists $stmt_maker_for{$stmt}; + $clear_screen_sub->(); + my $curr_val = $stmt_sleep_time_for{$stmt} || 0; + my $new_val = prompt('Specify a sleep delay after calling this SQL', $num_regex, $curr_val); + if ( $new_val ) { + $stmt_sleep_time_for{$stmt} = $new_val; + } + else { + delete $stmt_sleep_time_for{$stmt}; + } +} + +# edit_server_groups {{{3 +# Choose which server connections are in a server group. First choose a group, +# then choose which connections are in it. +sub edit_server_groups { + $clear_screen_sub->(); + my $mode = $config{mode}->{val}; + my $group = $modes{$mode}->{server_group}; + my %curr = %server_groups; + my $new = choose_or_create_server_group($group, 'to edit'); + $clear_screen_sub->(); + if ( exists $curr{$new} ) { + # Don't do this step if the user just created a new server group, + # because part of that process was to choose connections. + my $cxns = join(' ', @{$server_groups{$new}}); + my @conns = choose_or_create_connection($cxns, 'for this group'); + $server_groups{$new} = \@conns; + } +} + +# choose_server_groups {{{3 +sub choose_server_groups { + $clear_screen_sub->(); + my $mode = $config{mode}->{val}; + my $group = $modes{$mode}->{server_group}; + my $new = choose_or_create_server_group($group, 'for this mode'); + $modes{$mode}->{server_group} = $new if exists $server_groups{$new}; +} + +sub choose_or_create_server_group { + my ( $group, $prompt ) = @_; + my $new = ''; + + my @available = sort keys %server_groups; + + if ( @available ) { + print "You can enter the name of a new group to create it.\n"; + + $new = prompt_list( + "Choose a server group $prompt", + $group, + sub { return @available }, + { map { $_ => join(' ', @{$server_groups{$_}}) } @available }); + + $new =~ s/\s.*//; + + if ( !exists $server_groups{$new} ) { + my $answer = prompt("There is no server group called '$new'. Create it?", undef, "y"); + if ( $answer eq 'y' ) { + add_new_server_group($new); + } + } + } + else { + $new = add_new_server_group(); + } + return $new; +} + +sub choose_or_create_connection { + my ( $cxns, $prompt ) = @_; + print "You can enter the name of a new connection to create it.\n"; + + my @available = sort keys %connections; + my $new_cxns = prompt_list( + "Choose connections $prompt", + $cxns, + sub { return @available }, + { map { $_ => $connections{$_}->{dsn} } @available }); + + my @new = unique(grep { !exists $connections{$_} } $new_cxns =~ m/(\S+)/g); + foreach my $new ( @new ) { + my $answer = prompt("There is no connection called '$new'. Create it?", undef, "y"); + if ( $answer eq 'y' ) { + add_new_dsn($new); + } + } + + return unique(grep { exists $connections{$_} } split(/\s+/, $new_cxns)); +} + +# choose_servers {{{3 +sub choose_servers { + $clear_screen_sub->(); + my $mode = $config{mode}->{val}; + my $cxns = join(' ', get_connections()); + my @chosen = choose_or_create_connection($cxns, 'for this mode'); + $modes{$mode}->{connections} = \@chosen; + $modes{$mode}->{server_group} = ''; # Clear this because it overrides {connections} + get_connections(); # This will set the server group if it matches connections just chosen +} + +# display_license {{{3 +sub display_license { + $clear_screen_sub->(); + + print $innotop_license; + + pause(); +} + +# Data-retrieval functions {{{2 +# get_status_info {{{3 +# Get SHOW STATUS and SHOW VARIABLES together. +sub get_status_info { + my @cxns = @_; + if ( !$info_gotten{status}++ ) { + foreach my $cxn ( @cxns ) { + $vars{$cxn}->{$clock} ||= {}; + my $vars = $vars{$cxn}->{$clock}; + + my $sth = do_stmt($cxn, 'SHOW_STATUS') or next; + my $res = $sth->fetchall_arrayref(); + map { $vars->{$_->[0]} = $_->[1] || 0 } @$res; + + # Calculate hi-res uptime and add cxn to the hash. This duplicates get_driver_status, + # but it's most important to have consistency. + $vars->{Uptime_hires} ||= get_uptime($cxn); + $vars->{cxn} = $cxn; + + # Add SHOW VARIABLES to the hash. If we've gotten this info before, skip and re-use. + if ( $show_variables{$cxn} ) { + $res = $show_variables{$cxn}; + } + else { + $sth = do_stmt($cxn, 'SHOW_VARIABLES') or next; + $res = $sth->fetchall_arrayref(); + $res = {map { $_->[0] => $_->[1] || 0 } @$res}; + $show_variables{$cxn} = $res; + } + @{$vars}{keys %$res} = values %$res; + + # Create sparklines for QPS and Threads_running. As a consequence of + # this, we get QPS for free. TODO: remove QPS computation from + # elsewhere. + my $pre = $vars{$cxn}->{$clock - 1}; + if ( $pre && $pre->{Uptime_hires} ) { + my @prev_qps = ($pre->{SPARK_store_qps} || '') =~ m/(\S+)/g; + my @prev_run = ($pre->{SPARK_store_run} || '') =~ m/(\S+)/g; + + # Find out the values; throw away if too many; sparkify; store. + my $this_qps = (($vars->{Questions} || 0) - ($pre->{Questions} || 0))/ + ($vars->{Uptime_hires} - $pre->{Uptime_hires}); + push @prev_qps, $this_qps; + shift @prev_qps if @prev_qps > $config{spark}->{val}; + my $qps_spark = sparkify(@prev_qps); + $vars->{SPARK_qps} = $qps_spark; + $vars->{SPARK_store_qps} = join(' ', @prev_qps); + my $this_run = $vars->{Threads_running}; + push @prev_run, $this_run; + shift @prev_run if @prev_run > $config{spark}->{val}; + my $run_spark = sparkify(@prev_run); + $vars->{SPARK_run} = $run_spark; + $vars->{SPARK_store_run} = join(' ', @prev_run); + } + } + } +} + +# Chooses a thread for explaining, killing, etc... +# First arg is a func that can be called in grep. +sub choose_thread { + my ( $grep_cond, $prompt ) = @_; + + my %thread_for = map { + # Eliminate innotop's own threads. + $_ => $dbhs{$_}->{dbh} ? $dbhs{$_}->{dbh}->{mariadb_thread_id} : 0 + } keys %connections; + + my @candidates = grep { + $_->{id} != $thread_for{$_->{cxn}} && $grep_cond->($_) + } @current_queries; + return unless @candidates; + + # Find out which server. + my @cxns = unique map { $_->{cxn} } @candidates; + my ( $cxn ) = select_cxn('On which server', @cxns); + return unless $cxn && exists($connections{$cxn}); + + # Re-filter the list of candidates to only those on this server + @candidates = grep { $_->{cxn} eq $cxn } @candidates; + + # Find out which thread to do. + my $info; + if ( @candidates > 1 ) { + + # Sort longest-active first, then longest-idle. + my $sort_func = sub { + my ( $a, $b ) = @_; + return $a->{query} && !$b->{query} ? 1 + : $b->{query} && !$a->{query} ? -1 + : ($a->{time} || 0) cmp ($b->{time} || 0); + }; + my @threads = map { $_->{id} } reverse sort { $sort_func->($a, $b) } @candidates; + + print "\n"; + my $thread = prompt_list($prompt, + $threads[0], + sub { return @threads }); + return unless $thread && $thread =~ m/$int_regex/; + + # Find the info hash of that query on that server. + ( $info ) = grep { $thread == $_->{id} } @candidates; + } + else { + $info = $candidates[0]; + } + return $info; +} + +# analyze_query {{{3 +# Allows the user to show fulltext, explain, show optimized... +sub analyze_query { + my ( $action ) = @_; + + my $info = choose_thread( + sub { $_[0]->{query} }, + 'Select a thread to analyze', + ); + return unless $info; + + my %actions = ( + e => \&display_explain, + f => \&show_full_query, + o => \&show_optimized_query, + ); + do { + $actions{$action}->($info); + print "\n"; + $action = pause('Press e to explain, f for full query, o for optimized query'); + } while ( exists($actions{$action}) ); +} + +# inc {{{3 +# Returns the difference between two sets of variables/status/innodb stuff. +sub inc { + my ( $offset, $cxn ) = @_; + my $vars = $vars{$cxn}; + if ( $offset < 0 ) { + return $vars->{$clock}; + } + elsif ( exists $vars{$clock - $offset} && !exists $vars->{$clock - $offset - 1} ) { + return $vars->{$clock - $offset}; + } + my $cur = $vars->{$clock - $offset}; + my $pre = $vars->{$clock - $offset - 1}; + return { + # Numeric variables get subtracted, non-numeric get passed straight through. + map { + $_ => + ( (defined $cur->{$_} && $cur->{$_} =~ m/$num_regex/ && ($pre->{$_} || '') =~ m/$num_regex/ ) + ? $cur->{$_} - ($pre->{$_} || 0) + : $cur->{$_} ) + } keys %{$cur} + }; +} + +# extract_values {{{3 +# Arguments are a set of values (which may be incremental, derived from +# current and previous), current, and previous values. +# TODO: there are a few places that don't remember prev set so can't pass it. +sub extract_values { + my ( $set, $cur, $pre, $tbl ) = @_; + + # Hook in event listeners + foreach my $listener ( @{$event_listener_for{extract_values}} ) { + $listener->extract_values($set, $cur, $pre, $tbl); + } + + my $result = {}; + my $meta = $tbl_meta{$tbl}; + my $cols = $meta->{cols}; + foreach my $key ( keys %$cols ) { + my $info = $cols->{$key} + or die "Column '$key' doesn't exist in $tbl"; + die "No func defined for '$key' in $tbl" + unless $info->{func}; + eval { + $result->{$key} = $info->{func}->($set, $cur, $pre) + }; + if ( $EVAL_ERROR ) { + if ( $config{debug}->{val} ) { + die $EVAL_ERROR; + } + $result->{$key} = $info->{num} ? 0 : ''; + } + } + return $result; +} + +# get_processlist_stats {{{3 +# Inserts special values as though they are SHOW STATUS counters. +sub get_processlist_stats { + my @cxns = @_; + @current_queries = (); + if ( !$info_gotten{processlist_stats}++ ) { + foreach my $cxn ( @cxns ) { + my $max_query_time = 0; + my ($user_threads, $slaves, $longest_sql, $slave_sql, $locked); + $vars{$cxn}->{$clock} ||= {}; + my $vars = $vars{$cxn}->{$clock}; + $vars->{cxn} = $cxn; + my $stmt = do_stmt($cxn, 'PROCESSLIST_NO_IS') or next; + my $arr = $stmt->fetchall_arrayref({}); + my $cur = undef; + foreach my $thread ( @$arr ) { + if ( ($thread->{state} || '') =~ m/lock/i ) { + $locked++; + } + # Ignore non-user threads, but remember the SQL in case there is + # no user SQL. Ignore sleeping threads and SHOW PROCESSLIST + # threads. + if ( ($thread->{user} || '') =~ m/system user/ ) { + if ( $thread->{info} && $thread->{time} ) { + $slave_sql = $thread->{info}; + } + next; + } + next unless $thread->{command}; + if ( $thread->{command} eq 'Binlog Dump' ) { + $slaves++; + next; + } + next unless $thread->{command} eq 'Query'; + next unless $thread->{state} && $thread->{info}; + next if $thread->{info} =~ m#/\*innotop#; + $user_threads++; + if ( $thread->{time} > $max_query_time ) { + $max_query_time = $thread->{time}; + $longest_sql = $thread->{info}; + if ( $thread->{state} eq 'Checking table' ) { + $longest_sql = 'CHECK TABLE ' . $thread->{info}; + } + $cur = { + cxn => $cxn, + id => $thread->{id}, + db => $thread->{db}, + query => $thread->{info}, + time => $thread->{time}, + user => $thread->{user}, + host => $thread->{host}, + }; + $thread->{host} =~ s/:.*$//; + } + } + $vars->{Max_query_time} = $max_query_time; + $vars->{User_threads_running} = $user_threads; + $vars->{Slaves} = $slaves || 0; + $vars->{Longest_sql} = $longest_sql || $slave_sql || ''; + $vars->{Locked_count} = $locked || 0; + $vars->{Uptime_hires} ||= get_uptime($cxn); + push @current_queries, $cur if $cur; + } + } +} + +# get_full_processlist {{{3 +sub get_full_processlist { + my @cxns = @_; + my @result; + foreach my $cxn ( @cxns ) { + my $stmt = do_stmt($cxn, 'PROCESSLIST') or next; + my $arr = $stmt->fetchall_arrayref({}); + push @result, map { $_->{cxn} = $cxn; $_ } @$arr; + } + return @result; +} + +# get_open_tables {{{3 +sub get_open_tables { + my @cxns = @_; + my @result; + foreach my $cxn ( @cxns ) { + my $stmt = do_stmt($cxn, 'OPEN_TABLES') or next; + my $arr = $stmt->fetchall_arrayref({}); + push @result, map { $_->{cxn} = $cxn; $_ } @$arr; + } + return @result; +} + +# get_index_statistics {{{3 +sub get_index_statistics { + my @cxns = @_; + my @result; + foreach my $cxn ( @cxns ) { + my $stmt = do_stmt($cxn, 'INDEX_STATISTICS') or next; + my $arr = $stmt->fetchall_arrayref({}); + push @result, map { $_->{cxn} = $cxn; $_ } @$arr; + } + return @result; +} + +# get_index_table_statistics {{{3 +sub get_index_table_statistics { + my @cxns = @_; + my @result; + foreach my $cxn ( @cxns ) { + my $stmt = do_stmt($cxn, 'INDEX_TABLE_STATISTICS') or next; + my $arr = $stmt->fetchall_arrayref({}); + push @result, map { $_->{cxn} = $cxn; $_ } @$arr; + } + return @result; +} + +# get_table_statistics {{{3 +sub get_table_statistics { + my @cxns = @_; + my @result; + foreach my $cxn ( @cxns ) { + my $stmt = do_stmt($cxn, 'TABLE_STATISTICS') or next; + my $arr = $stmt->fetchall_arrayref({}); + push @result, map { $_->{cxn} = $cxn; $_ } @$arr; + } + return @result; +} + +# get_innodb_blocked_blocker {{{3 +sub get_innodb_blocked_blocker { + my @cxns = @_; + my @result; + foreach my $cxn ( @cxns ) { + my $stmt = do_stmt($cxn, 'INNODB_BLOCKED_BLOCKER') or next; + my $arr = $stmt->fetchall_arrayref({}); + push @result, map { $_->{cxn} = $cxn; $_ } @$arr; + } + return @result; +} + +# get_innodb_status {{{3 +sub get_innodb_status { + my ( $cxns, $addl_sections ) = @_; + if ( !$config{skip_innodb}->{val} && !$info_gotten{innodb_status}++ ) { + + # Determine which sections need to be parsed + my %sections_required = + map { $tbl_meta{$_}->{innodb} => 1 } + grep { $_ && $tbl_meta{$_}->{innodb} } + get_visible_tables(); + + # Add in any other sections the caller requested. + foreach my $sec ( @$addl_sections ) { + $sections_required{$sec} = 1; + } + + foreach my $cxn ( @$cxns ) { + my $innodb_status_text; + + if ( $file ) { # Try to fetch status text from the file. + my @stat = stat($file); + + # Initialize the file. + if ( !$file_mtime ) { + # Initialize to 130k from the end of the file (because the limit + # on the size of innodb status is 128k even with Google's patches) + # and try to grab the last status from the file. + sysseek($file, (-128 * 1_024), 2); + } + + # Read from the file. + my $buffer; + if ( !$file_mtime || $file_mtime != $stat[9] ) { + $file_data = ''; + while ( sysread($file, $buffer, 4096) ) { + $file_data .= $buffer; + } + $file_mtime = $stat[9]; + } + + # Delete everything but the last InnoDB status text from the file. + $file_data =~ s/\A.*(?=^=====================================\n...... ........ INNODB MONITOR OUTPUT)//ms; + $innodb_status_text = $file_data; + } + + else { + next if ($show_variables{$cxn}->{have_innodb} || 'YES') eq 'NO'; + my $stmt = do_stmt($cxn, 'INNODB_STATUS') or next; + $innodb_status_text = $stmt->fetchrow_hashref()->{status}; + } + + next unless $innodb_status_text + && substr($innodb_status_text, 0, 100) =~ m/INNODB MONITOR OUTPUT/; + + # Parse and merge into %vars storage + my %innodb_status = ( + $innodb_parser->get_status_hash( + $innodb_status_text, + $config{debug}->{val}, + \%sections_required, + 0, # don't parse full lock information + $show_variables{$cxn}->{version} + ) + ); + if ( !$innodb_status{IB_got_all} && $config{auto_wipe_dl}->{val} ) { + clear_deadlock($cxn); + } + + # Merge using a hash slice, which is the fastest way + $vars{$cxn}->{$clock} ||= {}; + my $hash = $vars{$cxn}->{$clock}; + @{$hash}{ keys %innodb_status } = values %innodb_status; + $hash->{cxn} = $cxn; + $hash->{Uptime_hires} ||= get_uptime($cxn); + } + } +} + +# clear_deadlock {{{3 +sub clear_deadlock { + my ( $cxn ) = @_; + return if $clearing_deadlocks++; + my $tbl = $connections{$cxn}->{dl_table}; + return unless $tbl; + + eval { + # disable binary logging for the session + do_query($cxn, "set SQL_LOG_BIN=0"); + + # Set up the table for creating a deadlock. + my $engine = version_ge($dbhs{$cxn}->{dbh}, '4.1.2') ? 'engine' : 'type'; + return unless do_query($cxn, "drop table if exists $tbl"); + return unless do_query($cxn, "create table $tbl(a int) $engine=innodb"); + return unless do_query($cxn, "delete from $tbl"); + return unless do_query($cxn, "insert into $tbl(a) values(0), (1)"); + return unless do_query($cxn, "commit"); # Or the children will block against the parent + + # Fork off two children to deadlock against each other. + my %children; + foreach my $child ( 0..1 ) { + my $pid = fork(); + if ( defined($pid) && $pid == 0 ) { # I am a child + deadlock_thread( $child, $tbl, $cxn ); + } + elsif ( !defined($pid) ) { + die("Unable to fork for clearing deadlocks!\n"); + } + # I already exited if I'm a child, so I'm the parent. + $children{$child} = $pid; + } + + # Wait for the children to exit. + foreach my $child ( keys %children ) { + my $pid = waitpid($children{$child}, 0); + } + + # Clean up. + do_query($cxn, "drop table $tbl"); + + # enable binary logging for the session again + # the session by itself will not be used anymore, but this is clean :) + do_query($cxn, "set SQL_LOG_BIN=1"); + + }; + if ( $EVAL_ERROR ) { + print $EVAL_ERROR; + pause(); + } + + $clearing_deadlocks = 0; +} + +sub get_master_logs { + my @cxns = @_; + my @result; + if ( !$info_gotten{master_logs}++ ) { + foreach my $cxn ( @cxns ) { + my $stmt = do_stmt($cxn, 'SHOW_MASTER_LOGS') or next; + push @result, @{$stmt->fetchall_arrayref({})}; + } + } + return @result; +} + +# get_master_slave_status {{{3 +# Inserts special counters as though they are SHOW STATUS counters. +sub get_master_slave_status { + my @cxns = @_; + if ( !$info_gotten{replication_status}++ ) { + foreach my $cxn ( @cxns ) { + $vars{$cxn}->{$clock} ||= {}; + my $vars = $vars{$cxn}->{$clock}; + $vars->{cxn} = $cxn; + $vars->{Uptime_hires} ||= get_uptime($cxn); + + my $stmt = do_stmt($cxn, 'SHOW_MASTER_STATUS') or next; + my $res = $stmt->fetchall_arrayref({})->[0]; + @{$vars}{ keys %$res } = values %$res; + + } + } +} + +# get_slave_status {{{3 +# Separated handling of slave status to support 5.7 and replication channels +sub get_slave_status { + my ($cxn, $channel) = @_; + my $chcxn = $channel . '=' . $cxn; + $vars{$chcxn}->{$clock} ||= {}; + my $vars = $vars{$chcxn}->{$clock}; + $vars->{chcxn} = $chcxn; + $vars->{Uptime_hires} ||= get_uptime($chcxn); + + if ( $channel =~ /no_channels/ ) { + my $stmt = do_stmt($cxn, 'SHOW_SLAVE_STATUS') or next; + my $res = $stmt->fetchall_arrayref({}); + if ( $res && @$res ) { + $res = $res->[0]; + @{$vars}{ keys %$res } = values %$res; + $vars->{Slave_ok} = + (($res->{slave_sql_running} || 'Yes') eq 'Yes' + && ($res->{slave_io_running} || 'Yes') eq 'Yes') ? 'Yes' : 'No'; + } + else { + $vars->{Slave_ok} = 'Off'; + } + } else { + my $dbh = connect_to_db($cxn); + my $sql = 'SHOW SLAVE STATUS FOR CHANNEL \'' . $channel . '\''; + my $stmt = $dbh->prepare($sql ) ; + $stmt->execute(); + my $res = $stmt->fetchall_arrayref({}); + if ( $res && @$res ) { + $res = $res->[0]; + @{$vars}{ keys %$res } = values %$res; + $vars->{Slave_ok} = + (($res->{slave_sql_running} || 'Yes') eq 'Yes' + && ($res->{slave_io_running} || 'Yes') eq 'Yes') ? 'Yes' : 'No'; + } + else { + $vars->{Slave_ok} = 'Off'; + } + } + } + + + + +sub is_func { + my ( $word ) = @_; + return defined(&$word) + || eval "my \$x= sub { $word }; 1" + || $EVAL_ERROR !~ m/^Bareword/; +} + +# Documentation {{{1 +# ############################################################################ +# I put this last as per the Dog book. +# ############################################################################ +=pod + +=head1 NAME + +innotop - MySQL and InnoDB transaction/status monitor. + +=head1 SYNOPSIS + +To monitor servers normally: + + innotop + +To monitor InnoDB status information from a file: + + innotop /var/log/mysql/mysqld.err + +To run innotop non-interactively in a pipe-and-filter configuration: + + innotop --count 5 -d 1 -n + +To monitor a database on another system using a particular username and password: + + innotop -u <username> -p <password> -h <hostname> + +=head1 DESCRIPTION + +innotop monitors MySQL servers. Each of its modes shows you a different aspect +of what's happening in the server. For example, there's a mode for monitoring +replication, one for queries, and one for transactions. innotop refreshes its +data periodically, so you see an updating view. + +innotop has lots of features for power users, but you can start and run it with +virtually no configuration. If you're just getting started, see +L<"QUICK-START">. Press '?' at any time while running innotop for +context-sensitive help. + +=head1 QUICK-START + +To start innotop, open a terminal or command prompt. If you have installed +innotop on your system, you should be able to just type "innotop" and press +Enter; otherwise, you will need to change to innotop's directory and type "perl +innotop". + +With no options specified, innotop will attempt to connect to a MySQL server on +localhost using mariadb_read_default_group=client for other connection +parameters. If you need to specify a different username and password, use the +-u and -p options, respectively. To monitor a MySQL database on another +host, use the -h option. + +After you've connected, innotop should show you something like the following: + + [RO] Query List (? for help) localhost, 01:11:19, 449.44 QPS, 14/7/163 con/run + + CXN When Load QPS Slow QCacheHit KCacheHit BpsIn BpsOut + localhost Total 0.00 1.07k 697 0.00% 98.17% 476.83k 242.83k + + CXN Cmd ID User Host DB Time Query + localhost Query 766446598 test 10.0.0.1 foo 00:02 INSERT INTO table ( + + +(This sample is truncated at the right so it will fit on a terminal when running +'man innotop') + +If your server is busy, you'll see more output. Notice the first line on the +screen, which tells you that readonly is set to true ([RO]), what mode you're +in and what server you're connected to. You can change to other modes with +keystrokes; press 'T' to switch to a list of InnoDB transactions, for example. + +Press the '?' key to see what keys are active in the current mode. You can +press any of these keys and innotop will either take the requested action or +prompt you for more input. If your system has Term::ReadLine support, you can +use TAB and other keys to auto-complete and edit input. + +To quit innotop, press the 'q' key. + +=head1 OPTIONS + +innotop is mostly configured via its configuration file, but some of the +configuration options can come from the command line. You can also specify a +file to monitor for InnoDB status output; see L<"MONITORING A FILE"> for more +details. + +You can negate some options by prefixing the option name with --no. For +example, --noinc (or --no-inc) negates L<"--inc">. + +=over + +=item --color + +Enable or disable terminal coloring. Corresponds to the L<"color"> config file +setting. + +=item --config + +Specifies a configuration file to read. This option is non-sticky, that is to +say it does not persist to the configuration file itself. + +=item --count + +Refresh only the specified number of times (ticks) before exiting. Each refresh +is a pause for L<"interval"> seconds, followed by requesting data from MySQL +connections and printing it to the terminal. + +=item --delay + +Specifies the amount of time to pause between ticks (refreshes). Corresponds to +the configuration option L<"interval">. + +=item --help + +Print a summary of command-line usage and exit. + +=item --host + +Host to connect to. + +=item --inc + +Specifies whether innotop should display absolute numbers or relative numbers +(offsets from their previous values). Corresponds to the configuration option +L<"status_inc">. + +=item --mode + +Specifies the mode in which innotop should start. Corresponds to the +configuration option L<"mode">. + +=item --nonint + +Enable non-interactive operation. See L<"NON-INTERACTIVE OPERATION"> for more. + +=item --password + +Password to use for connection. + +=item --port + +Port to use for connection. + +=item --skipcentral + +Don't read the central configuration file. + +=item --timestamp + +In -n mode, write a timestamp either before every screenful of output, or if +the option is given twice, at the start of every line. The format is controlled +by the timeformat config variable. + +=item --user + +User to use for connection. + +=item --version + +Output version information and exit. + +=item --write + +Sets the configuration option L<"readonly"> to 0, making innotop write the +running configuration to ~/.innotop/innotop.conf on exit, if no configuration +file was loaded at start-up. + +=back + +=head1 HOTKEYS + +innotop is interactive, and you control it with key-presses. + +=over + +=item * + +Uppercase keys switch between modes. + +=item * + +Lowercase keys initiate some action within the current mode. + +=item * + +Other keys do something special like change configuration or show the +innotop license. + +=back + +Press '?' at any time to see the currently active keys and what they do. + +=head1 MODES + +Each of innotop's modes retrieves and displays a particular type of data from +the servers you're monitoring. You switch between modes with uppercase keys. +The following is a brief description of each mode, in alphabetical order. To +switch to the mode, press the key listed in front of its heading in the +following list: + +=over + +=item A: Health Dashboard + +This mode displays a single table with one row per monitored server. The +columns show essential overview information about the server's health, and +coloration rules show whether replication is running or if there are any very +long-running queries or excessive replication delay. + +=item B: InnoDB Buffers + +This mode displays information about the InnoDB buffer pool, page statistics, +insert buffer, and adaptive hash index. The data comes from SHOW INNODB STATUS. + +This mode contains the L<"buffer_pool">, L<"page_statistics">, +L<"insert_buffers">, and L<"adaptive_hash_index"> tables by default. + +=item C: Command Summary + +This mode is similar to mytop's Command Summary mode. It shows the +L<"cmd_summary"> table, which looks something like the following: + + Command Summary (? for help) localhost, 25+07:16:43, 2.45 QPS, 3 thd, 5.0.40 + _____________________ Command Summary _____________________ + Name Value Pct Last Incr Pct + Select_scan 3244858 69.89% 2 100.00% + Select_range 1354177 29.17% 0 0.00% + Select_full_join 39479 0.85% 0 0.00% + Select_full_range_join 4097 0.09% 0 0.00% + Select_range_check 0 0.00% 0 0.00% + +The command summary table is built by extracting variables from +L<"STATUS_VARIABLES">. The variables must be numeric and must match the prefix +given by the L<"cmd_filter"> configuration variable. The variables are then +sorted by value descending and compared to the last variable, as shown above. +The percentage columns are percentage of the total of all variables in the +table, so you can see the relative weight of the variables. + +The example shows what you see if the prefix is "Select_". The default +prefix is "Com_". You can choose a prefix with the 's' key. + +It's rather like running SHOW VARIABLES LIKE "prefix%" with memory and +nice formatting. + +Values are aggregated across all servers. The Pct columns are not correctly +aggregated across multiple servers. This is a known limitation of the grouping +algorithm that may be fixed in the future. + +=item D: InnoDB Deadlocks + +This mode shows the transactions involved in the last InnoDB deadlock. A second +table shows the locks each transaction held and waited for. A deadlock is +caused by a cycle in the waits-for graph, so there should be two locks held and +one waited for unless the deadlock information is truncated. + +InnoDB puts deadlock information before some other information in the SHOW +INNODB STATUS output. If there are a lot of locks, the deadlock information can +grow very large, and there is a limit on the size of the SHOW INNODB +STATUS output. A large deadlock can fill the entire output, or even be +truncated, and prevent you from seeing other information at all. If you are +running innotop in another mode, for example T mode, and suddenly you don't see +anything, you might want to check and see if a deadlock has wiped out the data +you need. + +If it has, you can create a small deadlock to replace the large one. Use the +'w' key to 'wipe' the large deadlock with a small one. This will not work +unless you have defined a deadlock table for the connection (see L<"SERVER +CONNECTIONS">). + +You can also configure innotop to automatically detect when a large deadlock +needs to be replaced with a small one (see L<"auto_wipe_dl">). + +This mode displays the L<"deadlock_transactions"> and L<"deadlock_locks"> tables +by default. + +=item F: InnoDB Foreign Key Errors + +This mode shows the last InnoDB foreign key error information, such as the +table where it happened, when and who and what query caused it, and so on. + +InnoDB has a huge variety of foreign key error messages, and many of them are +just hard to parse. innotop doesn't always do the best job here, but there's +so much code devoted to parsing this messy, unparseable output that innotop is +likely never to be perfect in this regard. If innotop doesn't show you what +you need to see, just look at the status text directly. + +This mode displays the L<"fk_error"> table by default. + +=item I: InnoDB I/O Info + +This mode shows InnoDB's I/O statistics, including the I/O threads, pending I/O, +file I/O miscellaneous, and log statistics. It displays the L<"io_threads">, +L<"pending_io">, L<"file_io_misc">, and L<"log_statistics"> tables by default. + +=item K: InnoDB Lock Waits + +This mode shows information from InnoDB plugin's transaction and locking tables. +You can use it to find when a transaction is waiting for another, and kill the +blocking transaction. It displays the L<"innodb_blocked_blocker>" table. + +=item L: Locks + +This mode shows information about current locks. At the moment only InnoDB +locks are supported, and by default you'll only see locks for which transactions +are waiting. This information comes from the TRANSACTIONS section of the InnoDB +status text. If you have a very busy server, you may have frequent lock waits; +it helps to be able to see which tables and indexes are the "hot spot" for +locks. If your server is running pretty well, this mode should show nothing. + +You can configure MySQL and innotop to monitor not only locks for which a +transaction is waiting, but those currently held, too. You can do this with the +InnoDB Lock Monitor (L<http://dev.mysql.com/doc/en/innodb-monitor.html>). It's +not documented in the MySQL manual, but creating the lock monitor with the +following statement also affects the output of SHOW INNODB STATUS, which innotop +uses: + + CREATE TABLE innodb_lock_monitor(a int) ENGINE=INNODB; + +This causes InnoDB to print its output to the MySQL file every 16 seconds or so, +as stated in the manual, but it also makes the normal SHOW INNODB STATUS output +include lock information, which innotop can parse and display (that's the +undocumented feature). + +This means you can do what may have seemed impossible: to a limited extent +(InnoDB truncates some information in the output), you can see which transaction +holds the locks something else is waiting for. You can also enable and disable +the InnoDB Lock Monitor with the key mappings in this mode. + +This mode displays the L<"innodb_locks"> table by default. Here's a sample of +the screen when one connection is waiting for locks another connection holds: + + _________________________________ InnoDB Locks __________________________ + CXN ID Type Waiting Wait Active Mode DB Table Index + localhost 12 RECORD 1 00:10 00:10 X test t1 PRIMARY + localhost 12 TABLE 0 00:10 00:10 IX test t1 + localhost 12 RECORD 1 00:10 00:10 X test t1 PRIMARY + localhost 11 TABLE 0 00:00 00:25 IX test t1 + localhost 11 RECORD 0 00:00 00:25 X test t1 PRIMARY + +You can see the first connection, ID 12, is waiting for a lock on the PRIMARY +key on test.t1, and has been waiting for 10 seconds. The second connection +isn't waiting, because the Waiting column is 0, but it holds locks on the same +index. That tells you connection 11 is blocking connection 12. + +=item M: Master/Slave Replication Status + +This mode shows the output of SHOW SLAVE STATUS and SHOW MASTER STATUS in three +tables. The first two divide the slave's status into SQL and I/O thread status, +and the last shows master status. Filters are applied to eliminate non-slave +servers from the slave tables, and non-master servers from the master table. + +This mode displays the L<"slave_sql_status">, L<"slave_io_status">, and +L<"master_status"> tables by default. + +=item O: Open Tables + +This section comes from MySQL's SHOW OPEN TABLES command. By default it is +filtered to show tables which are in use by one or more queries, so you can +get a quick look at which tables are 'hot'. You can use this to guess which +tables might be locked implicitly. + +This mode displays the L<"open_tables"> mode by default. + +=item U: User Statistics + +This mode displays data that's available in Percona's enhanced version of MySQL +(also known as Percona Server with XtraDB). Specifically, it makes it easy to +enable and disable the so-called "user statistics." This feature gathers stats +on clients, threads, users, tables, and indexes and makes them available as +INFORMATION_SCHEMA tables. These are invaluable for understanding what your +server is doing. They are also available in MariaDB. + +The statistics supported so far are only from the TABLE_STATISTICS and +INDEX_STATISTICS tables added by Percona. There are three views: one of table stats, +one of index stats (which can be aggregated with the = key), and one of both. + +The server doesn't gather these stats by default. You have to set the variable +userstat_running to turn it on. You can do this easily with innotop from U mode, +with the 's' key. + +=item Q: Query List + +This mode displays the output from SHOW FULL PROCESSLIST, much like B<mytop>'s +query list mode. This mode does B<not> show InnoDB-related information. This +is probably one of the most useful modes for general usage. + +There is an informative header that shows general status information about +your server. You can toggle it on and off with the 'h' key. By default, +innotop hides inactive processes and its own process. You can toggle these on +and off with the 'i' and 'a' keys. + +You can EXPLAIN a query from this mode with the 'e' key. This displays the +query's full text, the results of EXPLAIN, and in newer MySQL versions, even +the optimized query resulting from EXPLAIN EXTENDED. innotop also tries to +rewrite certain queries to make them EXPLAIN-able. For example, INSERT/SELECT +statements are rewritable. + +This mode displays the L<"q_header"> and L<"processlist"> tables by default. + +=item R: InnoDB Row Operations and Semaphores + +This mode shows InnoDB row operations, row operation miscellaneous, semaphores, +and information from the wait array. It displays the L<"row_operations">, +L<"row_operation_misc">, L<"semaphores">, and L<"wait_array"> tables by default. + +=item S: Variables & Status + +This mode calculates statistics, such as queries per second, and prints them out +in several different styles. You can show absolute values, or incremental values +between ticks. + +You can switch between the views by pressing a key. The 's' key prints a +single line each time the screen updates, in the style of B<vmstat>. The 'g' +key changes the view to a graph of the same numbers, sort of like B<tload>. +The 'v' key changes the view to a pivoted table of variable names on the left, +with successive updates scrolling across the screen from left to right. You can +choose how many updates to put on the screen with the L<"num_status_sets"> +configuration variable. + +Headers may be abbreviated to fit on the screen in interactive operation. You +choose which variables to display with the 'c' key, which selects from +predefined sets, or lets you create your own sets. You can edit the current set +with the 'e' key. + +This mode doesn't really display any tables like other modes. Instead, it uses +a table definition to extract and format the data, but it then transforms the +result in special ways before outputting it. It uses the L<"var_status"> table +definition for this. + +=item T: InnoDB Transactions + +This mode shows transactions from the InnoDB monitor's output, in B<top>-like +format. This mode is the reason I wrote innotop. + +You can kill queries or processes with the 'k' and 'x' keys, and EXPLAIN a query +with the 'e' or 'f' keys. InnoDB doesn't print the full query in transactions, +so explaining may not work right if the query is truncated. + +The informational header can be toggled on and off with the 'h' key. By +default, innotop hides inactive transactions and its own transaction. You can +toggle this on and off with the 'i' and 'a' keys. + +This mode displays the L<"t_header"> and L<"innodb_transactions"> tables by +default. + +=back + +=head1 INNOTOP STATUS + +The first line innotop displays is a "status bar" of sorts. What it contains +depends on the mode you're in, and what servers you're monitoring. The first +few words are always [RO] (if readonly is set to 1), the innotop mode, such as +"InnoDB Txns" for T mode, followed by a reminder to press '?' for help at any +time. + +=head2 ONE SERVER + +The simplest case is when you're monitoring a single server. In this case, the +name of the connection is next on the status line. This is the name you gave +when you created the connection -- most likely the MySQL server's hostname. +This is followed by the server's uptime. + +If you're in an InnoDB mode, such as T or B, the next word is "InnoDB" followed +by some information about the SHOW INNODB STATUS output used to render the +screen. The first word is the number of seconds since the last SHOW INNODB +STATUS, which InnoDB uses to calculate some per-second statistics. The next is +a smiley face indicating whether the InnoDB output is truncated. If the smiley +face is a :-), all is well; there is no truncation. A :^| means the transaction +list is so long, InnoDB has only printed out some of the transactions. Finally, +a frown :-( means the output is incomplete, which is probably due to a deadlock +printing too much lock information (see L<"D: InnoDB Deadlocks">). + +The next two words indicate the server's queries per second (QPS) and how many +threads (connections) exist. Finally, the server's version number is the last +thing on the line. + +=head2 MULTIPLE SERVERS + +If you are monitoring multiple servers (see L<"SERVER CONNECTIONS">), the status +line does not show any details about individual servers. Instead, it shows the +names of the connections that are active. Again, these are connection names you +specified, which are likely to be the server's hostname. A connection that has +an error is prefixed with an exclamation point. + +If you are monitoring a group of servers (see L<"SERVER GROUPS">), the status +line shows the name of the group. If any connection in the group has an +error, the group's name is followed by the fraction of the connections that +don't have errors. + +See L<"ERROR HANDLING"> for more details about innotop's error handling. + +=head2 MONITORING A FILE + +If you give a filename on the command line, innotop will not connect to ANY +servers at all. It will watch the specified file for InnoDB status output and +use that as its data source. It will always show a single connection called +'file'. And since it can't connect to a server, it can't determine how long the +server it's monitoring has been up; so it calculates the server's uptime as time +since innotop started running. + +=head1 SERVER ADMINISTRATION + +While innotop is primarily a monitor that lets you watch and analyze your +servers, it can also send commands to servers. The most frequently useful +commands are killing queries and stopping or starting slaves. + +You can kill a connection, or in newer versions of MySQL kill a query but not a +connection, from L<"Q: Query List"> and L<"T: InnoDB Transactions"> modes. +Press 'k' to issue a KILL command, or 'x' to issue a KILL QUERY command. +innotop will prompt you for the server and/or connection ID to kill (innotop +does not prompt you if there is only one possible choice for any input). +innotop pre-selects the longest-running query, or the oldest connection. +Confirm the command with 'y'. + +In L<"M: Master/Slave Replication Status"> mode, you can start and stop slaves +with the 'a' and 'o' keys, respectively. You can send these commands to many +slaves at once. innotop fills in a default command of START SLAVE or STOP SLAVE +for you, but you can actually edit the command and send anything you wish, such +as SET GLOBAL SQL_SLAVE_SKIP_COUNTER=1 to make the slave skip one binlog event +when it starts. + +You can also ask innotop to calculate the earliest binlog in use by any slave +and issue a PURGE MASTER LOGS on the master. Use the 'b' key for this. innotop +will prompt you for a master to run the command on, then prompt you for the +connection names of that master's slaves (there is no way for innotop to +determine this reliably itself). innotop will find the minimum binlog in use by +these slave connections and suggest it as the argument to PURGE MASTER LOGS. + +in L<"U: User Statistics"> mode, you can use the 's' key to start and stop +the collection of the statistics data for TABLE_STATISTICS and similar. + +=head1 SERVER CONNECTIONS + +When you create a server connection using '@', innotop asks you for a series of +inputs, as follows: + +=over + +=item DSN + +A DSN is a Data Source Name, which is the initial argument passed to the DBI +module for connecting to a server. It is usually of the form + + DBI:MariaDB:;mariadb_read_default_group=mysql;host=HOSTNAME + +Since this DSN is passed to the DBD::MariaDB driver, you should read the driver's +documentation at L<https://metacpan.org/pod/DBD::MariaDB> for +the exact details on all the options you can pass the driver in the DSN. You +can read more about DBI at L<http://dbi.perl.org/docs/>, and especially at +L<http://search.cpan.org/~timb/DBI/DBI.pm>. + +The mariadb_read_default_group=mysql option lets the DBD driver read your MySQL +options files, such as ~/.my.cnf on UNIX-ish systems. You can use this to avoid +specifying a username or password for the connection. + +=item InnoDB Deadlock Table + +This optional item tells innotop a table name it can use to deliberately create +a small deadlock (see L<"D: InnoDB Deadlocks">). If you specify this option, +you just need to be sure the table doesn't exist, and that innotop can create +and drop the table with the InnoDB storage engine. You can safely omit or just +accept the default if you don't intend to use this. + +=item Username + +innotop will ask you if you want to specify a username. If you say 'y', it will +then prompt you for a user name. If you have a MySQL option file that specifies +your username, you don't have to specify a username. + +The username defaults to your login name on the system you're running innotop on. + +=item Password + +innotop will ask you if you want to specify a password. Like the username, the +password is optional, but there's an additional prompt that asks if you want to +save the password in the innotop configuration file. If you don't save it in +the configuration file, innotop will prompt you for a password each time it +starts. Passwords in the innotop configuration file are saved in plain text, +not encrypted in any way. + +=back + +Once you finish answering these questions, you should be connected to a server. +But innotop isn't limited to monitoring a single server; you can define many +server connections and switch between them by pressing the '@' key. See +L<"SWITCHING BETWEEN CONNECTIONS">. + +=head1 SERVER GROUPS + +If you have multiple MySQL instances, you can put them into named groups, such +as 'all', 'masters', and 'slaves', which innotop can monitor all together. + +You can choose which group to monitor with the '#' key, and you can press the +TAB key to switch to the next group. If you're not currently monitoring a +group, pressing TAB selects the first group. + +To create a group, press the '#' key and type the name of your new group, then +type the names of the connections you want the group to contain. + +=head1 SWITCHING BETWEEN CONNECTIONS + +innotop lets you quickly switch which servers you're monitoring. The most basic +way is by pressing the '@' key and typing the name(s) of the connection(s) you +want to use. This setting is per-mode, so you can monitor different connections +in each mode, and innotop remembers which connections you choose. + +You can quickly switch to the 'next' connection in alphabetical order with the +'n' key. If you're monitoring a server group (see L<"SERVER GROUPS">) this will +switch to the first connection. + +You can also type many connection names, and innotop will fetch and display data +from them all. Just separate the connection names with spaces, for example +"server1 server2." Again, if you type the name of a connection that doesn't +exist, innotop will prompt you for connection information and create the +connection. + +Another way to monitor multiple connections at once is with server groups. You +can use the TAB key to switch to the 'next' group in alphabetical order, or if +you're not monitoring any groups, TAB will switch to the first group. + +innotop does not fetch data in parallel from connections, so if you are +monitoring a large group or many connections, you may notice increased delay +between ticks. + +When you monitor more than one connection, innotop's status bar changes. See +L<"INNOTOP STATUS">. + +=head1 ERROR HANDLING + +Error handling is not that important when monitoring a single connection, but is +crucial when you have many active connections. A crashed server or lost +connection should not crash innotop. As a result, innotop will continue to run +even when there is an error; it just won't display any information from the +connection that had an error. Because of this, innotop's behavior might confuse +you. It's a feature, not a bug! + +innotop does not continue to query connections that have errors, because they +may slow innotop and make it hard to use, especially if the error is a problem +connecting and causes a long time-out. Instead, innotop retries the connection +occasionally to see if the error still exists. If so, it will wait until some +point in the future. The wait time increases in ticks as the Fibonacci series, +so it tries less frequently as time passes. + +Since errors might only happen in certain modes because of the SQL commands +issued in those modes, innotop keeps track of which mode caused the error. If +you switch to a different mode, innotop will retry the connection instead of +waiting. + +By default innotop will display the problem in red text at the bottom of the +first table on the screen. You can disable this behavior with the +L<"show_cxn_errors_in_tbl"> configuration option, which is enabled by default. +If the L<"debug"> option is enabled, innotop will display the error at the +bottom of every table, not just the first. And if L<"show_cxn_errors"> is +enabled, innotop will print the error text to STDOUT as well. Error messages +might only display in the mode that caused the error, depending on the mode and +whether innotop is avoiding querying that connection. + +=head1 NON-INTERACTIVE OPERATION + +You can run innotop in non-interactive mode, in which case it is entirely +controlled from the configuration file and command-line options. To start +innotop in non-interactive mode, give the L"<--nonint"> command-line option. +This changes innotop's behavior in the following ways: + +=over + +=item * + +Certain Perl modules are not loaded. Term::Readline is not loaded, since +innotop doesn't prompt interactively. Term::ANSIColor and Win32::Console::ANSI +modules are not loaded. Term::ReadKey is still used, since innotop may have to +prompt for connection passwords when starting up. + +=item * + +innotop does not clear the screen after each tick. + +=item * + +innotop does not persist any changes to the configuration file. + +=item * + +If L<"--count"> is given and innotop is in incremental mode (see L<"status_inc"> +and L<"--inc">), innotop actually refreshes one more time than specified so it +can print incremental statistics. This suppresses output during the first +tick, so innotop may appear to hang. + +=item * + +innotop only displays the first table in each mode. This is so the output can +be easily processed with other command-line utilities such as awk and sed. To +change which tables display in each mode, see L<"TABLES">. Since L<"Q: Query +List"> mode is so important, innotop automatically disables the L<"q_header"> +table. This ensures you'll see the L<"processlist"> table, even if you have +innotop configured to show the q_header table during interactive operation. +Similarly, in L<"T: InnoDB Transactions"> mode, the L<"t_header"> table is +suppressed so you see only the L<"innodb_transactions"> table. + +=item * + +All output is tab-separated instead of being column-aligned with whitespace, and +innotop prints the full contents of each table instead of only printing one +screenful at a time. + +=item * + +innotop only prints column headers once instead of every tick (see +L<"hide_hdr">). innotop does not print table captions (see +L<"display_table_captions">). innotop ensures there are no empty lines in the +output. + +=item * + +innotop does not honor the L<"shorten"> transformation, which normally shortens +some numbers to human-readable formats. + +=item * + +innotop does not print a status line (see L<"INNOTOP STATUS">). + +=back + +=head1 CONFIGURING + +Nearly everything about innotop is configurable. Most things are possible to +change with built-in commands, but you can also edit the configuration file. + +While running innotop, press the '$' key to bring up the configuration editing +dialog. Press another key to select the type of data you want to edit: + +=over + +=item S: Statement Sleep Times + +Edits SQL statement sleep delays, which make innotop pause for the specified +amount of time after executing a statement. See L<"SQL STATEMENTS"> for a +definition of each statement and what it does. By default innotop does not +delay after any statements. + +This feature is included so you can customize the side-effects caused by +monitoring your server. You may not see any effects, but some innotop users +have noticed that certain MySQL versions under very high load with InnoDB +enabled take longer than usual to execute SHOW GLOBAL STATUS. If innotop calls +SHOW FULL PROCESSLIST immediately afterward, the processlist contains more +queries than the machine actually averages at any given moment. Configuring +innotop to pause briefly after calling SHOW GLOBAL STATUS alleviates this +effect. + +Sleep times are stored in the L<"stmt_sleep_times"> section of the configuration +file. Fractional-second sleeps are supported, subject to your hardware's +limitations. + +=item c: Edit Columns + +Starts the table editor on one of the displayed tables. See L<"TABLE EDITOR">. +An alternative way to start the table editor without entering the configuration +dialog is with the '^' key. + +=item g: General Configuration + +Starts the configuration editor to edit global and mode-specific configuration +variables (see L<"MODES">). innotop prompts you to choose a variable from among +the global and mode-specific ones depending on the current mode. + +=item k: Row-Coloring Rules + +Starts the row-coloring rules editor on one of the displayed table(s). See +L<"COLORS"> for details. + +=item p: Manage Plugins + +Starts the plugin configuration editor. See L<"PLUGINS"> for details. + +=item s: Server Groups + +Lets you create and edit server groups. See L<"SERVER GROUPS">. + +=item t: Choose Displayed Tables + +Lets you choose which tables to display in this mode. See L<"MODES"> and +L<"TABLES">. + +=back + +=head1 CONFIGURATION FILE + +innotop's default configuration file locations are $HOME/.innotop and +/etc/innotop/innotop.conf, and they are looked for in that order. If the first +configuration file exists, the second will not be processed. Those can be +overridden with the L<"--config"> command-line option. You can edit it by hand +safely, however innotop reads the configuration file when it starts, and, if +readonly is set to 0, writes it out again when it exits. Thus, if readonly is +set to 0, any changes you make by hand while innotop is running will be lost. + +innotop doesn't store its entire configuration in the configuration file. It +has a huge set of default configuration values that it holds only in memory, +and the configuration file only overrides these defaults. When you customize a +default setting, innotop notices, and then stores the customizations into the +file. This keeps the file size down, makes it easier to edit, and makes +upgrades easier. + +A configuration file is read-only be default. You can override that with +L<"--write">. See L<"readonly">. + +The configuration file is arranged into sections like an INI file. Each +section begins with [section-name] and ends with [/section-name]. Each +section's entries have a different syntax depending on the data they need to +store. You can put comments in the file; any line that begins with a # +character is a comment. innotop will not read the comments, so it won't write +them back out to the file when it exits. Comments in read-only configuration +files are still useful, though. + +The first line in the file is innotop's version number. This lets innotop +notice when the file format is not backwards-compatible, and upgrade smoothly +without destroying your customized configuration. + +The following list describes each section of the configuration file and the data +it contains: + +=over + +=item general + +The 'general' section contains global configuration variables and variables that +may be mode-specific, but don't belong in any other section. The syntax is a +simple key=value list. innotop writes a comment above each value to help you +edit the file by hand. + +=over + +=item S_func + +Controls S mode presentation (see L<"S: Variables & Status">). If g, values are +graphed; if s, values are like vmstat; if p, values are in a pivoted table. + +=item S_set + +Specifies which set of variables to display in L<"S: Variables & Status"> mode. +See L<"VARIABLE SETS">. + +=item auto_wipe_dl + +Instructs innotop to automatically wipe large deadlocks when it notices them. +When this happens you may notice a slight delay. At the next tick, you will +usually see the information that was being truncated by the large deadlock. + +=item charset + +Specifies what kind of characters to allow through the L<"no_ctrl_char"> +transformation. This keeps non-printable characters from confusing a +terminal when you monitor queries that contain binary data, such as images. + +The default is 'ascii', which considers anything outside normal ASCII to be a +control character. The other allowable values are 'unicode' and 'none'. 'none' +considers every character a control character, which can be useful for +collapsing ALL text fields in queries. + +=item cmd_filter + +This is the prefix that filters variables in L<"C: Command Summary"> mode. + +=item color + +Whether terminal coloring is permitted. + +=item cxn_timeout + +On MySQL versions 4.0.3 and newer, this variable is used to set the connection's +timeout, so MySQL doesn't close the connection if it is not used for a while. +This might happen because a connection isn't monitored in a particular mode, for +example. + +=item debug + +This option enables more verbose errors and makes innotop more strict in some +places. It can help in debugging filters and other user-defined code. It also +makes innotop write a lot of information to L<"debugfile"> when there is a +crash. + +=item debugfile + +A file to which innotop will write information when there is a crash. See +L<"FILES">. + +=item display_table_captions + +innotop displays a table caption above most tables. This variable suppresses or +shows captions on all tables globally. Some tables are configured with the +hide_caption property, which overrides this. + +=item global + +Whether to show GLOBAL variables and status. innotop only tries to do this on +servers which support the GLOBAL option to SHOW VARIABLES and SHOW STATUS. In +some MySQL versions, you need certain privileges to do this; if you don't have +them, innotop will not be able to fetch any variable and status data. This +configuration variable lets you run innotop and fetch what data you can even +without the elevated privileges. + +I can no longer find or reproduce the situation where GLOBAL wasn't allowed, but +I know there was one. + +=item graph_char + +Defines the character to use when drawing graphs in L<"S: Variables & Status"> +mode. + +=item header_highlight + +Defines how to highlight column headers. This only works if Term::ANSIColor is +available. Valid values are 'bold' and 'underline'. + +=item hide_hdr + +Hides column headers globally. + +=item interval + +The interval at which innotop will refresh its data (ticks). The interval is +implemented as a sleep time between ticks, so the true interval will vary +depending on how long it takes innotop to fetch and render data. + +This variable accepts fractions of a second. + +=item mode + +The mode in which innotop should start. Allowable arguments are the same as the +key presses that select a mode interactively. See L<"MODES">. + +=item num_digits + +How many digits to show in fractional numbers and percents. This variable's +range is between 0 and 9 and can be set directly from L<"S: Variables & Status"> +mode with the '+' and '-' keys. It is used in the L<"set_precision">, +L<"shorten">, and L<"percent"> transformations. + +=item num_status_sets + +Controls how many sets of status variables to display in pivoted L<"S: Variables +& Status"> mode. It also controls the number of old sets of variables innotop +keeps in its memory, so the larger this variable is, the more memory innotop +uses. + +=item plugin_dir + +Specifies where plugins can be found. By default, innotop stores plugins in the +'plugins' subdirectory of your innotop configuration directory. + +=item readonly + +Whether the configuration file is readonly. This cannot be set interactively. + +=item show_cxn_errors + +Makes innotop print connection errors to STDOUT. See L<"ERROR HANDLING">. + +=item show_cxn_errors_in_tbl + +Makes innotop display connection errors as rows in the first table on screen. +See L<"ERROR HANDLING">. + +=item show_percent + +Adds a '%' character after the value returned by the L<"percent"> +transformation. + +=item show_statusbar + +Controls whether to show the status bar in the display. See L<"INNOTOP +STATUS">. + +=item skip_innodb + +Disables fetching SHOW INNODB STATUS, in case your server(s) do not have InnoDB +enabled and you don't want innotop to try to fetch it. This can also be useful +when you don't have the SUPER privilege, required to run SHOW INNODB STATUS. + +=item spark + +Specifies how wide a spark chart is. There are two ASCII spark charts in A +mode, showing QPS and User_threads_running. + +=item status_inc + +Whether to show absolute or incremental values for status variables. +Incremental values are calculated as an offset from the last value innotop saw +for that variable. This is a global setting, but will probably become +mode-specific at some point. Right now it is honored a bit inconsistently; some +modes don't pay attention to it. + +=item timeformat + +The C-style strftime()-compatible format for the timestamp line to be printed +in -n mode when -t is set. + +=back + +=item plugins + +This section holds a list of package names of active plugins. If the plugin +exists, innotop will activate it. See L<"PLUGINS"> for more information. + +=item filters + +This section holds user-defined filters (see L<"FILTERS">). Each line is in the +format filter_name=text='filter text' tbls='table list'. + +The filter text is the text of the subroutine's code. The table list is a list +of tables to which the filter can apply. By default, user-defined filters apply +to the table for which they were created, but you can manually override that by +editing the definition in the configuration file. + +=item active_filters + +This section stores which filters are active on each table. Each line is in the +format table_name=filter_list. + +=item tbl_meta + +This section stores user-defined or user-customized columns (see L<"COLUMNS">). +Each line is in the format col_name=properties, where the properties are a +name=quoted-value list. + +=item connections + +This section holds the server connections you have defined. Each line is in +the format name=properties, where the properties are a name=value list. The +properties are self-explanatory, and the only one that is treated specially is +'pass' which is only present if 'savepass' is set. This section of the +configuration file will be skipped if any DSN, username, or password +command-line options are used. See L<"SERVER CONNECTIONS">. + +=item active_connections + +This section holds a list of which connections are active in each mode. Each +line is in the format mode_name=connection_list. + +=item server_groups + +This section holds server groups. Each line is in the format +name=connection_list. See L<"SERVER GROUPS">. + +=item active_server_groups + +This section holds a list of which server group is active in each mode. Each +line is in the format mode_name=server_group. + +=item max_values_seen + +This section holds the maximum values seen for variables. This is used to scale +the graphs in L<"S: Variables & Status"> mode. Each line is in the format +name=value. + +=item active_columns + +This section holds table column lists. Each line is in the format +tbl_name=column_list. See L<"COLUMNS">. + +=item sort_cols + +This section holds the sort definition. Each line is in the format +tbl_name=column_list. If a column is prefixed with '-', that column sorts +descending. See L<"SORTING">. + +=item visible_tables + +This section defines which tables are visible in each mode. Each line is in the +format mode_name=table_list. See L<"TABLES">. + +=item varsets + +This section defines variable sets for use in L<"S: Status & Variables"> mode. +Each line is in the format name=variable_list. See L<"VARIABLE SETS">. + +=item colors + +This section defines colorization rules. Each line is in the format +tbl_name=property_list. See L<"COLORS">. + +=item stmt_sleep_times + +This section contains statement sleep times. Each line is in the format +statement_name=sleep_time. See L<"S: Statement Sleep Times">. + +=item group_by + +This section contains column lists for table group_by expressions. Each line is +in the format tbl_name=column_list. See L<"GROUPING">. + +=back + +=head1 CUSTOMIZING + +You can customize innotop a great deal. For example, you can: + +=over + +=item * + +Choose which tables to display, and in what order. + +=item * + +Choose which columns are in those tables, and create new columns. + +=item * + +Filter which rows display with built-in filters, user-defined filters, and +quick-filters. + +=item * + +Sort the rows to put important data first or group together related rows. + +=item * + +Highlight rows with color. + +=item * + +Customize the alignment, width, and formatting of columns, and apply +transformations to columns to extract parts of their values or format the values +as you wish (for example, shortening large numbers to familiar units). + +=item * + +Design your own expressions to extract and combine data as you need. This gives +you unlimited flexibility. + +=back + +All these and more are explained in the following sections. + +=head2 TABLES + +A table is what you'd expect: a collection of columns. It also has some other +properties, such as a caption. Filters, sorting rules, and colorization rules +belong to tables and are covered in later sections. + +Internally, table meta-data is defined in a data structure called %tbl_meta. +This hash holds all built-in table definitions, which contain a lot of default +instructions to innotop. The meta-data includes the caption, a list of columns +the user has customized, a list of columns, a list of visible columns, a list of +filters, color rules, a sort-column list, sort direction, and some information +about the table's data sources. Most of this is customizable via the table +editor (see L<"TABLE EDITOR">). + +You can choose which tables to show by pressing the '$' key. See L<"MODES"> and +L<"TABLES">. + +The table life-cycle is as follows: + +=over + +=item * + +Each table begins with a data source, which is an array of hashes. See below +for details on data sources. + +=item * + +Each element of the data source becomes a row in the final table. + +=item * + +For each element in the data source, innotop extracts values from the source and +creates a row. This row is another hash, which later steps will refer to as +$set. The values innotop extracts are determined by the table's columns. Each +column has an extraction subroutine, compiled from an expression (see +L<"EXPRESSIONS">). The resulting row is a hash whose keys are named the same as +the column name. + +=item * + +innotop filters the rows, removing those that don't need to be displayed. See +L<"FILTERS">. + +=item * + +innotop sorts the rows. See L<"SORTING">. + +=item * + +innotop groups the rows together, if specified. See L<"GROUPING">. + +=item * + +innotop colorizes the rows. See L<"COLORS">. + +=item * + +innotop transforms the column values in each row. See L<"TRANSFORMATIONS">. + +=item * + +innotop optionally pivots the rows (see L<"PIVOTING">), then filters and sorts +them. + +=item * + +innotop formats and justifies the rows as a table. During this step, innotop +applies further formatting to the column values, including alignment, maximum +and minimum widths. innotop also does final error checking to ensure there are +no crashes due to undefined values. innotop then adds a caption if specified, +and the table is ready to print. + +=back + +The lifecycle is slightly different if the table is pivoted, as noted above. To +clarify, if the table is pivoted, the process is extract, group, transform, +pivot, filter, sort, create. If it's not pivoted, the process is extract, +filter, sort, group, color, transform, create. This slightly convoluted process +doesn't map all that well to SQL, but pivoting complicates things pretty +thoroughly. Roughly speaking, filtering and sorting happen as late as needed to +effect the final result as you might expect, but as early as possible for +efficiency. + +Each built-in table is described below: + +=over + +=item adaptive_hash_index + +Displays data about InnoDB's adaptive hash index. Data source: +L<"STATUS_VARIABLES">. + +=item buffer_pool + +Displays data about InnoDB's buffer pool. Data source: L<"STATUS_VARIABLES">. + +=item cmd_summary + +Displays weighted status variables. Data source: L<"STATUS_VARIABLES">. + +=item deadlock_locks + +Shows which locks were held and waited for by the last detected deadlock. Data +source: L<"DEADLOCK_LOCKS">. + +=item deadlock_transactions + +Shows transactions involved in the last detected deadlock. Data source: +L<"DEADLOCK_TRANSACTIONS">. + +=item explain + +Shows the output of EXPLAIN. Data source: L<"EXPLAIN">. + +=item file_io_misc + +Displays data about InnoDB's file and I/O operations. Data source: +L<"STATUS_VARIABLES">. + +=item fk_error + +Displays various data about InnoDB's last foreign key error. Data source: +L<"STATUS_VARIABLES">. + +=item health_dashboard + +Displays an overall summary of servers, one server per line, for monitoring. +Data source: L<"STATUS_VARIABLES">, L<"MASTER_SLAVE">, L<"PROCESSLIST_STATS">. + +=item index_statistics + +Displays data from the INDEX_STATISTICS table in Percona-enhanced servers. + +=item index_table_statistics + +Displays data from the INDEX_STATISTICS and TABLE_STATISTICS tables in +Percona-enhanced servers. It joins the two together, grouped by the database +and table name. It is the default view in L<"U: User Statistics"> mode, +and makes it easy to see what tables are hot, how many rows are read from indexes, +how many changes are made, and how many changes are made to indexes. + +=item innodb_blocked_blocker + +Displays InnoDB locks and lock waits. Data source: L<"INNODB_BLOCKED_BLOCKER">. + +=item innodb_locks + +Displays InnoDB locks. Data source: L<"INNODB_LOCKS">. + +=item innodb_transactions + +Displays data about InnoDB's current transactions. Data source: +L<"INNODB_TRANSACTIONS">. + +=item insert_buffers + +Displays data about InnoDB's insert buffer. Data source: L<"STATUS_VARIABLES">. + +=item io_threads + +Displays data about InnoDB's I/O threads. Data source: L<"IO_THREADS">. + +=item log_statistics + +Displays data about InnoDB's logging system. Data source: L<"STATUS_VARIABLES">. + +=item master_status + +Displays replication master status. Data source: L<"STATUS_VARIABLES">. + +=item open_tables + +Displays open tables. Data source: L<"OPEN_TABLES">. + +=item page_statistics + +Displays InnoDB page statistics. Data source: L<"STATUS_VARIABLES">. + +=item pending_io + +Displays InnoDB pending I/O operations. Data source: L<"STATUS_VARIABLES">. + +=item processlist + +Displays current MySQL processes (threads/connections). Data source: +L<"PROCESSLIST">. + +=item q_header + +Displays various status values. Data source: L<"STATUS_VARIABLES">. + +=item row_operation_misc + +Displays data about InnoDB's row operations. Data source: +L<"STATUS_VARIABLES">. + +=item row_operations + +Displays data about InnoDB's row operations. Data source: +L<"STATUS_VARIABLES">. + +=item semaphores + +Displays data about InnoDB's semaphores and mutexes. Data source: +L<"STATUS_VARIABLES">. + +=item slave_io_status + +Displays data about the slave I/O thread. Data source: +L<"STATUS_VARIABLES">. + +=item slave_sql_status + +Displays data about the slave SQL thread. Data source: L<"STATUS_VARIABLES">. + +=item table_statistics + +Displays data from the TABLE_STATISTICS table in Percona-enhanced servers. + +=item t_header + +Displays various InnoDB status values. Data source: L<"STATUS_VARIABLES">. + +=item var_status + +Displays user-configurable data. Data source: L<"STATUS_VARIABLES">. + +=item wait_array + +Displays data about InnoDB's OS wait array. Data source: L<"OS_WAIT_ARRAY">. + +=back + +=head2 COLUMNS + +Columns belong to tables. You can choose a table's columns by pressing the '^' +key, which starts the L<"TABLE EDITOR"> and lets you choose and edit columns. +Pressing 'e' from within the table editor lets you edit the column's properties: + +=over + +=item * + +hdr: a column header. This appears in the first row of the table. + +=item * + +just: justification. '-' means left-justified and '' means right-justified, +just as with printf formatting codes (not a coincidence). + +=item * + +dec: whether to further align the column on the decimal point. + +=item * + +num: whether the column is numeric. This affects how values are sorted +(lexically or numerically). + +=item * + +label: a small note about the column, which appears in dialogs that help the +user choose columns. + +=item * + +src: an expression that innotop uses to extract the column's data from its +source (see L<"DATA SOURCES">). See L<"EXPRESSIONS"> for more on expressions. + +=item * + +minw: specifies a minimum display width. This helps stabilize the display, +which makes it easier to read if the data is changing frequently. + +=item * + +maxw: similar to minw. + +=item * + +trans: a list of column transformations. See L<"TRANSFORMATIONS">. + +=item * + +agg: an aggregate function. See L<"GROUPING">. The default is L<"first">. + +=item * + +aggonly: controls whether the column only shows when grouping is enabled on the +table (see L<"GROUPING">). By default, this is disabled. This means columns +will always be shown by default, whether grouping is enabled or not. If a +column's aggonly is set true, the column will appear when you toggle grouping on +the table. Several columns are set this way, such as the count column on +L<"processlist"> and L<"innodb_transactions">, so you don't see a count when the +grouping isn't enabled, but you do when it is. + +=item * + +agghide: the reverse of aggonly. The column is hidden when grouping is enabled. + +=back + +=head2 FILTERS + +Filters remove rows from the display. They behave much like a WHERE clause in +SQL. innotop has several built-in filters, which remove irrelevant information +like inactive queries, but you can define your own as well. innotop also lets +you create quick-filters, which do not get saved to the configuration file, and +are just an easy way to quickly view only some rows. + +You can enable or disable a filter on any table. Press the '%' key (mnemonic: % +looks kind of like a line being filtered between two circles) and choose which +table you want to filter, if asked. You'll then see a list of possible filters +and a list of filters currently enabled for that table. Type the names of +filters you want to apply and press Enter. + +=head3 USER-DEFINED FILTERS + +If you type a name that doesn't exist, innotop will prompt you to create the +filter. Filters are easy to create if you know Perl, and not hard if you don't. +What you're doing is creating a subroutine that returns true if the row should +be displayed. The row is a hash reference passed to your subroutine as $set. + +For example, imagine you want to filter the processlist table so you only see +queries that have been running more than five minutes. Type a new name for your +filter, and when prompted for the subroutine body, press TAB to initiate your +terminal's auto-completion. You'll see the names of the columns in the +L<"processlist"> table (innotop generally tries to help you with auto-completion +lists). You want to filter on the 'time' column. Type the text "$set->{time} > +300" to return true when the query is more than five minutes old. That's all +you need to do. + +In other words, the code you're typing is surrounded by an implicit context, +which looks like this: + + sub filter { + my ( $set ) = @_; + # YOUR CODE HERE + } + +If your filter doesn't work, or if something else suddenly behaves differently, +you might have made an error in your filter, and innotop is silently catching +the error. Try enabling L<"debug"> to make innotop throw an error instead. + +=head3 QUICK-FILTERS + +innotop's quick-filters are a shortcut to create a temporary filter that doesn't +persist when you restart innotop. To create a quick-filter, press the '/' key. +innotop will prompt you for the column name and filter text. Again, you can use +auto-completion on column names. The filter text can be just the text you want +to "search for." For example, to filter the L<"processlist"> table on queries +that refer to the products table, type '/' and then 'info product'. Internally, +the filter is compiled into a subroutine like this: + + sub filter { + my ( $set ) = @_; + $set->{info} =~ m/product/; + } + +The filter text can actually be any Perl regular expression, but of course a +literal string like 'product' works fine as a regular expression. + +What if you want the filter to discard matching rows, rather than showing +matching rows? If you're familiar with Perl regular expressions, you might +guess how to do this. You have to use a zero-width negative lookahead +assertion. If you don't know what that means, don't worry. Let's filter out +all rows where the command is Gandalf. Type the following: + + 1. / + 2. cmd ^(?!Gandalf) + +Behind the scenes innotop compiles the quick-filter into a specially tagged +filter that is otherwise like any other filter. It just isn't saved to the +configuration file. + +To clear quick-filters, press the '\' key and innotop will clear them all at +once. + +=head2 SORTING + +innotop has sensible built-in defaults to sort the most important rows to the +top of the table. Like anything else in innotop, you can customize how any +table is sorted. + +To start the sort dialog, start the L<"TABLE EDITOR"> with the '^' key, choose a +table if necessary, and press the 's' key. You'll see a list of columns you can +use in the sort expression and the current sort expression, if any. Enter a +list of columns by which you want to sort and press Enter. If you want to +reverse sort, prefix the column name with a minus sign. For example, if you +want to sort by column a ascending, then column b descending, type 'a -b'. You +can also explicitly add a + in front of columns you want to sort ascending, but +it's not required. + +Some modes have keys mapped to open this dialog directly, and to quickly reverse +sort direction. Press '?' as usual to see which keys are mapped in any mode. + +=head2 GROUPING + +innotop can group, or aggregate, rows together (the terms are used +interchangeably). This is quite similar to an SQL GROUP BY clause. You can +specify to group on certain columns, or if you don't specify any, the entire set +of rows is treated as one group. This is quite like SQL so far, but unlike SQL, +you can also select un-grouped columns. innotop actually aggregates every +column. If you don't explicitly specify a grouping function, the default is +'first'. This is basically a convenience so you don't have to specify an +aggregate function for every column you want in the result. + +You can quickly toggle grouping on a table with the '=' key, which toggles its +aggregate property. This property doesn't persist to the config file. + +The columns by which the table is grouped are specified in its group_by +property. When you turn grouping on, innotop places the group_by columns at the +far left of the table, even if they're not supposed to be visible. The rest of +the visible columns appear in order after them. + +Two tables have default group_by lists and a count column built in: +L<"processlist"> and L<"innodb_transactions">. The grouping is by connection +and status, so you can quickly see how many queries or transactions are in a +given status on each server you're monitoring. The time columns are aggregated +as a sum; other columns are left at the default 'first' aggregation. + +By default, the table shown in L<"S: Variables & Status"> mode also uses +grouping so you can monitor variables and status across many servers. The +default aggregation function in this mode is 'avg'. + +Valid grouping functions are defined in the %agg_funcs hash. They include + +=over + +=item first + +Returns the first element in the group. + +=item count + +Returns the number of elements in the group, including undefined elements, much +like SQL's COUNT(*). + +=item avg + +Returns the average of defined elements in the group. + +=item sum + +Returns the sum of elements in the group. + +=back + +Here's an example of grouping at work. Suppose you have a very busy server with +hundreds of open connections, and you want to see how many connections are in +what status. Using the built-in grouping rules, you can press 'Q' to enter +L<"Q: Query List"> mode. Press '=' to toggle grouping (if necessary, select the +L<"processlist"> table when prompted). + +Your display might now look like the following: + + Query List (? for help) localhost, 32:33, 0.11 QPS, 1 thd, 5.0.38-log + + CXN Cmd Cnt ID User Host Time Query + localhost Query 49 12933 webusr localhost 19:38 SELECT * FROM + localhost Sending Da 23 2383 webusr localhost 12:43 SELECT col1, + localhost Sleep 120 140 webusr localhost 5:18:12 + localhost Statistics 12 19213 webusr localhost 01:19 SELECT * FROM + +That's actually quite a worrisome picture. You've got a lot of idle connections +(Sleep), and some connections executing queries (Query and Sending Data). +That's okay, but you also have a lot in Statistics status, collectively spending +over a minute. That means the query optimizer is having a really hard time +generating execution plans for your statements. Something is wrong; it should +normally take milliseconds to plan queries. You might not have seen this pattern if you +didn't look at your connections in aggregate. (This is a made-up example, but +it can happen in real life). + +=head2 PIVOTING + +innotop can pivot a table for more compact display, similar to a Pivot Table in +a spreadsheet (also known as a crosstab). Pivoting a table makes columns into +rows. Assume you start with this table: + + foo bar + === === + 1 3 + 2 4 + +After pivoting, the table will look like this: + + name set0 set1 + ==== ==== ==== + foo 1 2 + bar 3 4 + +To get reasonable results, you might need to group as well as pivoting. +innotop currently does this for L<"S: Variables & Status"> mode. + +=head2 COLORS + +By default, innotop highlights rows with color so you can see at a glance which +rows are more important. You can customize the colorization rules and add your +own to any table. Open the table editor with the '^' key, choose a table if +needed, and press 'o' to open the color editor dialog. + +The color editor dialog displays the rules applied to the table, in the order +they are evaluated. Each row is evaluated against each rule to see if the rule +matches the row; if it does, the row gets the specified color, and no further +rules are evaluated. The rules look like the following: + + state eq Locked black on_red + cmd eq Sleep white + user eq system user white + cmd eq Connect white + cmd eq Binlog Dump white + time > 600 red + time > 120 yellow + time > 60 green + time > 30 cyan + +This is the default rule set for the L<"processlist"> table. In order of +priority, these rules make locked queries black on a red background, "gray out" +connections from replication and sleeping queries, and make queries turn from +cyan to red as they run longer. + +(For some reason, the ANSI color code "white" is actually a light gray. Your +terminal's display may vary; experiment to find colors you like). + +You can use keystrokes to move the rules up and down, which re-orders their +priority. You can also delete rules and add new ones. If you add a new rule, +innotop prompts you for the column, an operator for the comparison, a value +against which to compare the column, and a color to assign if the rule matches. +There is auto-completion and prompting at each step. + +The value in the third step needs to be correctly quoted. innotop does not try +to quote the value because it doesn't know whether it should treat the value as +a string or a number. If you want to compare the column against a string, as +for example in the first rule above, you should enter 'Locked' surrounded by +quotes. If you get an error message about a bareword, you probably should have +quoted something. + +=head2 EXPRESSIONS + +Expressions are at the core of how innotop works, and are what enables you to +extend innotop as you wish. Recall the table lifecycle explained in +L<"TABLES">. Expressions are used in the earliest step, where it extracts +values from a data source to form rows. + +It does this by calling a subroutine for each column, passing it the source data +set, a set of current values, and a set of previous values. These are all +needed so the subroutine can calculate things like the difference between this +tick and the previous tick. + +The subroutines that extract the data from the set are compiled from +expressions. This gives significantly more power than just naming the values to +fill the columns, because it allows the column's value to be calculated from +whatever data is necessary, but avoids the need to write complicated and lengthy +Perl code. + +innotop begins with a string of text that can look as simple as a value's name +or as complicated as a full-fledged Perl expression. It looks at each +'bareword' token in the string and decides whether it's supposed to be a key +into the $set hash. A bareword is an unquoted value that isn't already +surrounded by code-ish things like dollar signs or curly brackets. If innotop +decides that the bareword isn't a function or other valid Perl code, it converts +it into a hash access. After the whole string is processed, innotop compiles a +subroutine, like this: + + sub compute_column_value { + my ( $set, $cur, $pre ) = @_; + my $val = # EXPANDED STRING GOES HERE + return $val; + } + +Here's a concrete example, taken from the header table L<"q_header"> in L<"Q: +Query List"> mode. This expression calculates the qps, or Queries Per Second, +column's values, from the values returned by SHOW STATUS: + + Questions/Uptime_hires + +innotop decides both words are barewords, and transforms this expression into +the following Perl code: + + $set->{Questions}/$set->{Uptime_hires} + +When surrounded by the rest of the subroutine's code, this is executable Perl +that calculates a high-resolution queries-per-second value. + +The arguments to the subroutine are named $set, $cur, and $pre. In most cases, +$set and $cur will be the same values. However, if L<"status_inc"> is set, $cur +will not be the same as $set, because $set will already contain values that are +the incremental difference between $cur and $pre. + +Every column in innotop is computed by subroutines compiled in the same fashion. +There is no difference between innotop's built-in columns and user-defined +columns. This keeps things consistent and predictable. + +=head2 TRANSFORMATIONS + +Transformations change how a value is rendered. For example, they can take a +number of seconds and display it in H:M:S format. The following transformations +are defined: + +=over + +=item commify + +Adds commas to large numbers every three decimal places. + +=item distill + +Distills SQL into verb-noun-noun format for quick comprehension. + +=item dulint_to_int + +Accepts two unsigned integers and converts them into a single longlong. This is +useful for certain operations with InnoDB, which uses two integers as +transaction identifiers, for example. + +=item fuzzy_time + +Converts a number of seconds into a friendly, readable value like "1h35m". + +=item no_ctrl_char + +Removes quoted control characters from the value. This is affected by the +L<"charset"> configuration variable. + +This transformation only operates within quoted strings, for example, values to +a SET clause in an UPDATE statement. It will not alter the UPDATE statement, +but will collapse the quoted string to [BINARY] or [TEXT], depending on the +charset. + +=item percent + +Converts a number to a percentage by multiplying it by two, formatting it with +L<"num_digits"> digits after the decimal point, and optionally adding a percent +sign (see L<"show_percent">). + +=item secs_to_time + +Formats a number of seconds as time in days+hours:minutes:seconds format. + +=item set_precision + +Formats numbers with L<"num_digits"> number of digits after the decimal point. + +=item shorten + +Formats a number as a unit of 1024 (k/M/G/T) and with L<"num_digits"> number of +digits after the decimal point. + +=back + +=head2 TABLE EDITOR + +The innotop table editor lets you customize tables with keystrokes. You start +the table editor with the '^' key. If there's more than one table on the +screen, it will prompt you to choose one of them. Once you do, innotop will +show you something like this: + + Editing table definition for Buffer Pool. Press ? for help, q to quit. + + name hdr label src + cxn CXN Connection from which cxn + buf_pool_size Size Buffer pool size IB_bp_buf_poo + buf_free Free Bufs Buffers free in the b IB_bp_buf_fre + pages_total Pages Pages total IB_bp_pages_t + pages_modified Dirty Pages Pages modified (dirty IB_bp_pages_m + buf_pool_hit_rate Hit Rate Buffer pool hit rate IB_bp_buf_poo + total_mem_alloc Memory Total memory allocate IB_bp_total_m + add_pool_alloc Add'l Pool Additional pool alloca IB_bp_add_poo + +The first line shows which table you're editing, and reminds you again to press +'?' for a list of key mappings. The rest is a tabular representation of the +table's columns, because that's likely what you're trying to edit. However, you +can edit more than just the table's columns; this screen can start the filter +editor, color rule editor, and more. + +Each row in the display shows a single column in the table you're editing, along +with a couple of its properties such as its header and source expression (see +L<"EXPRESSIONS">). + +The key mappings are Vim-style, as in many other places. Pressing 'j' and 'k' +moves the highlight up or down. You can then (d)elete or (e)dit the highlighted +column. You can also (a)dd a column to the table. This actually just activates +one of the columns already defined for the table; it prompts you to choose from +among the columns available but not currently displayed. Finally, you can +re-order the columns with the '+' and '-' keys. + +You can do more than just edit the columns with the table editor, you can also +edit other properties, such as the table's sort expression and group-by +expression. Press '?' to see the full list, of course. + +If you want to really customize and create your own column, as opposed to just +activating a built-in one that's not currently displayed, press the (n)ew key, +and innotop will prompt you for the information it needs: + +=over + +=item * + +The column name: this needs to be a word without any funny characters, e.g. just +letters, numbers and underscores. + +=item * + +The column header: this is the label that appears at the top of the column, in +the table header. This can have spaces and funny characters, but be careful not +to make it too wide and waste space on-screen. + +=item * + +The column's data source: this is an expression that determines what data from +the source (see L<"TABLES">) innotop will put into the column. This can just be +the name of an item in the source, or it can be a more complex expression, as +described in L<"EXPRESSIONS">. + +=back + +Once you've entered the required data, your table has a new column. There is no +difference between this column and the built-in ones; it can have all the same +properties and behaviors. innotop will write the column's definition to the +configuration file, so it will persist across sessions. + +Here's an example: suppose you want to track how many times your slaves have +retried transactions. According to the MySQL manual, the +Slave_retried_transactions status variable gives you that data: "The total +number of times since startup that the replication slave SQL thread has retried +transactions. This variable was added in version 5.0.4." This is appropriate to +add to the L<"slave_sql_status"> table. + +To add the column, switch to the replication-monitoring mode with the 'M' key, +and press the '^' key to start the table editor. When prompted, choose +slave_sql_status as the table, then press 'n' to create the column. Type +'retries' as the column name, 'Retries' as the column header, and +'Slave_retried_transactions' as the source. Now the column is created, and you +see the table editor screen again. Press 'q' to exit the table editor, and +you'll see your column at the end of the table. + +=head1 VARIABLE SETS + +Variable sets are used in L<"S: Variables & Status"> mode to define more easily +what variables you want to monitor. Behind the scenes they are compiled to a +list of expressions, and then into a column list so they can be treated just +like columns in any other table, in terms of data extraction and +transformations. However, you're protected from the tedious details by a syntax +that ought to feel very natural to you: a SQL SELECT list. + +The data source for variable sets, and indeed the entire S mode, is the +combination of SHOW STATUS, SHOW VARIABLES, and SHOW INNODB STATUS. Imagine +that you had a huge table with one column per variable returned from those +statements. That's the data source for variable sets. You can now query this +data source just like you'd expect. For example: + + Questions, Uptime, Questions/Uptime as QPS + +Behind the scenes innotop will split that variable set into three expressions, +compile them and turn them into a table definition, then extract as usual. This +becomes a "variable set," or a "list of variables you want to monitor." + +innotop lets you name and save your variable sets, and writes them to the +configuration file. You can choose which variable set you want to see with the +'c' key, or activate the next and previous sets with the '>' and '<' keys. +There are many built-in variable sets as well, which should give you a good +start for creating your own. Press 'e' to edit the current variable set, or +just to see how it's defined. To create a new one, just press 'c' and type its +name. + +You may want to use some of the functions listed in L<"TRANSFORMATIONS"> to help +format the results. In particular, L<"set_precision"> is often useful to limit +the number of digits you see. Extending the above example, here's how: + + Questions, Uptime, set_precision(Questions/Uptime) as QPS + +Actually, this still needs a little more work. If your L<"interval"> is less +than one second, you might be dividing by zero because Uptime is incremental in +this mode by default. Instead, use Uptime_hires: + + Questions, Uptime, set_precision(Questions/Uptime_hires) as QPS + +This example is simple, but it shows how easy it is to choose which variables +you want to monitor. + +=head1 PLUGINS + +innotop has a simple but powerful plugin mechanism by which you can extend +or modify its existing functionality, and add new functionality. innotop's +plugin functionality is event-based: plugins register themselves to be called +when events happen. They then have a chance to influence the event. + +An innotop plugin is a Perl module (.pm) file placed in innotop's L<"plugin_dir"> +directory. On UNIX systems, you can place a symbolic link to the module instead +of putting the actual file there. innotop automatically discovers files named C<*.pm>. If +there is a corresponding entry in the L<"plugins"> configuration file section, +innotop loads and activates the plugin. + +The module must conform to innotop's plugin interface. Additionally, the source +code of the module must be written in such a way that innotop can inspect the +file and determine the package name and description. + +=head2 Package Source Convention + +innotop inspects the plugin module's source to determine the Perl package name. +It looks for a line of the form "package Foo;" and if found, considers the +plugin's package name to be Foo. Of course the package name can be a valid Perl +package name such as Foo::Bar, with double colons (::) and so on. + +It also looks for a description in the source code, to make the plugin editor +more human-friendly. The description is a comment line of the form "# +description: Foo", where "Foo" is the text innotop will consider to be the +plugin's description. + +=head2 Plugin Interface + +The innotop plugin interface is quite simple: innotop expects the plugin to be +an object-oriented module it can call certain methods on. The methods are + +=over + +=item new(%variables) + +This is the plugin's constructor. It is passed a hash of innotop's variables, +which it can manipulate (see L<"Plugin Variables">). It must return a reference +to the newly created plugin object. + +At construction time, innotop has only loaded the general configuration and +created the default built-in variables with their default contents (which is +quite a lot). Therefore, the state of the program is exactly as in the innotop +source code, plus the configuration variables from the L<"general"> section in +the config file. + +If your plugin manipulates the variables, it is changing global data, which is +shared by innotop and all plugins. Plugins are loaded in the order they're +listed in the config file. Your plugin may load before or after another plugin, +so there is a potential for conflict or interaction between plugins if they +modify data other plugins use or modify. + +=item register_for_events() + +This method must return a list of events in which the plugin is interested, if +any. See L<"Plugin Events"> for the defined events. If the plugin returns an +event that's not defined, the event is ignored. + +=item event handlers + +The plugin must implement a method named the same as each event for which it has +registered. In other words, if the plugin returns qw(foo bar) from +register_for_events(), it must have foo() and bar() methods. These methods are +callbacks for the events. See L<"Plugin Events"> for more details about each +event. + +=back + +=head2 Plugin Variables + +The plugin's constructor is passed a hash of innotop's variables, which it can +manipulate. It is probably a good idea if the plugin object saves a copy of it +for later use. The variables are defined in the innotop variable +%pluggable_vars, and are as follows: + +=over + +=item action_for + +A hashref of key mappings. These are innotop's global hot-keys. + +=item agg_funcs + +A hashref of functions that can be used for grouping. See L<"GROUPING">. + +=item config + +The global configuration hash. + +=item connections + +A hashref of connection specifications. These are just specifications of how to +connect to a server. + +=item dbhs + +A hashref of innotop's database connections. These are actual DBI connection +objects. + +=item filters + +A hashref of filters applied to table rows. See L<"FILTERS"> for more. + +=item modes + +A hashref of modes. See L<"MODES"> for more. + +=item server_groups + +A hashref of server groups. See L<"SERVER GROUPS">. + +=item tbl_meta + +A hashref of innotop's table meta-data, with one entry per table (see +L<"TABLES"> for more information). + +=item trans_funcs + +A hashref of transformation functions. See L<"TRANSFORMATIONS">. + +=item var_sets + +A hashref of variable sets. See L<"VARIABLE SETS">. + +=back + +=head2 Plugin Events + +Each event is defined somewhere in the innotop source code. When innotop runs +that code, it executes the callback function for each plugin that expressed its +interest in the event. innotop passes some data for each event. The events are +defined in the %event_listener_for variable, and are as follows: + +=over + +=item extract_values($set, $cur, $pre, $tbl) + +This event occurs inside the function that extracts values from a data source. +The arguments are the set of values, the current values, the previous values, +and the table name. + +=item set_to_tbl + +Events are defined at many places in this subroutine, which is responsible for +turning an arrayref of hashrefs into an arrayref of lines that can be printed to +the screen. The events all pass the same data: an arrayref of rows and the name +of the table being created. The events are set_to_tbl_pre_filter, +set_to_tbl_pre_sort,set_to_tbl_pre_group, set_to_tbl_pre_colorize, +set_to_tbl_pre_transform, set_to_tbl_pre_pivot, set_to_tbl_pre_create, +set_to_tbl_post_create. + +=item draw_screen($lines) + +This event occurs inside the subroutine that prints the lines to the screen. +$lines is an arrayref of strings. + +=back + +=head2 Simple Plugin Example + +The easiest way to explain the plugin functionality is probably with a simple +example. The following module adds a column to the beginning of every table and +sets its value to 1. (If you copy and paste this example code, be sure to remove +the first space from each line; lines such as '# description' must not start with +whitespace). + + use strict; + use warnings FATAL => 'all'; + + package Innotop::Plugin::Example; + # description: Adds an 'example' column to every table + + sub new { + my ( $class, %vars ) = @_; + # Store reference to innotop's variables in $self + my $self = bless { %vars }, $class; + + # Design the example column + my $col = { + hdr => 'Example', + just => '', + dec => 0, + num => 1, + label => 'Example', + src => 'example', # Get data from this column in the data source + tbl => '', + trans => [], + }; + + # Add the column to every table. + my $tbl_meta = $vars{tbl_meta}; + foreach my $tbl ( values %$tbl_meta ) { + # Add the column to the list of defined columns + $tbl->{cols}->{example} = $col; + # Add the column to the list of visible columns + unshift @{$tbl->{visible}}, 'example'; + } + + # Be sure to return a reference to the object. + return $self; + } + + # I'd like to be called when a data set is being rendered into a table, please. + sub register_for_events { + my ( $self ) = @_; + return qw(set_to_tbl_pre_filter); + } + + # This method will be called when the event fires. + sub set_to_tbl_pre_filter { + my ( $self, $rows, $tbl ) = @_; + # Set the example column's data source to the value 1. + foreach my $row ( @$rows ) { + $row->{example} = 1; + } + } + + 1; + +=head2 Plugin Editor + +The plugin editor lets you view the plugins innotop discovered and activate or +deactivate them. Start the editor by pressing $ to start the configuration +editor from any mode. Press the 'p' key to start the plugin editor. You'll see +a list of plugins innotop discovered. You can use the 'j' and 'k' keys to move +the highlight to the desired one, then press the * key to toggle it active or +inactive. Exit the editor and restart innotop for the changes to take effect. + +=head1 SQL STATEMENTS + +innotop uses a limited set of SQL statements to retrieve data from MySQL for +display. The statements are customized depending on the server version against +which they are executed; for example, on MySQL 5 and newer, INNODB_STATUS +executes "SHOW ENGINE INNODB STATUS", while on earlier versions it executes +"SHOW INNODB STATUS". The statements are as follows: + + Statement SQL executed + =================== =============================== + INDEX_STATISTICS SELECT * FROM INFORMATION_SCHEMA.INDEX_STATISTICS + INNODB_STATUS SHOW [ENGINE] INNODB STATUS + KILL_CONNECTION KILL + KILL_QUERY KILL QUERY + OPEN_TABLES SHOW OPEN TABLES + PROCESSLIST SHOW FULL PROCESSLIST + SHOW_MASTER_LOGS SHOW MASTER LOGS + SHOW_MASTER_STATUS SHOW MASTER STATUS + SHOW_SLAVE_STATUS SHOW SLAVE STATUS + SHOW_STATUS SHOW [GLOBAL] STATUS + SHOW_VARIABLES SHOW [GLOBAL] VARIABLES + TABLE_STATISTICS SELECT * FROM INFORMATION_SCHEMA.TABLE_STATISTICS + +=head1 DATA SOURCES + +Each time innotop extracts values to create a table (see L<"EXPRESSIONS"> and +L<"TABLES">), it does so from a particular data source. Largely because of the +complex data extracted from SHOW INNODB STATUS, this is slightly messy. SHOW +INNODB STATUS contains a mixture of single values and repeated values that form +nested data sets. + +Whenever innotop fetches data from MySQL, it adds two extra bits to each set: +cxn and Uptime_hires. cxn is the name of the connection from which the data +came. Uptime_hires is a high-resolution version of the server's Uptime status +variable, which is important if your L<"interval"> setting is sub-second. + +Here are the kinds of data sources from which data is extracted: + +=over + +=item STATUS_VARIABLES + +This is the broadest category, into which the most kinds of data fall. It +begins with the combination of SHOW STATUS and SHOW VARIABLES, but other sources +may be included as needed, for example, SHOW MASTER STATUS and SHOW SLAVE +STATUS, as well as many of the non-repeated values from SHOW INNODB STATUS. + +=item DEADLOCK_LOCKS + +This data is extracted from the transaction list in the LATEST DETECTED DEADLOCK +section of SHOW INNODB STATUS. It is nested two levels deep: transactions, then +locks. + +=item DEADLOCK_TRANSACTIONS + +This data is from the transaction list in the LATEST DETECTED DEADLOCK +section of SHOW INNODB STATUS. It is nested one level deep. + +=item EXPLAIN + +This data is from the result set returned by EXPLAIN. + +=item INNODB_BLOCKED_BLOCKER + +This data is from the INFORMATION_SCHEMA tables related to InnoDB locks and +the processlist. + +=item INNODB_TRANSACTIONS + +This data is from the TRANSACTIONS section of SHOW INNODB STATUS. + +=item IO_THREADS + +This data is from the list of threads in the FILE I/O section of SHOW INNODB +STATUS. + +=item INNODB_LOCKS + +This data is from the TRANSACTIONS section of SHOW INNODB STATUS and is nested +two levels deep. + +=item MASTER_SLAVE + +This data is from the combination of SHOW MASTER STATUS and SHOW SLAVE STATUS. + +=item OPEN_TABLES + +This data is from SHOW OPEN TABLES. + +=item PROCESSLIST + +This data is from SHOW FULL PROCESSLIST. + +=item PROCESSLIST_STATS + +This data is from SHOW FULL PROCESSLIST and computes stats such as the maximum time +a user query has been running, and how many user queries are running. A "user +query" excludes replication threads. + +=item OS_WAIT_ARRAY + +This data is from the SEMAPHORES section of SHOW INNODB STATUS and is nested one +level deep. It comes from the lines that look like this: + + --Thread 1568861104 has waited at btr0cur.c line 424 .... + +=back + +=head1 MYSQL PRIVILEGES + +=over + +=item * + +You must connect to MySQL as a user who has the SUPER privilege for many of the +functions. + +=item * + +If you don't have the SUPER privilege, you can still run some functions, but you +won't necessarily see all the same data. + +=item * + +You need the PROCESS privilege to see the list of currently running queries in Q +mode. + +=item * + +You need special privileges to start and stop slave servers. + +=item * + +You need appropriate privileges to create and drop the deadlock tables if needed +(see L<"SERVER CONNECTIONS">). + +=back + +=head1 SYSTEM REQUIREMENTS + +You need Perl to run innotop, of course. You also need a few Perl modules: DBI, +DBD::MariaDB, Term::ReadKey, and Time::HiRes. These should be included with most +Perl distributions, but in case they are not, I recommend using versions +distributed with your operating system or Perl distribution, not from CPAN. +Term::ReadKey in particular has been known to cause problems if installed from +CPAN. + +If you have Term::ANSIColor, innotop will use it to format headers more readably +and compactly. (Under Microsoft Windows, you also need Win32::Console::ANSI for +terminal formatting codes to be honored). If you install Term::ReadLine, +preferably Term::ReadLine::Gnu, you'll get nice auto-completion support. + +I run innotop on Gentoo GNU/Linux, Debian and Ubuntu, and I've had feedback from +people successfully running it on Red Hat, CentOS, Solaris, and Mac OSX. I +don't see any reason why it won't work on other UNIX-ish operating systems, but +I don't know for sure. It also runs on Windows under ActivePerl without +problem. + +innotop has been used on MySQL versions 3.23.58, 4.0.27, 4.1.0, 4.1.22, 5.0.26, +5.1.15, and 5.2.3. If it doesn't run correctly for you, that is a bug that +should be reported. + +=head1 FILES + +$HOMEDIR/.innotop and/or /etc/innotop are used to store +configuration information. Files include the configuration file innotop.conf, +the core_dump file which contains verbose error messages if L<"debug"> is +enabled, and the plugins/ subdirectory. + +=head1 GLOSSARY OF TERMS + +=over + +=item tick + +A tick is a refresh event, when innotop re-fetches data from connections and +displays it. + +=back + +=head1 ACKNOWLEDGEMENTS + +The following people and organizations are acknowledged for various reasons. +Hopefully no one has been forgotten. + +Aaron Racine, +Allen K. Smith, +Aurimas Mikalauskas, +Bartosz Fenski, +Brian Miezejewski, +Christian Hammers, +Cyril Scetbon, +Dane Miller, +David Multer, +Dr. Frank Ullrich, +Giuseppe Maxia, +Google.com Site Reliability Engineers, +Google Code, +Jan Pieter Kunst, +Jari Aalto, +Jay Pipes, +Jeremy Zawodny, +Johan Idren, +Kristian Kohntopp, +Lenz Grimmer, +Maciej Dobrzanski, +Michiel Betel, +MySQL AB, +Paul McCullagh, +Sebastien Estienne, +Sourceforge.net, +Steven Kreuzer, +The Gentoo MySQL Team, +Trevor Price, +Yaar Schnitman, +and probably more people that have not been included. + +(If your name has been misspelled, it's probably out of fear of putting +international characters into this documentation; earlier versions of Perl might +not be able to compile it then). + +=head1 COPYRIGHT, LICENSE AND WARRANTY + +This program is copyright (c) 2006 Baron Schwartz. +Feedback and improvements are welcome. + +THIS PROGRAM IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED +WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF +MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + +This program is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free Software +Foundation, version 2; OR the Perl Artistic License. On UNIX and similar +systems, you can issue `man perlgpl' or `man perlartistic' to read these +licenses. + +You should have received a copy of the GNU General Public License along with +this program; if not, write to the Free Software Foundation, Inc., 51 Franklin +Street, Fifth Floor, Boston, MA 02110-1335 USA. + +Execute innotop and press '!' to see this information at any time. + +=head1 AUTHOR + +Originally written by Baron Schwartz; currently maintained by Aaron Racine. + +=head1 BUGS + +You can report bugs, ask for improvements, and get other help and support at +L<https://github.com/innotop/innotop>. There are mailing lists, a source code +browser, a bug tracker, etc. Please use these instead of contacting the +maintainer or author directly, as it makes our job easier and benefits others if the +discussions are permanent and public. Of course, if you need to contact us in +private, please do. + +=cut diff --git a/debian/additions/innotop/innotop.1 b/debian/additions/innotop/innotop.1 new file mode 100644 index 00000000..62a9aed6 --- /dev/null +++ b/debian/additions/innotop/innotop.1 @@ -0,0 +1,2200 @@ +.\" Automatically generated by Pod::Man 2.28 (Pod::Simple 3.28) +.\" +.\" Standard preamble: +.\" ======================================================================== +.de Sp \" Vertical space (when we can't use .PP) +.if t .sp .5v +.if n .sp +.. +.de Vb \" Begin verbatim text +.ft CW +.nf +.ne \\$1 +.. +.de Ve \" End verbatim text +.ft R +.fi +.. +.\" Set up some character translations and predefined strings. \*(-- will +.\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left +.\" double quote, and \*(R" will give a right double quote. \*(C+ will +.\" give a nicer C++. Capital omega is used to do unbreakable dashes and +.\" therefore won't be available. \*(C` and \*(C' expand to `' in nroff, +.\" nothing in troff, for use with C<>. +.tr \(*W- +.ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p' +.ie n \{\ +. ds -- \(*W- +. ds PI pi +. if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch +. if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch +. ds L" "" +. ds R" "" +. ds C` "" +. ds C' "" +'br\} +.el\{\ +. ds -- \|\(em\| +. ds PI \(*p +. ds L" `` +. ds R" '' +. ds C` +. ds C' +'br\} +.\" +.\" Escape single quotes in literal strings from groff's Unicode transform. +.ie \n(.g .ds Aq \(aq +.el .ds Aq ' +.\" +.\" If the F register is turned on, we'll generate index entries on stderr for +.\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index +.\" entries marked with X<> in POD. Of course, you'll have to process the +.\" output yourself in some meaningful fashion. +.\" +.\" Avoid warning from groff about undefined register 'F'. +.de IX +.. +.nr rF 0 +.if \n(.g .if rF .nr rF 1 +.if (\n(rF:(\n(.g==0)) \{ +. if \nF \{ +. de IX +. tm Index:\\$1\t\\n%\t"\\$2" +.. +. if !\nF==2 \{ +. nr % 0 +. nr F 2 +. \} +. \} +.\} +.rr rF +.\" +.\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2). +.\" Fear. Run. Save yourself. No user-serviceable parts. +. \" fudge factors for nroff and troff +.if n \{\ +. ds #H 0 +. ds #V .8m +. ds #F .3m +. ds #[ \f1 +. ds #] \fP +.\} +.if t \{\ +. ds #H ((1u-(\\\\n(.fu%2u))*.13m) +. ds #V .6m +. ds #F 0 +. ds #[ \& +. ds #] \& +.\} +. \" simple accents for nroff and troff +.if n \{\ +. ds ' \& +. ds ` \& +. ds ^ \& +. ds , \& +. ds ~ ~ +. ds / +.\} +.if t \{\ +. ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u" +. ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u' +. ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u' +. ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u' +. ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u' +. ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u' +.\} +. \" troff and (daisy-wheel) nroff accents +.ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V' +.ds 8 \h'\*(#H'\(*b\h'-\*(#H' +.ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#] +.ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H' +.ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u' +.ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#] +.ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#] +.ds ae a\h'-(\w'a'u*4/10)'e +.ds Ae A\h'-(\w'A'u*4/10)'E +. \" corrections for vroff +.if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u' +.if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u' +. \" for low resolution devices (crt and lpr) +.if \n(.H>23 .if \n(.V>19 \ +\{\ +. ds : e +. ds 8 ss +. ds o a +. ds d- d\h'-1'\(ga +. ds D- D\h'-1'\(hy +. ds th \o'bp' +. ds Th \o'LP' +. ds ae ae +. ds Ae AE +.\} +.rm #[ #] #H #V #F C +.\" ======================================================================== +.\" +.IX Title "INNOTOP 1" +.TH INNOTOP 1 "2017-01-23" "perl v5.20.2" "User Contributed Perl Documentation" +.\" For nroff, turn off justification. Always turn off hyphenation; it makes +.\" way too many mistakes in technical documents. +.if n .ad l +.nh +.SH "NAME" +innotop \- MySQL and InnoDB transaction/status monitor. +.SH "SYNOPSIS" +.IX Header "SYNOPSIS" +To monitor servers normally: +.PP +.Vb 1 +\& innotop +.Ve +.PP +To monitor InnoDB status information from a file: +.PP +.Vb 1 +\& innotop /var/log/mysql/mysqld.err +.Ve +.PP +To run innotop non-interactively in a pipe-and-filter configuration: +.PP +.Vb 1 +\& innotop \-\-count 5 \-d 1 \-n +.Ve +.PP +To monitor a database on another system using a particular username and password: +.PP +.Vb 1 +\& innotop \-u <username> \-p <password> \-h <hostname> +.Ve +.SH "DESCRIPTION" +.IX Header "DESCRIPTION" +innotop monitors MySQL servers. Each of its modes shows you a different aspect +of what's happening in the server. For example, there's a mode for monitoring +replication, one for queries, and one for transactions. innotop refreshes its +data periodically, so you see an updating view. +.PP +innotop has lots of features for power users, but you can start and run it with +virtually no configuration. If you're just getting started, see +\&\*(L"QUICK-START\*(R". Press '?' at any time while running innotop for +context-sensitive help. +.SH "QUICK-START" +.IX Header "QUICK-START" +To start innotop, open a terminal or command prompt. If you have installed +innotop on your system, you should be able to just type \*(L"innotop\*(R" and press +Enter; otherwise, you will need to change to innotop's directory and type \*(L"perl +innotop\*(R". +.PP +With no options specified, innotop will attempt to connect to a MySQL server on +localhost using mariadb_read_default_group=client for other connection +parameters. If you need to specify a different username and password, use the +\&\-u and \-p options, respectively. To monitor a MySQL database on another +host, use the \-h option. +.PP +After you've connected, innotop should show you something like the following: +.PP +.Vb 1 +\& [RO] Query List (? for help) localhost, 01:11:19, 449.44 QPS, 14/7/163 con/run +\& +\& CXN When Load QPS Slow QCacheHit KCacheHit BpsIn BpsOut +\& localhost Total 0.00 1.07k 697 0.00% 98.17% 476.83k 242.83k +\& +\& CXN Cmd ID User Host DB Time Query +\& localhost Query 766446598 test 10.0.0.1 foo 00:02 INSERT INTO table ( +.Ve +.PP +(This sample is truncated at the right so it will fit on a terminal when running +\&'man innotop') +.PP +If your server is busy, you'll see more output. Notice the first line on the +screen, which tells you that readonly is set to true ([\s-1RO\s0]), what mode you're +in and what server you're connected to. You can change to other modes with +keystrokes; press 'T' to switch to a list of InnoDB transactions, for example. +.PP +Press the '?' key to see what keys are active in the current mode. You can +press any of these keys and innotop will either take the requested action or +prompt you for more input. If your system has Term::ReadLine support, you can +use \s-1TAB\s0 and other keys to auto-complete and edit input. +.PP +To quit innotop, press the 'q' key. +.SH "OPTIONS" +.IX Header "OPTIONS" +innotop is mostly configured via its configuration file, but some of the +configuration options can come from the command line. You can also specify a +file to monitor for InnoDB status output; see \*(L"\s-1MONITORING A FILE\*(R"\s0 for more +details. +.PP +You can negate some options by prefixing the option name with \-\-no. For +example, \-\-noinc (or \-\-no\-inc) negates \*(L"\-\-inc\*(R". +.IP "\-\-color" 4 +.IX Item "--color" +Enable or disable terminal coloring. Corresponds to the \*(L"color\*(R" config file +setting. +.IP "\-\-config" 4 +.IX Item "--config" +Specifies a configuration file to read. This option is non-sticky, that is to +say it does not persist to the configuration file itself. +.IP "\-\-count" 4 +.IX Item "--count" +Refresh only the specified number of times (ticks) before exiting. Each refresh +is a pause for \*(L"interval\*(R" seconds, followed by requesting data from MySQL +connections and printing it to the terminal. +.IP "\-\-delay" 4 +.IX Item "--delay" +Specifies the amount of time to pause between ticks (refreshes). Corresponds to +the configuration option \*(L"interval\*(R". +.IP "\-\-help" 4 +.IX Item "--help" +Print a summary of command-line usage and exit. +.IP "\-\-host" 4 +.IX Item "--host" +Host to connect to. +.IP "\-\-inc" 4 +.IX Item "--inc" +Specifies whether innotop should display absolute numbers or relative numbers +(offsets from their previous values). Corresponds to the configuration option +\&\*(L"status_inc\*(R". +.IP "\-\-mode" 4 +.IX Item "--mode" +Specifies the mode in which innotop should start. Corresponds to the +configuration option \*(L"mode\*(R". +.IP "\-\-nonint" 4 +.IX Item "--nonint" +Enable non-interactive operation. See \*(L"NON-INTERACTIVE \s-1OPERATION\*(R"\s0 for more. +.IP "\-\-password" 4 +.IX Item "--password" +Password to use for connection. +.IP "\-\-port" 4 +.IX Item "--port" +Port to use for connection. +.IP "\-\-skipcentral" 4 +.IX Item "--skipcentral" +Don't read the central configuration file. +.IP "\-\-timestamp" 4 +.IX Item "--timestamp" +In \-n mode, write a timestamp either before every screenful of output, or if +the option is given twice, at the start of every line. The format is controlled +by the timeformat config variable. +.IP "\-\-user" 4 +.IX Item "--user" +User to use for connection. +.IP "\-\-version" 4 +.IX Item "--version" +Output version information and exit. +.IP "\-\-write" 4 +.IX Item "--write" +Sets the configuration option \*(L"readonly\*(R" to 0, making innotop write the +running configuration to ~/.innotop/innotop.conf on exit, if no configuration +file was loaded at start-up. +.SH "HOTKEYS" +.IX Header "HOTKEYS" +innotop is interactive, and you control it with key-presses. +.IP "\(bu" 4 +Uppercase keys switch between modes. +.IP "\(bu" 4 +Lowercase keys initiate some action within the current mode. +.IP "\(bu" 4 +Other keys do something special like change configuration or show the +innotop license. +.PP +Press '?' at any time to see the currently active keys and what they do. +.SH "MODES" +.IX Header "MODES" +Each of innotop's modes retrieves and displays a particular type of data from +the servers you're monitoring. You switch between modes with uppercase keys. +The following is a brief description of each mode, in alphabetical order. To +switch to the mode, press the key listed in front of its heading in the +following list: +.IP "A: Health Dashboard" 4 +.IX Item "A: Health Dashboard" +This mode displays a single table with one row per monitored server. The +columns show essential overview information about the server's health, and +coloration rules show whether replication is running or if there are any very +long-running queries or excessive replication delay. +.IP "B: InnoDB Buffers" 4 +.IX Item "B: InnoDB Buffers" +This mode displays information about the InnoDB buffer pool, page statistics, +insert buffer, and adaptive hash index. The data comes from \s-1SHOW INNODB STATUS.\s0 +.Sp +This mode contains the \*(L"buffer_pool\*(R", \*(L"page_statistics\*(R", +\&\*(L"insert_buffers\*(R", and \*(L"adaptive_hash_index\*(R" tables by default. +.IP "C: Command Summary" 4 +.IX Item "C: Command Summary" +This mode is similar to mytop's Command Summary mode. It shows the +\&\*(L"cmd_summary\*(R" table, which looks something like the following: +.Sp +.Vb 8 +\& Command Summary (? for help) localhost, 25+07:16:43, 2.45 QPS, 3 thd, 5.0.40 +\& _\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_ Command Summary _\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_ +\& Name Value Pct Last Incr Pct +\& Select_scan 3244858 69.89% 2 100.00% +\& Select_range 1354177 29.17% 0 0.00% +\& Select_full_join 39479 0.85% 0 0.00% +\& Select_full_range_join 4097 0.09% 0 0.00% +\& Select_range_check 0 0.00% 0 0.00% +.Ve +.Sp +The command summary table is built by extracting variables from +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. The variables must be numeric and must match the prefix +given by the \*(L"cmd_filter\*(R" configuration variable. The variables are then +sorted by value descending and compared to the last variable, as shown above. +The percentage columns are percentage of the total of all variables in the +table, so you can see the relative weight of the variables. +.Sp +The example shows what you see if the prefix is \*(L"Select_\*(R". The default +prefix is \*(L"Com_\*(R". You can choose a prefix with the 's' key. +.Sp +It's rather like running \s-1SHOW VARIABLES LIKE \s0\*(L"prefix%\*(R" with memory and +nice formatting. +.Sp +Values are aggregated across all servers. The Pct columns are not correctly +aggregated across multiple servers. This is a known limitation of the grouping +algorithm that may be fixed in the future. +.IP "D: InnoDB Deadlocks" 4 +.IX Item "D: InnoDB Deadlocks" +This mode shows the transactions involved in the last InnoDB deadlock. A second +table shows the locks each transaction held and waited for. A deadlock is +caused by a cycle in the waits-for graph, so there should be two locks held and +one waited for unless the deadlock information is truncated. +.Sp +InnoDB puts deadlock information before some other information in the \s-1SHOW +INNODB STATUS\s0 output. If there are a lot of locks, the deadlock information can +grow very large, and there is a limit on the size of the \s-1SHOW INNODB +STATUS\s0 output. A large deadlock can fill the entire output, or even be +truncated, and prevent you from seeing other information at all. If you are +running innotop in another mode, for example T mode, and suddenly you don't see +anything, you might want to check and see if a deadlock has wiped out the data +you need. +.Sp +If it has, you can create a small deadlock to replace the large one. Use the +\&'w' key to 'wipe' the large deadlock with a small one. This will not work +unless you have defined a deadlock table for the connection (see \*(L"\s-1SERVER +CONNECTIONS\*(R"\s0). +.Sp +You can also configure innotop to automatically detect when a large deadlock +needs to be replaced with a small one (see \*(L"auto_wipe_dl\*(R"). +.Sp +This mode displays the \*(L"deadlock_transactions\*(R" and \*(L"deadlock_locks\*(R" tables +by default. +.IP "F: InnoDB Foreign Key Errors" 4 +.IX Item "F: InnoDB Foreign Key Errors" +This mode shows the last InnoDB foreign key error information, such as the +table where it happened, when and who and what query caused it, and so on. +.Sp +InnoDB has a huge variety of foreign key error messages, and many of them are +just hard to parse. innotop doesn't always do the best job here, but there's +so much code devoted to parsing this messy, unparseable output that innotop is +likely never to be perfect in this regard. If innotop doesn't show you what +you need to see, just look at the status text directly. +.Sp +This mode displays the \*(L"fk_error\*(R" table by default. +.IP "I: InnoDB I/O Info" 4 +.IX Item "I: InnoDB I/O Info" +This mode shows InnoDB's I/O statistics, including the I/O threads, pending I/O, +file I/O miscellaneous, and log statistics. It displays the \*(L"io_threads\*(R", +\&\*(L"pending_io\*(R", \*(L"file_io_misc\*(R", and \*(L"log_statistics\*(R" tables by default. +.IP "K: InnoDB Lock Waits" 4 +.IX Item "K: InnoDB Lock Waits" +This mode shows information from InnoDB plugin's transaction and locking tables. +You can use it to find when a transaction is waiting for another, and kill the +blocking transaction. It displays the "innodb_blocked_blocker" table. +.IP "L: Locks" 4 +.IX Item "L: Locks" +This mode shows information about current locks. At the moment only InnoDB +locks are supported, and by default you'll only see locks for which transactions +are waiting. This information comes from the \s-1TRANSACTIONS\s0 section of the InnoDB +status text. If you have a very busy server, you may have frequent lock waits; +it helps to be able to see which tables and indexes are the \*(L"hot spot\*(R" for +locks. If your server is running pretty well, this mode should show nothing. +.Sp +You can configure MySQL and innotop to monitor not only locks for which a +transaction is waiting, but those currently held, too. You can do this with the +InnoDB Lock Monitor (<http://dev.mysql.com/doc/en/innodb\-monitor.html>). It's +not documented in the MySQL manual, but creating the lock monitor with the +following statement also affects the output of \s-1SHOW INNODB STATUS,\s0 which innotop +uses: +.Sp +.Vb 1 +\& CREATE TABLE innodb_lock_monitor(a int) ENGINE=INNODB; +.Ve +.Sp +This causes InnoDB to print its output to the MySQL file every 16 seconds or so, +as stated in the manual, but it also makes the normal \s-1SHOW INNODB STATUS\s0 output +include lock information, which innotop can parse and display (that's the +undocumented feature). +.Sp +This means you can do what may have seemed impossible: to a limited extent +(InnoDB truncates some information in the output), you can see which transaction +holds the locks something else is waiting for. You can also enable and disable +the InnoDB Lock Monitor with the key mappings in this mode. +.Sp +This mode displays the \*(L"innodb_locks\*(R" table by default. Here's a sample of +the screen when one connection is waiting for locks another connection holds: +.Sp +.Vb 7 +\& _\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_ InnoDB Locks _\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_\|_ +\& CXN ID Type Waiting Wait Active Mode DB Table Index +\& localhost 12 RECORD 1 00:10 00:10 X test t1 PRIMARY +\& localhost 12 TABLE 0 00:10 00:10 IX test t1 +\& localhost 12 RECORD 1 00:10 00:10 X test t1 PRIMARY +\& localhost 11 TABLE 0 00:00 00:25 IX test t1 +\& localhost 11 RECORD 0 00:00 00:25 X test t1 PRIMARY +.Ve +.Sp +You can see the first connection, \s-1ID 12,\s0 is waiting for a lock on the \s-1PRIMARY\s0 +key on test.t1, and has been waiting for 10 seconds. The second connection +isn't waiting, because the Waiting column is 0, but it holds locks on the same +index. That tells you connection 11 is blocking connection 12. +.IP "M: Master/Slave Replication Status" 4 +.IX Item "M: Master/Slave Replication Status" +This mode shows the output of \s-1SHOW SLAVE STATUS\s0 and \s-1SHOW MASTER STATUS\s0 in three +tables. The first two divide the slave's status into \s-1SQL\s0 and I/O thread status, +and the last shows master status. Filters are applied to eliminate non-slave +servers from the slave tables, and non-master servers from the master table. +.Sp +This mode displays the \*(L"slave_sql_status\*(R", \*(L"slave_io_status\*(R", and +\&\*(L"master_status\*(R" tables by default. +.IP "O: Open Tables" 4 +.IX Item "O: Open Tables" +This section comes from MySQL's \s-1SHOW OPEN TABLES\s0 command. By default it is +filtered to show tables which are in use by one or more queries, so you can +get a quick look at which tables are 'hot'. You can use this to guess which +tables might be locked implicitly. +.Sp +This mode displays the \*(L"open_tables\*(R" mode by default. +.IP "U: User Statistics" 4 +.IX Item "U: User Statistics" +This mode displays data that's available in Percona's enhanced version of MySQL +(also known as Percona Server with XtraDB). Specifically, it makes it easy to +enable and disable the so-called \*(L"user statistics.\*(R" This feature gathers stats +on clients, threads, users, tables, and indexes and makes them available as +\&\s-1INFORMATION_SCHEMA\s0 tables. These are invaluable for understanding what your +server is doing. They are also available in MariaDB. +.Sp +The statistics supported so far are only from the \s-1TABLE_STATISTICS\s0 and +\&\s-1INDEX_STATISTICS\s0 tables added by Percona. There are three views: one of table stats, +one of index stats (which can be aggregated with the = key), and one of both. +.Sp +The server doesn't gather these stats by default. You have to set the variable +userstat_running to turn it on. You can do this easily with innotop from U mode, +with the 's' key. +.IP "Q: Query List" 4 +.IX Item "Q: Query List" +This mode displays the output from \s-1SHOW FULL PROCESSLIST,\s0 much like \fBmytop\fR's +query list mode. This mode does \fBnot\fR show InnoDB-related information. This +is probably one of the most useful modes for general usage. +.Sp +There is an informative header that shows general status information about +your server. You can toggle it on and off with the 'h' key. By default, +innotop hides inactive processes and its own process. You can toggle these on +and off with the 'i' and 'a' keys. +.Sp +You can \s-1EXPLAIN\s0 a query from this mode with the 'e' key. This displays the +query's full text, the results of \s-1EXPLAIN,\s0 and in newer MySQL versions, even +the optimized query resulting from \s-1EXPLAIN EXTENDED. \s0 innotop also tries to +rewrite certain queries to make them EXPLAIN-able. For example, \s-1INSERT/SELECT\s0 +statements are rewritable. +.Sp +This mode displays the \*(L"q_header\*(R" and \*(L"processlist\*(R" tables by default. +.IP "R: InnoDB Row Operations and Semaphores" 4 +.IX Item "R: InnoDB Row Operations and Semaphores" +This mode shows InnoDB row operations, row operation miscellaneous, semaphores, +and information from the wait array. It displays the \*(L"row_operations\*(R", +\&\*(L"row_operation_misc\*(R", \*(L"semaphores\*(R", and \*(L"wait_array\*(R" tables by default. +.IP "S: Variables & Status" 4 +.IX Item "S: Variables & Status" +This mode calculates statistics, such as queries per second, and prints them out +in several different styles. You can show absolute values, or incremental values +between ticks. +.Sp +You can switch between the views by pressing a key. The 's' key prints a +single line each time the screen updates, in the style of \fBvmstat\fR. The 'g' +key changes the view to a graph of the same numbers, sort of like \fBtload\fR. +The 'v' key changes the view to a pivoted table of variable names on the left, +with successive updates scrolling across the screen from left to right. You can +choose how many updates to put on the screen with the \*(L"num_status_sets\*(R" +configuration variable. +.Sp +Headers may be abbreviated to fit on the screen in interactive operation. You +choose which variables to display with the 'c' key, which selects from +predefined sets, or lets you create your own sets. You can edit the current set +with the 'e' key. +.Sp +This mode doesn't really display any tables like other modes. Instead, it uses +a table definition to extract and format the data, but it then transforms the +result in special ways before outputting it. It uses the \*(L"var_status\*(R" table +definition for this. +.IP "T: InnoDB Transactions" 4 +.IX Item "T: InnoDB Transactions" +This mode shows transactions from the InnoDB monitor's output, in \fBtop\fR\-like +format. This mode is the reason I wrote innotop. +.Sp +You can kill queries or processes with the 'k' and 'x' keys, and \s-1EXPLAIN\s0 a query +with the 'e' or 'f' keys. InnoDB doesn't print the full query in transactions, +so explaining may not work right if the query is truncated. +.Sp +The informational header can be toggled on and off with the 'h' key. By +default, innotop hides inactive transactions and its own transaction. You can +toggle this on and off with the 'i' and 'a' keys. +.Sp +This mode displays the \*(L"t_header\*(R" and \*(L"innodb_transactions\*(R" tables by +default. +.SH "INNOTOP STATUS" +.IX Header "INNOTOP STATUS" +The first line innotop displays is a \*(L"status bar\*(R" of sorts. What it contains +depends on the mode you're in, and what servers you're monitoring. The first +few words are always [\s-1RO\s0] (if readonly is set to 1), the innotop mode, such as +\&\*(L"InnoDB Txns\*(R" for T mode, followed by a reminder to press '?' for help at any +time. +.SS "\s-1ONE SERVER\s0" +.IX Subsection "ONE SERVER" +The simplest case is when you're monitoring a single server. In this case, the +name of the connection is next on the status line. This is the name you gave +when you created the connection \*(-- most likely the MySQL server's hostname. +This is followed by the server's uptime. +.PP +If you're in an InnoDB mode, such as T or B, the next word is \*(L"InnoDB\*(R" followed +by some information about the \s-1SHOW INNODB STATUS\s0 output used to render the +screen. The first word is the number of seconds since the last \s-1SHOW INNODB +STATUS,\s0 which InnoDB uses to calculate some per-second statistics. The next is +a smiley face indicating whether the InnoDB output is truncated. If the smiley +face is a :\-), all is well; there is no truncation. A :^| means the transaction +list is so long, InnoDB has only printed out some of the transactions. Finally, +a frown :\-( means the output is incomplete, which is probably due to a deadlock +printing too much lock information (see \*(L"D: InnoDB Deadlocks\*(R"). +.PP +The next two words indicate the server's queries per second (\s-1QPS\s0) and how many +threads (connections) exist. Finally, the server's version number is the last +thing on the line. +.SS "\s-1MULTIPLE SERVERS\s0" +.IX Subsection "MULTIPLE SERVERS" +If you are monitoring multiple servers (see \*(L"\s-1SERVER CONNECTIONS\*(R"\s0), the status +line does not show any details about individual servers. Instead, it shows the +names of the connections that are active. Again, these are connection names you +specified, which are likely to be the server's hostname. A connection that has +an error is prefixed with an exclamation point. +.PP +If you are monitoring a group of servers (see \*(L"\s-1SERVER GROUPS\*(R"\s0), the status +line shows the name of the group. If any connection in the group has an +error, the group's name is followed by the fraction of the connections that +don't have errors. +.PP +See \*(L"\s-1ERROR HANDLING\*(R"\s0 for more details about innotop's error handling. +.SS "\s-1MONITORING A FILE\s0" +.IX Subsection "MONITORING A FILE" +If you give a filename on the command line, innotop will not connect to \s-1ANY\s0 +servers at all. It will watch the specified file for InnoDB status output and +use that as its data source. It will always show a single connection called +\&'file'. And since it can't connect to a server, it can't determine how long the +server it's monitoring has been up; so it calculates the server's uptime as time +since innotop started running. +.SH "SERVER ADMINISTRATION" +.IX Header "SERVER ADMINISTRATION" +While innotop is primarily a monitor that lets you watch and analyze your +servers, it can also send commands to servers. The most frequently useful +commands are killing queries and stopping or starting slaves. +.PP +You can kill a connection, or in newer versions of MySQL kill a query but not a +connection, from \*(L"Q: Query List\*(R" and \*(L"T: InnoDB Transactions\*(R" modes. +Press 'k' to issue a \s-1KILL\s0 command, or 'x' to issue a \s-1KILL QUERY\s0 command. +innotop will prompt you for the server and/or connection \s-1ID\s0 to kill (innotop +does not prompt you if there is only one possible choice for any input). +innotop pre-selects the longest-running query, or the oldest connection. +Confirm the command with 'y'. +.PP +In \*(L"Slave Replication Status\*(R"\*(L" in \*(R"M: Master mode, you can start and stop slaves +with the 'a' and 'o' keys, respectively. You can send these commands to many +slaves at once. innotop fills in a default command of \s-1START SLAVE\s0 or \s-1STOP SLAVE\s0 +for you, but you can actually edit the command and send anything you wish, such +as \s-1SET GLOBAL\s0 SQL_SLAVE_SKIP_COUNTER=1 to make the slave skip one binlog event +when it starts. +.PP +You can also ask innotop to calculate the earliest binlog in use by any slave +and issue a \s-1PURGE MASTER LOGS\s0 on the master. Use the 'b' key for this. innotop +will prompt you for a master to run the command on, then prompt you for the +connection names of that master's slaves (there is no way for innotop to +determine this reliably itself). innotop will find the minimum binlog in use by +these slave connections and suggest it as the argument to \s-1PURGE MASTER LOGS.\s0 +.PP +in \*(L"U: User Statistics\*(R" mode, you can use the 's' key to start and stop +the collection of the statistics data for \s-1TABLE_STATISTICS\s0 and similar. +.SH "SERVER CONNECTIONS" +.IX Header "SERVER CONNECTIONS" +When you create a server connection using '@', innotop asks you for a series of +inputs, as follows: +.IP "\s-1DSN\s0" 4 +.IX Item "DSN" +A \s-1DSN\s0 is a Data Source Name, which is the initial argument passed to the \s-1DBI\s0 +module for connecting to a server. It is usually of the form +.Sp +.Vb 1 +\& DBI:MariaDB:;mariadb_read_default_group=mysql;host=HOSTNAME +.Ve +.Sp +Since this \s-1DSN\s0 is passed to the DBD::MariaDB driver, you should read the driver's +documentation at <https://metacpan.org/pod/DBD::MariaDB> for +the exact details on all the options you can pass the driver in the \s-1DSN. \s0 You +can read more about \s-1DBI\s0 at <http://dbi.perl.org/docs/>, and especially at +<http://search.cpan.org/~timb/DBI/DBI.pm>. +.Sp +The mariadb_read_default_group=mysql option lets the \s-1DBD\s0 driver read your MySQL +options files, such as ~/.my.cnf on UNIX-ish systems. You can use this to avoid +specifying a username or password for the connection. +.IP "InnoDB Deadlock Table" 4 +.IX Item "InnoDB Deadlock Table" +This optional item tells innotop a table name it can use to deliberately create +a small deadlock (see \*(L"D: InnoDB Deadlocks\*(R"). If you specify this option, +you just need to be sure the table doesn't exist, and that innotop can create +and drop the table with the InnoDB storage engine. You can safely omit or just +accept the default if you don't intend to use this. +.IP "Username" 4 +.IX Item "Username" +innotop will ask you if you want to specify a username. If you say 'y', it will +then prompt you for a user name. If you have a MySQL option file that specifies +your username, you don't have to specify a username. +.Sp +The username defaults to your login name on the system you're running innotop on. +.IP "Password" 4 +.IX Item "Password" +innotop will ask you if you want to specify a password. Like the username, the +password is optional, but there's an additional prompt that asks if you want to +save the password in the innotop configuration file. If you don't save it in +the configuration file, innotop will prompt you for a password each time it +starts. Passwords in the innotop configuration file are saved in plain text, +not encrypted in any way. +.PP +Once you finish answering these questions, you should be connected to a server. +But innotop isn't limited to monitoring a single server; you can define many +server connections and switch between them by pressing the '@' key. See +\&\*(L"\s-1SWITCHING BETWEEN CONNECTIONS\*(R"\s0. +.SH "SERVER GROUPS" +.IX Header "SERVER GROUPS" +If you have multiple MySQL instances, you can put them into named groups, such +as 'all', 'masters', and 'slaves', which innotop can monitor all together. +.PP +You can choose which group to monitor with the '#' key, and you can press the +\&\s-1TAB\s0 key to switch to the next group. If you're not currently monitoring a +group, pressing \s-1TAB\s0 selects the first group. +.PP +To create a group, press the '#' key and type the name of your new group, then +type the names of the connections you want the group to contain. +.SH "SWITCHING BETWEEN CONNECTIONS" +.IX Header "SWITCHING BETWEEN CONNECTIONS" +innotop lets you quickly switch which servers you're monitoring. The most basic +way is by pressing the '@' key and typing the name(s) of the connection(s) you +want to use. This setting is per-mode, so you can monitor different connections +in each mode, and innotop remembers which connections you choose. +.PP +You can quickly switch to the 'next' connection in alphabetical order with the +\&'n' key. If you're monitoring a server group (see \*(L"\s-1SERVER GROUPS\*(R"\s0) this will +switch to the first connection. +.PP +You can also type many connection names, and innotop will fetch and display data +from them all. Just separate the connection names with spaces, for example +\&\*(L"server1 server2.\*(R" Again, if you type the name of a connection that doesn't +exist, innotop will prompt you for connection information and create the +connection. +.PP +Another way to monitor multiple connections at once is with server groups. You +can use the \s-1TAB\s0 key to switch to the 'next' group in alphabetical order, or if +you're not monitoring any groups, \s-1TAB\s0 will switch to the first group. +.PP +innotop does not fetch data in parallel from connections, so if you are +monitoring a large group or many connections, you may notice increased delay +between ticks. +.PP +When you monitor more than one connection, innotop's status bar changes. See +\&\*(L"\s-1INNOTOP STATUS\*(R"\s0. +.SH "ERROR HANDLING" +.IX Header "ERROR HANDLING" +Error handling is not that important when monitoring a single connection, but is +crucial when you have many active connections. A crashed server or lost +connection should not crash innotop. As a result, innotop will continue to run +even when there is an error; it just won't display any information from the +connection that had an error. Because of this, innotop's behavior might confuse +you. It's a feature, not a bug! +.PP +innotop does not continue to query connections that have errors, because they +may slow innotop and make it hard to use, especially if the error is a problem +connecting and causes a long time-out. Instead, innotop retries the connection +occasionally to see if the error still exists. If so, it will wait until some +point in the future. The wait time increases in ticks as the Fibonacci series, +so it tries less frequently as time passes. +.PP +Since errors might only happen in certain modes because of the \s-1SQL\s0 commands +issued in those modes, innotop keeps track of which mode caused the error. If +you switch to a different mode, innotop will retry the connection instead of +waiting. +.PP +By default innotop will display the problem in red text at the bottom of the +first table on the screen. You can disable this behavior with the +\&\*(L"show_cxn_errors_in_tbl\*(R" configuration option, which is enabled by default. +If the \*(L"debug\*(R" option is enabled, innotop will display the error at the +bottom of every table, not just the first. And if \*(L"show_cxn_errors\*(R" is +enabled, innotop will print the error text to \s-1STDOUT\s0 as well. Error messages +might only display in the mode that caused the error, depending on the mode and +whether innotop is avoiding querying that connection. +.SH "NON-INTERACTIVE OPERATION" +.IX Header "NON-INTERACTIVE OPERATION" +You can run innotop in non-interactive mode, in which case it is entirely +controlled from the configuration file and command-line options. To start +innotop in non-interactive mode, give the L\*(L"<\-\-nonint\*(R"> command-line option. +This changes innotop's behavior in the following ways: +.IP "\(bu" 4 +Certain Perl modules are not loaded. Term::Readline is not loaded, since +innotop doesn't prompt interactively. Term::ANSIColor and Win32::Console::ANSI +modules are not loaded. Term::ReadKey is still used, since innotop may have to +prompt for connection passwords when starting up. +.IP "\(bu" 4 +innotop does not clear the screen after each tick. +.IP "\(bu" 4 +innotop does not persist any changes to the configuration file. +.IP "\(bu" 4 +If \*(L"\-\-count\*(R" is given and innotop is in incremental mode (see \*(L"status_inc\*(R" +and \*(L"\-\-inc\*(R"), innotop actually refreshes one more time than specified so it +can print incremental statistics. This suppresses output during the first +tick, so innotop may appear to hang. +.IP "\(bu" 4 +innotop only displays the first table in each mode. This is so the output can +be easily processed with other command-line utilities such as awk and sed. To +change which tables display in each mode, see \*(L"\s-1TABLES\*(R"\s0. Since \*(L"Q: Query +List\*(R" mode is so important, innotop automatically disables the \*(L"q_header\*(R" +table. This ensures you'll see the \*(L"processlist\*(R" table, even if you have +innotop configured to show the q_header table during interactive operation. +Similarly, in \*(L"T: InnoDB Transactions\*(R" mode, the \*(L"t_header\*(R" table is +suppressed so you see only the \*(L"innodb_transactions\*(R" table. +.IP "\(bu" 4 +All output is tab-separated instead of being column-aligned with whitespace, and +innotop prints the full contents of each table instead of only printing one +screenful at a time. +.IP "\(bu" 4 +innotop only prints column headers once instead of every tick (see +\&\*(L"hide_hdr\*(R"). innotop does not print table captions (see +\&\*(L"display_table_captions\*(R"). innotop ensures there are no empty lines in the +output. +.IP "\(bu" 4 +innotop does not honor the \*(L"shorten\*(R" transformation, which normally shortens +some numbers to human-readable formats. +.IP "\(bu" 4 +innotop does not print a status line (see \*(L"\s-1INNOTOP STATUS\*(R"\s0). +.SH "CONFIGURING" +.IX Header "CONFIGURING" +Nearly everything about innotop is configurable. Most things are possible to +change with built-in commands, but you can also edit the configuration file. +.PP +While running innotop, press the '$' key to bring up the configuration editing +dialog. Press another key to select the type of data you want to edit: +.IP "S: Statement Sleep Times" 4 +.IX Item "S: Statement Sleep Times" +Edits \s-1SQL\s0 statement sleep delays, which make innotop pause for the specified +amount of time after executing a statement. See \*(L"\s-1SQL STATEMENTS\*(R"\s0 for a +definition of each statement and what it does. By default innotop does not +delay after any statements. +.Sp +This feature is included so you can customize the side-effects caused by +monitoring your server. You may not see any effects, but some innotop users +have noticed that certain MySQL versions under very high load with InnoDB +enabled take longer than usual to execute \s-1SHOW GLOBAL STATUS. \s0 If innotop calls +\&\s-1SHOW FULL PROCESSLIST\s0 immediately afterward, the processlist contains more +queries than the machine actually averages at any given moment. Configuring +innotop to pause briefly after calling \s-1SHOW GLOBAL STATUS\s0 alleviates this +effect. +.Sp +Sleep times are stored in the \*(L"stmt_sleep_times\*(R" section of the configuration +file. Fractional-second sleeps are supported, subject to your hardware's +limitations. +.IP "c: Edit Columns" 4 +.IX Item "c: Edit Columns" +Starts the table editor on one of the displayed tables. See \*(L"\s-1TABLE EDITOR\*(R"\s0. +An alternative way to start the table editor without entering the configuration +dialog is with the '^' key. +.IP "g: General Configuration" 4 +.IX Item "g: General Configuration" +Starts the configuration editor to edit global and mode-specific configuration +variables (see \*(L"\s-1MODES\*(R"\s0). innotop prompts you to choose a variable from among +the global and mode-specific ones depending on the current mode. +.IP "k: Row-Coloring Rules" 4 +.IX Item "k: Row-Coloring Rules" +Starts the row-coloring rules editor on one of the displayed table(s). See +\&\*(L"\s-1COLORS\*(R"\s0 for details. +.IP "p: Manage Plugins" 4 +.IX Item "p: Manage Plugins" +Starts the plugin configuration editor. See \*(L"\s-1PLUGINS\*(R"\s0 for details. +.IP "s: Server Groups" 4 +.IX Item "s: Server Groups" +Lets you create and edit server groups. See \*(L"\s-1SERVER GROUPS\*(R"\s0. +.IP "t: Choose Displayed Tables" 4 +.IX Item "t: Choose Displayed Tables" +Lets you choose which tables to display in this mode. See \*(L"\s-1MODES\*(R"\s0 and +\&\*(L"\s-1TABLES\*(R"\s0. +.SH "CONFIGURATION FILE" +.IX Header "CONFIGURATION FILE" +innotop's default configuration file locations are \f(CW$HOME\fR/.innotop and +/etc/innotop/innotop.conf, and they are looked for in that order. If the first +configuration file exists, the second will not be processed. Those can be +overridden with the \*(L"\-\-config\*(R" command-line option. You can edit it by hand +safely, however innotop reads the configuration file when it starts, and, if +readonly is set to 0, writes it out again when it exits. Thus, if readonly is +set to 0, any changes you make by hand while innotop is running will be lost. +.PP +innotop doesn't store its entire configuration in the configuration file. It +has a huge set of default configuration values that it holds only in memory, +and the configuration file only overrides these defaults. When you customize a +default setting, innotop notices, and then stores the customizations into the +file. This keeps the file size down, makes it easier to edit, and makes +upgrades easier. +.PP +A configuration file is read-only be default. You can override that with +\&\*(L"\-\-write\*(R". See \*(L"readonly\*(R". +.PP +The configuration file is arranged into sections like an \s-1INI\s0 file. Each +section begins with [section\-name] and ends with [/section\-name]. Each +section's entries have a different syntax depending on the data they need to +store. You can put comments in the file; any line that begins with a # +character is a comment. innotop will not read the comments, so it won't write +them back out to the file when it exits. Comments in read-only configuration +files are still useful, though. +.PP +The first line in the file is innotop's version number. This lets innotop +notice when the file format is not backwards-compatible, and upgrade smoothly +without destroying your customized configuration. +.PP +The following list describes each section of the configuration file and the data +it contains: +.IP "general" 4 +.IX Item "general" +The 'general' section contains global configuration variables and variables that +may be mode-specific, but don't belong in any other section. The syntax is a +simple key=value list. innotop writes a comment above each value to help you +edit the file by hand. +.RS 4 +.IP "S_func" 4 +.IX Item "S_func" +Controls S mode presentation (see \*(L"S: Variables & Status\*(R"). If g, values are +graphed; if s, values are like vmstat; if p, values are in a pivoted table. +.IP "S_set" 4 +.IX Item "S_set" +Specifies which set of variables to display in \*(L"S: Variables & Status\*(R" mode. +See \*(L"\s-1VARIABLE SETS\*(R"\s0. +.IP "auto_wipe_dl" 4 +.IX Item "auto_wipe_dl" +Instructs innotop to automatically wipe large deadlocks when it notices them. +When this happens you may notice a slight delay. At the next tick, you will +usually see the information that was being truncated by the large deadlock. +.IP "charset" 4 +.IX Item "charset" +Specifies what kind of characters to allow through the \*(L"no_ctrl_char\*(R" +transformation. This keeps non-printable characters from confusing a +terminal when you monitor queries that contain binary data, such as images. +.Sp +The default is 'ascii', which considers anything outside normal \s-1ASCII\s0 to be a +control character. The other allowable values are 'unicode' and 'none'. 'none' +considers every character a control character, which can be useful for +collapsing \s-1ALL\s0 text fields in queries. +.IP "cmd_filter" 4 +.IX Item "cmd_filter" +This is the prefix that filters variables in \*(L"C: Command Summary\*(R" mode. +.IP "color" 4 +.IX Item "color" +Whether terminal coloring is permitted. +.IP "cxn_timeout" 4 +.IX Item "cxn_timeout" +On MySQL versions 4.0.3 and newer, this variable is used to set the connection's +timeout, so MySQL doesn't close the connection if it is not used for a while. +This might happen because a connection isn't monitored in a particular mode, for +example. +.IP "debug" 4 +.IX Item "debug" +This option enables more verbose errors and makes innotop more strict in some +places. It can help in debugging filters and other user-defined code. It also +makes innotop write a lot of information to \*(L"debugfile\*(R" when there is a +crash. +.IP "debugfile" 4 +.IX Item "debugfile" +A file to which innotop will write information when there is a crash. See +\&\*(L"\s-1FILES\*(R"\s0. +.IP "display_table_captions" 4 +.IX Item "display_table_captions" +innotop displays a table caption above most tables. This variable suppresses or +shows captions on all tables globally. Some tables are configured with the +hide_caption property, which overrides this. +.IP "global" 4 +.IX Item "global" +Whether to show \s-1GLOBAL\s0 variables and status. innotop only tries to do this on +servers which support the \s-1GLOBAL\s0 option to \s-1SHOW VARIABLES\s0 and \s-1SHOW STATUS. \s0 In +some MySQL versions, you need certain privileges to do this; if you don't have +them, innotop will not be able to fetch any variable and status data. This +configuration variable lets you run innotop and fetch what data you can even +without the elevated privileges. +.Sp +I can no longer find or reproduce the situation where \s-1GLOBAL\s0 wasn't allowed, but +I know there was one. +.IP "graph_char" 4 +.IX Item "graph_char" +Defines the character to use when drawing graphs in \*(L"S: Variables & Status\*(R" +mode. +.IP "header_highlight" 4 +.IX Item "header_highlight" +Defines how to highlight column headers. This only works if Term::ANSIColor is +available. Valid values are 'bold' and 'underline'. +.IP "hide_hdr" 4 +.IX Item "hide_hdr" +Hides column headers globally. +.IP "interval" 4 +.IX Item "interval" +The interval at which innotop will refresh its data (ticks). The interval is +implemented as a sleep time between ticks, so the true interval will vary +depending on how long it takes innotop to fetch and render data. +.Sp +This variable accepts fractions of a second. +.IP "mode" 4 +.IX Item "mode" +The mode in which innotop should start. Allowable arguments are the same as the +key presses that select a mode interactively. See \*(L"\s-1MODES\*(R"\s0. +.IP "num_digits" 4 +.IX Item "num_digits" +How many digits to show in fractional numbers and percents. This variable's +range is between 0 and 9 and can be set directly from \*(L"S: Variables & Status\*(R" +mode with the '+' and '\-' keys. It is used in the \*(L"set_precision\*(R", +\&\*(L"shorten\*(R", and \*(L"percent\*(R" transformations. +.IP "num_status_sets" 4 +.IX Item "num_status_sets" +Controls how many sets of status variables to display in pivoted \*(L"S: Variables +& Status\*(R" mode. It also controls the number of old sets of variables innotop +keeps in its memory, so the larger this variable is, the more memory innotop +uses. +.IP "plugin_dir" 4 +.IX Item "plugin_dir" +Specifies where plugins can be found. By default, innotop stores plugins in the +\&'plugins' subdirectory of your innotop configuration directory. +.IP "readonly" 4 +.IX Item "readonly" +Whether the configuration file is readonly. This cannot be set interactively. +.IP "show_cxn_errors" 4 +.IX Item "show_cxn_errors" +Makes innotop print connection errors to \s-1STDOUT. \s0 See \*(L"\s-1ERROR HANDLING\*(R"\s0. +.IP "show_cxn_errors_in_tbl" 4 +.IX Item "show_cxn_errors_in_tbl" +Makes innotop display connection errors as rows in the first table on screen. +See \*(L"\s-1ERROR HANDLING\*(R"\s0. +.IP "show_percent" 4 +.IX Item "show_percent" +Adds a '%' character after the value returned by the \*(L"percent\*(R" +transformation. +.IP "show_statusbar" 4 +.IX Item "show_statusbar" +Controls whether to show the status bar in the display. See \*(L"\s-1INNOTOP +STATUS\*(R"\s0. +.IP "skip_innodb" 4 +.IX Item "skip_innodb" +Disables fetching \s-1SHOW INNODB STATUS,\s0 in case your server(s) do not have InnoDB +enabled and you don't want innotop to try to fetch it. This can also be useful +when you don't have the \s-1SUPER\s0 privilege, required to run \s-1SHOW INNODB STATUS.\s0 +.IP "spark" 4 +.IX Item "spark" +Specifies how wide a spark chart is. There are two \s-1ASCII\s0 spark charts in A +mode, showing \s-1QPS\s0 and User_threads_running. +.IP "status_inc" 4 +.IX Item "status_inc" +Whether to show absolute or incremental values for status variables. +Incremental values are calculated as an offset from the last value innotop saw +for that variable. This is a global setting, but will probably become +mode-specific at some point. Right now it is honored a bit inconsistently; some +modes don't pay attention to it. +.IP "timeformat" 4 +.IX Item "timeformat" +The C\-style \fIstrftime()\fR\-compatible format for the timestamp line to be printed +in \-n mode when \-t is set. +.RE +.RS 4 +.RE +.IP "plugins" 4 +.IX Item "plugins" +This section holds a list of package names of active plugins. If the plugin +exists, innotop will activate it. See \*(L"\s-1PLUGINS\*(R"\s0 for more information. +.IP "filters" 4 +.IX Item "filters" +This section holds user-defined filters (see \*(L"\s-1FILTERS\*(R"\s0). Each line is in the +format filter_name=text='filter text' tbls='table list'. +.Sp +The filter text is the text of the subroutine's code. The table list is a list +of tables to which the filter can apply. By default, user-defined filters apply +to the table for which they were created, but you can manually override that by +editing the definition in the configuration file. +.IP "active_filters" 4 +.IX Item "active_filters" +This section stores which filters are active on each table. Each line is in the +format table_name=filter_list. +.IP "tbl_meta" 4 +.IX Item "tbl_meta" +This section stores user-defined or user-customized columns (see \*(L"\s-1COLUMNS\*(R"\s0). +Each line is in the format col_name=properties, where the properties are a +name=quoted\-value list. +.IP "connections" 4 +.IX Item "connections" +This section holds the server connections you have defined. Each line is in +the format name=properties, where the properties are a name=value list. The +properties are self-explanatory, and the only one that is treated specially is +\&'pass' which is only present if 'savepass' is set. This section of the +configuration file will be skipped if any \s-1DSN,\s0 username, or password +command-line options are used. See \*(L"\s-1SERVER CONNECTIONS\*(R"\s0. +.IP "active_connections" 4 +.IX Item "active_connections" +This section holds a list of which connections are active in each mode. Each +line is in the format mode_name=connection_list. +.IP "server_groups" 4 +.IX Item "server_groups" +This section holds server groups. Each line is in the format +name=connection_list. See \*(L"\s-1SERVER GROUPS\*(R"\s0. +.IP "active_server_groups" 4 +.IX Item "active_server_groups" +This section holds a list of which server group is active in each mode. Each +line is in the format mode_name=server_group. +.IP "max_values_seen" 4 +.IX Item "max_values_seen" +This section holds the maximum values seen for variables. This is used to scale +the graphs in \*(L"S: Variables & Status\*(R" mode. Each line is in the format +name=value. +.IP "active_columns" 4 +.IX Item "active_columns" +This section holds table column lists. Each line is in the format +tbl_name=column_list. See \*(L"\s-1COLUMNS\*(R"\s0. +.IP "sort_cols" 4 +.IX Item "sort_cols" +This section holds the sort definition. Each line is in the format +tbl_name=column_list. If a column is prefixed with '\-', that column sorts +descending. See \*(L"\s-1SORTING\*(R"\s0. +.IP "visible_tables" 4 +.IX Item "visible_tables" +This section defines which tables are visible in each mode. Each line is in the +format mode_name=table_list. See \*(L"\s-1TABLES\*(R"\s0. +.IP "varsets" 4 +.IX Item "varsets" +This section defines variable sets for use in \*(L"S: Status & Variables\*(R" mode. +Each line is in the format name=variable_list. See \*(L"\s-1VARIABLE SETS\*(R"\s0. +.IP "colors" 4 +.IX Item "colors" +This section defines colorization rules. Each line is in the format +tbl_name=property_list. See \*(L"\s-1COLORS\*(R"\s0. +.IP "stmt_sleep_times" 4 +.IX Item "stmt_sleep_times" +This section contains statement sleep times. Each line is in the format +statement_name=sleep_time. See \*(L"S: Statement Sleep Times\*(R". +.IP "group_by" 4 +.IX Item "group_by" +This section contains column lists for table group_by expressions. Each line is +in the format tbl_name=column_list. See \*(L"\s-1GROUPING\*(R"\s0. +.SH "CUSTOMIZING" +.IX Header "CUSTOMIZING" +You can customize innotop a great deal. For example, you can: +.IP "\(bu" 4 +Choose which tables to display, and in what order. +.IP "\(bu" 4 +Choose which columns are in those tables, and create new columns. +.IP "\(bu" 4 +Filter which rows display with built-in filters, user-defined filters, and +quick-filters. +.IP "\(bu" 4 +Sort the rows to put important data first or group together related rows. +.IP "\(bu" 4 +Highlight rows with color. +.IP "\(bu" 4 +Customize the alignment, width, and formatting of columns, and apply +transformations to columns to extract parts of their values or format the values +as you wish (for example, shortening large numbers to familiar units). +.IP "\(bu" 4 +Design your own expressions to extract and combine data as you need. This gives +you unlimited flexibility. +.PP +All these and more are explained in the following sections. +.SS "\s-1TABLES\s0" +.IX Subsection "TABLES" +A table is what you'd expect: a collection of columns. It also has some other +properties, such as a caption. Filters, sorting rules, and colorization rules +belong to tables and are covered in later sections. +.PP +Internally, table meta-data is defined in a data structure called \f(CW%tbl_meta\fR. +This hash holds all built-in table definitions, which contain a lot of default +instructions to innotop. The meta-data includes the caption, a list of columns +the user has customized, a list of columns, a list of visible columns, a list of +filters, color rules, a sort-column list, sort direction, and some information +about the table's data sources. Most of this is customizable via the table +editor (see \*(L"\s-1TABLE EDITOR\*(R"\s0). +.PP +You can choose which tables to show by pressing the '$' key. See \*(L"\s-1MODES\*(R"\s0 and +\&\*(L"\s-1TABLES\*(R"\s0. +.PP +The table life-cycle is as follows: +.IP "\(bu" 4 +Each table begins with a data source, which is an array of hashes. See below +for details on data sources. +.IP "\(bu" 4 +Each element of the data source becomes a row in the final table. +.IP "\(bu" 4 +For each element in the data source, innotop extracts values from the source and +creates a row. This row is another hash, which later steps will refer to as +\&\f(CW$set\fR. The values innotop extracts are determined by the table's columns. Each +column has an extraction subroutine, compiled from an expression (see +\&\*(L"\s-1EXPRESSIONS\*(R"\s0). The resulting row is a hash whose keys are named the same as +the column name. +.IP "\(bu" 4 +innotop filters the rows, removing those that don't need to be displayed. See +\&\*(L"\s-1FILTERS\*(R"\s0. +.IP "\(bu" 4 +innotop sorts the rows. See \*(L"\s-1SORTING\*(R"\s0. +.IP "\(bu" 4 +innotop groups the rows together, if specified. See \*(L"\s-1GROUPING\*(R"\s0. +.IP "\(bu" 4 +innotop colorizes the rows. See \*(L"\s-1COLORS\*(R"\s0. +.IP "\(bu" 4 +innotop transforms the column values in each row. See \*(L"\s-1TRANSFORMATIONS\*(R"\s0. +.IP "\(bu" 4 +innotop optionally pivots the rows (see \*(L"\s-1PIVOTING\*(R"\s0), then filters and sorts +them. +.IP "\(bu" 4 +innotop formats and justifies the rows as a table. During this step, innotop +applies further formatting to the column values, including alignment, maximum +and minimum widths. innotop also does final error checking to ensure there are +no crashes due to undefined values. innotop then adds a caption if specified, +and the table is ready to print. +.PP +The lifecycle is slightly different if the table is pivoted, as noted above. To +clarify, if the table is pivoted, the process is extract, group, transform, +pivot, filter, sort, create. If it's not pivoted, the process is extract, +filter, sort, group, color, transform, create. This slightly convoluted process +doesn't map all that well to \s-1SQL,\s0 but pivoting complicates things pretty +thoroughly. Roughly speaking, filtering and sorting happen as late as needed to +effect the final result as you might expect, but as early as possible for +efficiency. +.PP +Each built-in table is described below: +.IP "adaptive_hash_index" 4 +.IX Item "adaptive_hash_index" +Displays data about InnoDB's adaptive hash index. Data source: +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "buffer_pool" 4 +.IX Item "buffer_pool" +Displays data about InnoDB's buffer pool. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "cmd_summary" 4 +.IX Item "cmd_summary" +Displays weighted status variables. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "deadlock_locks" 4 +.IX Item "deadlock_locks" +Shows which locks were held and waited for by the last detected deadlock. Data +source: \*(L"\s-1DEADLOCK_LOCKS\*(R"\s0. +.IP "deadlock_transactions" 4 +.IX Item "deadlock_transactions" +Shows transactions involved in the last detected deadlock. Data source: +\&\*(L"\s-1DEADLOCK_TRANSACTIONS\*(R"\s0. +.IP "explain" 4 +.IX Item "explain" +Shows the output of \s-1EXPLAIN. \s0 Data source: \*(L"\s-1EXPLAIN\*(R"\s0. +.IP "file_io_misc" 4 +.IX Item "file_io_misc" +Displays data about InnoDB's file and I/O operations. Data source: +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "fk_error" 4 +.IX Item "fk_error" +Displays various data about InnoDB's last foreign key error. Data source: +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "health_dashboard" 4 +.IX Item "health_dashboard" +Displays an overall summary of servers, one server per line, for monitoring. +Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0, \*(L"\s-1MASTER_SLAVE\*(R"\s0, \*(L"\s-1PROCESSLIST_STATS\*(R"\s0. +.IP "index_statistics" 4 +.IX Item "index_statistics" +Displays data from the \s-1INDEX_STATISTICS\s0 table in Percona-enhanced servers. +.IP "index_table_statistics" 4 +.IX Item "index_table_statistics" +Displays data from the \s-1INDEX_STATISTICS\s0 and \s-1TABLE_STATISTICS\s0 tables in +Percona-enhanced servers. It joins the two together, grouped by the database +and table name. It is the default view in \*(L"U: User Statistics\*(R" mode, +and makes it easy to see what tables are hot, how many rows are read from indexes, +how many changes are made, and how many changes are made to indexes. +.IP "innodb_blocked_blocker" 4 +.IX Item "innodb_blocked_blocker" +Displays InnoDB locks and lock waits. Data source: \*(L"\s-1INNODB_BLOCKED_BLOCKER\*(R"\s0. +.IP "innodb_locks" 4 +.IX Item "innodb_locks" +Displays InnoDB locks. Data source: \*(L"\s-1INNODB_LOCKS\*(R"\s0. +.IP "innodb_transactions" 4 +.IX Item "innodb_transactions" +Displays data about InnoDB's current transactions. Data source: +\&\*(L"\s-1INNODB_TRANSACTIONS\*(R"\s0. +.IP "insert_buffers" 4 +.IX Item "insert_buffers" +Displays data about InnoDB's insert buffer. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "io_threads" 4 +.IX Item "io_threads" +Displays data about InnoDB's I/O threads. Data source: \*(L"\s-1IO_THREADS\*(R"\s0. +.IP "log_statistics" 4 +.IX Item "log_statistics" +Displays data about InnoDB's logging system. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "master_status" 4 +.IX Item "master_status" +Displays replication master status. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "open_tables" 4 +.IX Item "open_tables" +Displays open tables. Data source: \*(L"\s-1OPEN_TABLES\*(R"\s0. +.IP "page_statistics" 4 +.IX Item "page_statistics" +Displays InnoDB page statistics. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "pending_io" 4 +.IX Item "pending_io" +Displays InnoDB pending I/O operations. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "processlist" 4 +.IX Item "processlist" +Displays current MySQL processes (threads/connections). Data source: +\&\*(L"\s-1PROCESSLIST\*(R"\s0. +.IP "q_header" 4 +.IX Item "q_header" +Displays various status values. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "row_operation_misc" 4 +.IX Item "row_operation_misc" +Displays data about InnoDB's row operations. Data source: +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "row_operations" 4 +.IX Item "row_operations" +Displays data about InnoDB's row operations. Data source: +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "semaphores" 4 +.IX Item "semaphores" +Displays data about InnoDB's semaphores and mutexes. Data source: +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "slave_io_status" 4 +.IX Item "slave_io_status" +Displays data about the slave I/O thread. Data source: +\&\*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "slave_sql_status" 4 +.IX Item "slave_sql_status" +Displays data about the slave \s-1SQL\s0 thread. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "table_statistics" 4 +.IX Item "table_statistics" +Displays data from the \s-1TABLE_STATISTICS\s0 table in Percona-enhanced servers. +.IP "t_header" 4 +.IX Item "t_header" +Displays various InnoDB status values. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "var_status" 4 +.IX Item "var_status" +Displays user-configurable data. Data source: \*(L"\s-1STATUS_VARIABLES\*(R"\s0. +.IP "wait_array" 4 +.IX Item "wait_array" +Displays data about InnoDB's \s-1OS\s0 wait array. Data source: \*(L"\s-1OS_WAIT_ARRAY\*(R"\s0. +.SS "\s-1COLUMNS\s0" +.IX Subsection "COLUMNS" +Columns belong to tables. You can choose a table's columns by pressing the '^' +key, which starts the \*(L"\s-1TABLE EDITOR\*(R"\s0 and lets you choose and edit columns. +Pressing 'e' from within the table editor lets you edit the column's properties: +.IP "\(bu" 4 +hdr: a column header. This appears in the first row of the table. +.IP "\(bu" 4 +just: justification. '\-' means left-justified and '' means right-justified, +just as with printf formatting codes (not a coincidence). +.IP "\(bu" 4 +dec: whether to further align the column on the decimal point. +.IP "\(bu" 4 +num: whether the column is numeric. This affects how values are sorted +(lexically or numerically). +.IP "\(bu" 4 +label: a small note about the column, which appears in dialogs that help the +user choose columns. +.IP "\(bu" 4 +src: an expression that innotop uses to extract the column's data from its +source (see \*(L"\s-1DATA SOURCES\*(R"\s0). See \*(L"\s-1EXPRESSIONS\*(R"\s0 for more on expressions. +.IP "\(bu" 4 +minw: specifies a minimum display width. This helps stabilize the display, +which makes it easier to read if the data is changing frequently. +.IP "\(bu" 4 +maxw: similar to minw. +.IP "\(bu" 4 +trans: a list of column transformations. See \*(L"\s-1TRANSFORMATIONS\*(R"\s0. +.IP "\(bu" 4 +agg: an aggregate function. See \*(L"\s-1GROUPING\*(R"\s0. The default is \*(L"first\*(R". +.IP "\(bu" 4 +aggonly: controls whether the column only shows when grouping is enabled on the +table (see \*(L"\s-1GROUPING\*(R"\s0). By default, this is disabled. This means columns +will always be shown by default, whether grouping is enabled or not. If a +column's aggonly is set true, the column will appear when you toggle grouping on +the table. Several columns are set this way, such as the count column on +\&\*(L"processlist\*(R" and \*(L"innodb_transactions\*(R", so you don't see a count when the +grouping isn't enabled, but you do when it is. +.IP "\(bu" 4 +agghide: the reverse of aggonly. The column is hidden when grouping is enabled. +.SS "\s-1FILTERS\s0" +.IX Subsection "FILTERS" +Filters remove rows from the display. They behave much like a \s-1WHERE\s0 clause in +\&\s-1SQL. \s0 innotop has several built-in filters, which remove irrelevant information +like inactive queries, but you can define your own as well. innotop also lets +you create quick-filters, which do not get saved to the configuration file, and +are just an easy way to quickly view only some rows. +.PP +You can enable or disable a filter on any table. Press the '%' key (mnemonic: % +looks kind of like a line being filtered between two circles) and choose which +table you want to filter, if asked. You'll then see a list of possible filters +and a list of filters currently enabled for that table. Type the names of +filters you want to apply and press Enter. +.PP +\fIUSER-DEFINED \s-1FILTERS\s0\fR +.IX Subsection "USER-DEFINED FILTERS" +.PP +If you type a name that doesn't exist, innotop will prompt you to create the +filter. Filters are easy to create if you know Perl, and not hard if you don't. +What you're doing is creating a subroutine that returns true if the row should +be displayed. The row is a hash reference passed to your subroutine as \f(CW$set\fR. +.PP +For example, imagine you want to filter the processlist table so you only see +queries that have been running more than five minutes. Type a new name for your +filter, and when prompted for the subroutine body, press \s-1TAB\s0 to initiate your +terminal's auto-completion. You'll see the names of the columns in the +\&\*(L"processlist\*(R" table (innotop generally tries to help you with auto-completion +lists). You want to filter on the 'time' column. Type the text \*(L"$set\->{time} > +300\*(R" to return true when the query is more than five minutes old. That's all +you need to do. +.PP +In other words, the code you're typing is surrounded by an implicit context, +which looks like this: +.PP +.Vb 4 +\& sub filter { +\& my ( $set ) = @_; +\& # YOUR CODE HERE +\& } +.Ve +.PP +If your filter doesn't work, or if something else suddenly behaves differently, +you might have made an error in your filter, and innotop is silently catching +the error. Try enabling \*(L"debug\*(R" to make innotop throw an error instead. +.PP +\fIQUICK-FILTERS\fR +.IX Subsection "QUICK-FILTERS" +.PP +innotop's quick-filters are a shortcut to create a temporary filter that doesn't +persist when you restart innotop. To create a quick-filter, press the '/' key. +innotop will prompt you for the column name and filter text. Again, you can use +auto-completion on column names. The filter text can be just the text you want +to \*(L"search for.\*(R" For example, to filter the \*(L"processlist\*(R" table on queries +that refer to the products table, type '/' and then 'info product'. Internally, +the filter is compiled into a subroutine like this: +.PP +.Vb 4 +\& sub filter { +\& my ( $set ) = @_; +\& $set\->{info} =~ m/product/; +\& } +.Ve +.PP +The filter text can actually be any Perl regular expression, but of course a +literal string like 'product' works fine as a regular expression. +.PP +What if you want the filter to discard matching rows, rather than showing +matching rows? If you're familiar with Perl regular expressions, you might +guess how to do this. You have to use a zero-width negative lookahead +assertion. If you don't know what that means, don't worry. Let's filter out +all rows where the command is Gandalf. Type the following: +.PP +.Vb 2 +\& 1. / +\& 2. cmd ^(?!Gandalf) +.Ve +.PP +Behind the scenes innotop compiles the quick-filter into a specially tagged +filter that is otherwise like any other filter. It just isn't saved to the +configuration file. +.PP +To clear quick-filters, press the '\e' key and innotop will clear them all at +once. +.SS "\s-1SORTING\s0" +.IX Subsection "SORTING" +innotop has sensible built-in defaults to sort the most important rows to the +top of the table. Like anything else in innotop, you can customize how any +table is sorted. +.PP +To start the sort dialog, start the \*(L"\s-1TABLE EDITOR\*(R"\s0 with the '^' key, choose a +table if necessary, and press the 's' key. You'll see a list of columns you can +use in the sort expression and the current sort expression, if any. Enter a +list of columns by which you want to sort and press Enter. If you want to +reverse sort, prefix the column name with a minus sign. For example, if you +want to sort by column a ascending, then column b descending, type 'a \-b'. You +can also explicitly add a + in front of columns you want to sort ascending, but +it's not required. +.PP +Some modes have keys mapped to open this dialog directly, and to quickly reverse +sort direction. Press '?' as usual to see which keys are mapped in any mode. +.SS "\s-1GROUPING\s0" +.IX Subsection "GROUPING" +innotop can group, or aggregate, rows together (the terms are used +interchangeably). This is quite similar to an \s-1SQL GROUP BY\s0 clause. You can +specify to group on certain columns, or if you don't specify any, the entire set +of rows is treated as one group. This is quite like \s-1SQL\s0 so far, but unlike \s-1SQL,\s0 +you can also select un-grouped columns. innotop actually aggregates every +column. If you don't explicitly specify a grouping function, the default is +\&'first'. This is basically a convenience so you don't have to specify an +aggregate function for every column you want in the result. +.PP +You can quickly toggle grouping on a table with the '=' key, which toggles its +aggregate property. This property doesn't persist to the config file. +.PP +The columns by which the table is grouped are specified in its group_by +property. When you turn grouping on, innotop places the group_by columns at the +far left of the table, even if they're not supposed to be visible. The rest of +the visible columns appear in order after them. +.PP +Two tables have default group_by lists and a count column built in: +\&\*(L"processlist\*(R" and \*(L"innodb_transactions\*(R". The grouping is by connection +and status, so you can quickly see how many queries or transactions are in a +given status on each server you're monitoring. The time columns are aggregated +as a sum; other columns are left at the default 'first' aggregation. +.PP +By default, the table shown in \*(L"S: Variables & Status\*(R" mode also uses +grouping so you can monitor variables and status across many servers. The +default aggregation function in this mode is 'avg'. +.PP +Valid grouping functions are defined in the \f(CW%agg_funcs\fR hash. They include +.IP "first" 4 +.IX Item "first" +Returns the first element in the group. +.IP "count" 4 +.IX Item "count" +Returns the number of elements in the group, including undefined elements, much +like \s-1SQL\s0's \s-1COUNT\s0(*). +.IP "avg" 4 +.IX Item "avg" +Returns the average of defined elements in the group. +.IP "sum" 4 +.IX Item "sum" +Returns the sum of elements in the group. +.PP +Here's an example of grouping at work. Suppose you have a very busy server with +hundreds of open connections, and you want to see how many connections are in +what status. Using the built-in grouping rules, you can press 'Q' to enter +\&\*(L"Q: Query List\*(R" mode. Press '=' to toggle grouping (if necessary, select the +\&\*(L"processlist\*(R" table when prompted). +.PP +Your display might now look like the following: +.PP +.Vb 1 +\& Query List (? for help) localhost, 32:33, 0.11 QPS, 1 thd, 5.0.38\-log +\& +\& CXN Cmd Cnt ID User Host Time Query +\& localhost Query 49 12933 webusr localhost 19:38 SELECT * FROM +\& localhost Sending Da 23 2383 webusr localhost 12:43 SELECT col1, +\& localhost Sleep 120 140 webusr localhost 5:18:12 +\& localhost Statistics 12 19213 webusr localhost 01:19 SELECT * FROM +.Ve +.PP +That's actually quite a worrisome picture. You've got a lot of idle connections +(Sleep), and some connections executing queries (Query and Sending Data). +That's okay, but you also have a lot in Statistics status, collectively spending +over a minute. That means the query optimizer is having a really hard time +generating execution plans for your statements. Something is wrong; it should +normally take milliseconds to plan queries. You might not have seen this pattern if you +didn't look at your connections in aggregate. (This is a made-up example, but +it can happen in real life). +.SS "\s-1PIVOTING\s0" +.IX Subsection "PIVOTING" +innotop can pivot a table for more compact display, similar to a Pivot Table in +a spreadsheet (also known as a crosstab). Pivoting a table makes columns into +rows. Assume you start with this table: +.PP +.Vb 4 +\& foo bar +\& === === +\& 1 3 +\& 2 4 +.Ve +.PP +After pivoting, the table will look like this: +.PP +.Vb 4 +\& name set0 set1 +\& ==== ==== ==== +\& foo 1 2 +\& bar 3 4 +.Ve +.PP +To get reasonable results, you might need to group as well as pivoting. +innotop currently does this for \*(L"S: Variables & Status\*(R" mode. +.SS "\s-1COLORS\s0" +.IX Subsection "COLORS" +By default, innotop highlights rows with color so you can see at a glance which +rows are more important. You can customize the colorization rules and add your +own to any table. Open the table editor with the '^' key, choose a table if +needed, and press 'o' to open the color editor dialog. +.PP +The color editor dialog displays the rules applied to the table, in the order +they are evaluated. Each row is evaluated against each rule to see if the rule +matches the row; if it does, the row gets the specified color, and no further +rules are evaluated. The rules look like the following: +.PP +.Vb 9 +\& state eq Locked black on_red +\& cmd eq Sleep white +\& user eq system user white +\& cmd eq Connect white +\& cmd eq Binlog Dump white +\& time > 600 red +\& time > 120 yellow +\& time > 60 green +\& time > 30 cyan +.Ve +.PP +This is the default rule set for the \*(L"processlist\*(R" table. In order of +priority, these rules make locked queries black on a red background, \*(L"gray out\*(R" +connections from replication and sleeping queries, and make queries turn from +cyan to red as they run longer. +.PP +(For some reason, the \s-1ANSI\s0 color code \*(L"white\*(R" is actually a light gray. Your +terminal's display may vary; experiment to find colors you like). +.PP +You can use keystrokes to move the rules up and down, which re-orders their +priority. You can also delete rules and add new ones. If you add a new rule, +innotop prompts you for the column, an operator for the comparison, a value +against which to compare the column, and a color to assign if the rule matches. +There is auto-completion and prompting at each step. +.PP +The value in the third step needs to be correctly quoted. innotop does not try +to quote the value because it doesn't know whether it should treat the value as +a string or a number. If you want to compare the column against a string, as +for example in the first rule above, you should enter 'Locked' surrounded by +quotes. If you get an error message about a bareword, you probably should have +quoted something. +.SS "\s-1EXPRESSIONS\s0" +.IX Subsection "EXPRESSIONS" +Expressions are at the core of how innotop works, and are what enables you to +extend innotop as you wish. Recall the table lifecycle explained in +\&\*(L"\s-1TABLES\*(R"\s0. Expressions are used in the earliest step, where it extracts +values from a data source to form rows. +.PP +It does this by calling a subroutine for each column, passing it the source data +set, a set of current values, and a set of previous values. These are all +needed so the subroutine can calculate things like the difference between this +tick and the previous tick. +.PP +The subroutines that extract the data from the set are compiled from +expressions. This gives significantly more power than just naming the values to +fill the columns, because it allows the column's value to be calculated from +whatever data is necessary, but avoids the need to write complicated and lengthy +Perl code. +.PP +innotop begins with a string of text that can look as simple as a value's name +or as complicated as a full-fledged Perl expression. It looks at each +\&'bareword' token in the string and decides whether it's supposed to be a key +into the \f(CW$set\fR hash. A bareword is an unquoted value that isn't already +surrounded by code-ish things like dollar signs or curly brackets. If innotop +decides that the bareword isn't a function or other valid Perl code, it converts +it into a hash access. After the whole string is processed, innotop compiles a +subroutine, like this: +.PP +.Vb 5 +\& sub compute_column_value { +\& my ( $set, $cur, $pre ) = @_; +\& my $val = # EXPANDED STRING GOES HERE +\& return $val; +\& } +.Ve +.PP +Here's a concrete example, taken from the header table \*(L"q_header\*(R" in \*(L"Q: +Query List\*(R" mode. This expression calculates the qps, or Queries Per Second, +column's values, from the values returned by \s-1SHOW STATUS:\s0 +.PP +.Vb 1 +\& Questions/Uptime_hires +.Ve +.PP +innotop decides both words are barewords, and transforms this expression into +the following Perl code: +.PP +.Vb 1 +\& $set\->{Questions}/$set\->{Uptime_hires} +.Ve +.PP +When surrounded by the rest of the subroutine's code, this is executable Perl +that calculates a high-resolution queries-per-second value. +.PP +The arguments to the subroutine are named \f(CW$set\fR, \f(CW$cur\fR, and \f(CW$pre\fR. In most cases, +\&\f(CW$set\fR and \f(CW$cur\fR will be the same values. However, if \*(L"status_inc\*(R" is set, \f(CW$cur\fR +will not be the same as \f(CW$set\fR, because \f(CW$set\fR will already contain values that are +the incremental difference between \f(CW$cur\fR and \f(CW$pre\fR. +.PP +Every column in innotop is computed by subroutines compiled in the same fashion. +There is no difference between innotop's built-in columns and user-defined +columns. This keeps things consistent and predictable. +.SS "\s-1TRANSFORMATIONS\s0" +.IX Subsection "TRANSFORMATIONS" +Transformations change how a value is rendered. For example, they can take a +number of seconds and display it in H:M:S format. The following transformations +are defined: +.IP "commify" 4 +.IX Item "commify" +Adds commas to large numbers every three decimal places. +.IP "distill" 4 +.IX Item "distill" +Distills \s-1SQL\s0 into verb-noun-noun format for quick comprehension. +.IP "dulint_to_int" 4 +.IX Item "dulint_to_int" +Accepts two unsigned integers and converts them into a single longlong. This is +useful for certain operations with InnoDB, which uses two integers as +transaction identifiers, for example. +.IP "fuzzy_time" 4 +.IX Item "fuzzy_time" +Converts a number of seconds into a friendly, readable value like \*(L"1h35m\*(R". +.IP "no_ctrl_char" 4 +.IX Item "no_ctrl_char" +Removes quoted control characters from the value. This is affected by the +\&\*(L"charset\*(R" configuration variable. +.Sp +This transformation only operates within quoted strings, for example, values to +a \s-1SET\s0 clause in an \s-1UPDATE\s0 statement. It will not alter the \s-1UPDATE\s0 statement, +but will collapse the quoted string to [\s-1BINARY\s0] or [\s-1TEXT\s0], depending on the +charset. +.IP "percent" 4 +.IX Item "percent" +Converts a number to a percentage by multiplying it by two, formatting it with +\&\*(L"num_digits\*(R" digits after the decimal point, and optionally adding a percent +sign (see \*(L"show_percent\*(R"). +.IP "secs_to_time" 4 +.IX Item "secs_to_time" +Formats a number of seconds as time in days+hours:minutes:seconds format. +.IP "set_precision" 4 +.IX Item "set_precision" +Formats numbers with \*(L"num_digits\*(R" number of digits after the decimal point. +.IP "shorten" 4 +.IX Item "shorten" +Formats a number as a unit of 1024 (k/M/G/T) and with \*(L"num_digits\*(R" number of +digits after the decimal point. +.SS "\s-1TABLE EDITOR\s0" +.IX Subsection "TABLE EDITOR" +The innotop table editor lets you customize tables with keystrokes. You start +the table editor with the '^' key. If there's more than one table on the +screen, it will prompt you to choose one of them. Once you do, innotop will +show you something like this: +.PP +.Vb 1 +\& Editing table definition for Buffer Pool. Press ? for help, q to quit. +\& +\& name hdr label src +\& cxn CXN Connection from which cxn +\& buf_pool_size Size Buffer pool size IB_bp_buf_poo +\& buf_free Free Bufs Buffers free in the b IB_bp_buf_fre +\& pages_total Pages Pages total IB_bp_pages_t +\& pages_modified Dirty Pages Pages modified (dirty IB_bp_pages_m +\& buf_pool_hit_rate Hit Rate Buffer pool hit rate IB_bp_buf_poo +\& total_mem_alloc Memory Total memory allocate IB_bp_total_m +\& add_pool_alloc Add\*(Aql Pool Additional pool alloca IB_bp_add_poo +.Ve +.PP +The first line shows which table you're editing, and reminds you again to press +\&'?' for a list of key mappings. The rest is a tabular representation of the +table's columns, because that's likely what you're trying to edit. However, you +can edit more than just the table's columns; this screen can start the filter +editor, color rule editor, and more. +.PP +Each row in the display shows a single column in the table you're editing, along +with a couple of its properties such as its header and source expression (see +\&\*(L"\s-1EXPRESSIONS\*(R"\s0). +.PP +The key mappings are Vim-style, as in many other places. Pressing 'j' and 'k' +moves the highlight up or down. You can then (d)elete or (e)dit the highlighted +column. You can also (a)dd a column to the table. This actually just activates +one of the columns already defined for the table; it prompts you to choose from +among the columns available but not currently displayed. Finally, you can +re-order the columns with the '+' and '\-' keys. +.PP +You can do more than just edit the columns with the table editor, you can also +edit other properties, such as the table's sort expression and group-by +expression. Press '?' to see the full list, of course. +.PP +If you want to really customize and create your own column, as opposed to just +activating a built-in one that's not currently displayed, press the (n)ew key, +and innotop will prompt you for the information it needs: +.IP "\(bu" 4 +The column name: this needs to be a word without any funny characters, e.g. just +letters, numbers and underscores. +.IP "\(bu" 4 +The column header: this is the label that appears at the top of the column, in +the table header. This can have spaces and funny characters, but be careful not +to make it too wide and waste space on-screen. +.IP "\(bu" 4 +The column's data source: this is an expression that determines what data from +the source (see \*(L"\s-1TABLES\*(R"\s0) innotop will put into the column. This can just be +the name of an item in the source, or it can be a more complex expression, as +described in \*(L"\s-1EXPRESSIONS\*(R"\s0. +.PP +Once you've entered the required data, your table has a new column. There is no +difference between this column and the built-in ones; it can have all the same +properties and behaviors. innotop will write the column's definition to the +configuration file, so it will persist across sessions. +.PP +Here's an example: suppose you want to track how many times your slaves have +retried transactions. According to the MySQL manual, the +Slave_retried_transactions status variable gives you that data: \*(L"The total +number of times since startup that the replication slave \s-1SQL\s0 thread has retried +transactions. This variable was added in version 5.0.4.\*(R" This is appropriate to +add to the \*(L"slave_sql_status\*(R" table. +.PP +To add the column, switch to the replication-monitoring mode with the 'M' key, +and press the '^' key to start the table editor. When prompted, choose +slave_sql_status as the table, then press 'n' to create the column. Type +\&'retries' as the column name, 'Retries' as the column header, and +\&'Slave_retried_transactions' as the source. Now the column is created, and you +see the table editor screen again. Press 'q' to exit the table editor, and +you'll see your column at the end of the table. +.SH "VARIABLE SETS" +.IX Header "VARIABLE SETS" +Variable sets are used in \*(L"S: Variables & Status\*(R" mode to define more easily +what variables you want to monitor. Behind the scenes they are compiled to a +list of expressions, and then into a column list so they can be treated just +like columns in any other table, in terms of data extraction and +transformations. However, you're protected from the tedious details by a syntax +that ought to feel very natural to you: a \s-1SQL SELECT\s0 list. +.PP +The data source for variable sets, and indeed the entire S mode, is the +combination of \s-1SHOW STATUS, SHOW VARIABLES,\s0 and \s-1SHOW INNODB STATUS. \s0 Imagine +that you had a huge table with one column per variable returned from those +statements. That's the data source for variable sets. You can now query this +data source just like you'd expect. For example: +.PP +.Vb 1 +\& Questions, Uptime, Questions/Uptime as QPS +.Ve +.PP +Behind the scenes innotop will split that variable set into three expressions, +compile them and turn them into a table definition, then extract as usual. This +becomes a \*(L"variable set,\*(R" or a \*(L"list of variables you want to monitor.\*(R" +.PP +innotop lets you name and save your variable sets, and writes them to the +configuration file. You can choose which variable set you want to see with the +\&'c' key, or activate the next and previous sets with the '>' and '<' keys. +There are many built-in variable sets as well, which should give you a good +start for creating your own. Press 'e' to edit the current variable set, or +just to see how it's defined. To create a new one, just press 'c' and type its +name. +.PP +You may want to use some of the functions listed in \*(L"\s-1TRANSFORMATIONS\*(R"\s0 to help +format the results. In particular, \*(L"set_precision\*(R" is often useful to limit +the number of digits you see. Extending the above example, here's how: +.PP +.Vb 1 +\& Questions, Uptime, set_precision(Questions/Uptime) as QPS +.Ve +.PP +Actually, this still needs a little more work. If your \*(L"interval\*(R" is less +than one second, you might be dividing by zero because Uptime is incremental in +this mode by default. Instead, use Uptime_hires: +.PP +.Vb 1 +\& Questions, Uptime, set_precision(Questions/Uptime_hires) as QPS +.Ve +.PP +This example is simple, but it shows how easy it is to choose which variables +you want to monitor. +.SH "PLUGINS" +.IX Header "PLUGINS" +innotop has a simple but powerful plugin mechanism by which you can extend +or modify its existing functionality, and add new functionality. innotop's +plugin functionality is event-based: plugins register themselves to be called +when events happen. They then have a chance to influence the event. +.PP +An innotop plugin is a Perl module (.pm) file placed in innotop's \*(L"plugin_dir\*(R" +directory. On \s-1UNIX\s0 systems, you can place a symbolic link to the module instead +of putting the actual file there. innotop automatically discovers files named \f(CW\*(C`*.pm\*(C'\fR. If +there is a corresponding entry in the \*(L"plugins\*(R" configuration file section, +innotop loads and activates the plugin. +.PP +The module must conform to innotop's plugin interface. Additionally, the source +code of the module must be written in such a way that innotop can inspect the +file and determine the package name and description. +.SS "Package Source Convention" +.IX Subsection "Package Source Convention" +innotop inspects the plugin module's source to determine the Perl package name. +It looks for a line of the form \*(L"package Foo;\*(R" and if found, considers the +plugin's package name to be Foo. Of course the package name can be a valid Perl +package name such as Foo::Bar, with double colons (::) and so on. +.PP +It also looks for a description in the source code, to make the plugin editor +more human-friendly. The description is a comment line of the form \*(L"# +description: Foo\*(R", where \*(L"Foo\*(R" is the text innotop will consider to be the +plugin's description. +.SS "Plugin Interface" +.IX Subsection "Plugin Interface" +The innotop plugin interface is quite simple: innotop expects the plugin to be +an object-oriented module it can call certain methods on. The methods are +.IP "new(%variables)" 4 +.IX Item "new(%variables)" +This is the plugin's constructor. It is passed a hash of innotop's variables, +which it can manipulate (see \*(L"Plugin Variables\*(R"). It must return a reference +to the newly created plugin object. +.Sp +At construction time, innotop has only loaded the general configuration and +created the default built-in variables with their default contents (which is +quite a lot). Therefore, the state of the program is exactly as in the innotop +source code, plus the configuration variables from the \*(L"general\*(R" section in +the config file. +.Sp +If your plugin manipulates the variables, it is changing global data, which is +shared by innotop and all plugins. Plugins are loaded in the order they're +listed in the config file. Your plugin may load before or after another plugin, +so there is a potential for conflict or interaction between plugins if they +modify data other plugins use or modify. +.IP "\fIregister_for_events()\fR" 4 +.IX Item "register_for_events()" +This method must return a list of events in which the plugin is interested, if +any. See \*(L"Plugin Events\*(R" for the defined events. If the plugin returns an +event that's not defined, the event is ignored. +.IP "event handlers" 4 +.IX Item "event handlers" +The plugin must implement a method named the same as each event for which it has +registered. In other words, if the plugin returns qw(foo bar) from +\&\fIregister_for_events()\fR, it must have \fIfoo()\fR and \fIbar()\fR methods. These methods are +callbacks for the events. See \*(L"Plugin Events\*(R" for more details about each +event. +.SS "Plugin Variables" +.IX Subsection "Plugin Variables" +The plugin's constructor is passed a hash of innotop's variables, which it can +manipulate. It is probably a good idea if the plugin object saves a copy of it +for later use. The variables are defined in the innotop variable +\&\f(CW%pluggable_vars\fR, and are as follows: +.IP "action_for" 4 +.IX Item "action_for" +A hashref of key mappings. These are innotop's global hot-keys. +.IP "agg_funcs" 4 +.IX Item "agg_funcs" +A hashref of functions that can be used for grouping. See \*(L"\s-1GROUPING\*(R"\s0. +.IP "config" 4 +.IX Item "config" +The global configuration hash. +.IP "connections" 4 +.IX Item "connections" +A hashref of connection specifications. These are just specifications of how to +connect to a server. +.IP "dbhs" 4 +.IX Item "dbhs" +A hashref of innotop's database connections. These are actual \s-1DBI\s0 connection +objects. +.IP "filters" 4 +.IX Item "filters" +A hashref of filters applied to table rows. See \*(L"\s-1FILTERS\*(R"\s0 for more. +.IP "modes" 4 +.IX Item "modes" +A hashref of modes. See \*(L"\s-1MODES\*(R"\s0 for more. +.IP "server_groups" 4 +.IX Item "server_groups" +A hashref of server groups. See \*(L"\s-1SERVER GROUPS\*(R"\s0. +.IP "tbl_meta" 4 +.IX Item "tbl_meta" +A hashref of innotop's table meta-data, with one entry per table (see +\&\*(L"\s-1TABLES\*(R"\s0 for more information). +.IP "trans_funcs" 4 +.IX Item "trans_funcs" +A hashref of transformation functions. See \*(L"\s-1TRANSFORMATIONS\*(R"\s0. +.IP "var_sets" 4 +.IX Item "var_sets" +A hashref of variable sets. See \*(L"\s-1VARIABLE SETS\*(R"\s0. +.SS "Plugin Events" +.IX Subsection "Plugin Events" +Each event is defined somewhere in the innotop source code. When innotop runs +that code, it executes the callback function for each plugin that expressed its +interest in the event. innotop passes some data for each event. The events are +defined in the \f(CW%event_listener_for\fR variable, and are as follows: +.ie n .IP "extract_values($set, $cur, $pre, $tbl)" 4 +.el .IP "extract_values($set, \f(CW$cur\fR, \f(CW$pre\fR, \f(CW$tbl\fR)" 4 +.IX Item "extract_values($set, $cur, $pre, $tbl)" +This event occurs inside the function that extracts values from a data source. +The arguments are the set of values, the current values, the previous values, +and the table name. +.IP "set_to_tbl" 4 +.IX Item "set_to_tbl" +Events are defined at many places in this subroutine, which is responsible for +turning an arrayref of hashrefs into an arrayref of lines that can be printed to +the screen. The events all pass the same data: an arrayref of rows and the name +of the table being created. The events are set_to_tbl_pre_filter, +set_to_tbl_pre_sort,set_to_tbl_pre_group, set_to_tbl_pre_colorize, +set_to_tbl_pre_transform, set_to_tbl_pre_pivot, set_to_tbl_pre_create, +set_to_tbl_post_create. +.IP "draw_screen($lines)" 4 +.IX Item "draw_screen($lines)" +This event occurs inside the subroutine that prints the lines to the screen. +\&\f(CW$lines\fR is an arrayref of strings. +.SS "Simple Plugin Example" +.IX Subsection "Simple Plugin Example" +The easiest way to explain the plugin functionality is probably with a simple +example. The following module adds a column to the beginning of every table and +sets its value to 1. (If you copy and paste this example code, be sure to remove +the first space from each line; lines such as '# description' must not start with +whitespace). +.PP +.Vb 2 +\& use strict; +\& use warnings FATAL => \*(Aqall\*(Aq; +\& +\& package Innotop::Plugin::Example; +\& # description: Adds an \*(Aqexample\*(Aq column to every table +\& +\& sub new { +\& my ( $class, %vars ) = @_; +\& # Store reference to innotop\*(Aqs variables in $self +\& my $self = bless { %vars }, $class; +\& +\& # Design the example column +\& my $col = { +\& hdr => \*(AqExample\*(Aq, +\& just => \*(Aq\*(Aq, +\& dec => 0, +\& num => 1, +\& label => \*(AqExample\*(Aq, +\& src => \*(Aqexample\*(Aq, # Get data from this column in the data source +\& tbl => \*(Aq\*(Aq, +\& trans => [], +\& }; +\& +\& # Add the column to every table. +\& my $tbl_meta = $vars{tbl_meta}; +\& foreach my $tbl ( values %$tbl_meta ) { +\& # Add the column to the list of defined columns +\& $tbl\->{cols}\->{example} = $col; +\& # Add the column to the list of visible columns +\& unshift @{$tbl\->{visible}}, \*(Aqexample\*(Aq; +\& } +\& +\& # Be sure to return a reference to the object. +\& return $self; +\& } +\& +\& # I\*(Aqd like to be called when a data set is being rendered into a table, please. +\& sub register_for_events { +\& my ( $self ) = @_; +\& return qw(set_to_tbl_pre_filter); +\& } +\& +\& # This method will be called when the event fires. +\& sub set_to_tbl_pre_filter { +\& my ( $self, $rows, $tbl ) = @_; +\& # Set the example column\*(Aqs data source to the value 1. +\& foreach my $row ( @$rows ) { +\& $row\->{example} = 1; +\& } +\& } +\& +\& 1; +.Ve +.SS "Plugin Editor" +.IX Subsection "Plugin Editor" +The plugin editor lets you view the plugins innotop discovered and activate or +deactivate them. Start the editor by pressing $ to start the configuration +editor from any mode. Press the 'p' key to start the plugin editor. You'll see +a list of plugins innotop discovered. You can use the 'j' and 'k' keys to move +the highlight to the desired one, then press the * key to toggle it active or +inactive. Exit the editor and restart innotop for the changes to take effect. +.SH "SQL STATEMENTS" +.IX Header "SQL STATEMENTS" +innotop uses a limited set of \s-1SQL\s0 statements to retrieve data from MySQL for +display. The statements are customized depending on the server version against +which they are executed; for example, on MySQL 5 and newer, \s-1INNODB_STATUS\s0 +executes \*(L"\s-1SHOW ENGINE INNODB STATUS\*(R",\s0 while on earlier versions it executes +\&\*(L"\s-1SHOW INNODB STATUS\*(R". \s0 The statements are as follows: +.PP +.Vb 10 +\& Statement SQL executed +\& =================== =============================== +\& INDEX_STATISTICS SELECT * FROM INFORMATION_SCHEMA.INDEX_STATISTICS +\& INNODB_STATUS SHOW [ENGINE] INNODB STATUS +\& KILL_CONNECTION KILL +\& KILL_QUERY KILL QUERY +\& OPEN_TABLES SHOW OPEN TABLES +\& PROCESSLIST SHOW FULL PROCESSLIST +\& SHOW_MASTER_LOGS SHOW MASTER LOGS +\& SHOW_MASTER_STATUS SHOW MASTER STATUS +\& SHOW_SLAVE_STATUS SHOW SLAVE STATUS +\& SHOW_STATUS SHOW [GLOBAL] STATUS +\& SHOW_VARIABLES SHOW [GLOBAL] VARIABLES +\& TABLE_STATISTICS SELECT * FROM INFORMATION_SCHEMA.TABLE_STATISTICS +.Ve +.SH "DATA SOURCES" +.IX Header "DATA SOURCES" +Each time innotop extracts values to create a table (see \*(L"\s-1EXPRESSIONS\*(R"\s0 and +\&\*(L"\s-1TABLES\*(R"\s0), it does so from a particular data source. Largely because of the +complex data extracted from \s-1SHOW INNODB STATUS,\s0 this is slightly messy. \s-1SHOW +INNODB STATUS\s0 contains a mixture of single values and repeated values that form +nested data sets. +.PP +Whenever innotop fetches data from MySQL, it adds two extra bits to each set: +cxn and Uptime_hires. cxn is the name of the connection from which the data +came. Uptime_hires is a high-resolution version of the server's Uptime status +variable, which is important if your \*(L"interval\*(R" setting is sub-second. +.PP +Here are the kinds of data sources from which data is extracted: +.IP "\s-1STATUS_VARIABLES\s0" 4 +.IX Item "STATUS_VARIABLES" +This is the broadest category, into which the most kinds of data fall. It +begins with the combination of \s-1SHOW STATUS\s0 and \s-1SHOW VARIABLES,\s0 but other sources +may be included as needed, for example, \s-1SHOW MASTER STATUS\s0 and \s-1SHOW SLAVE +STATUS,\s0 as well as many of the non-repeated values from \s-1SHOW INNODB STATUS.\s0 +.IP "\s-1DEADLOCK_LOCKS\s0" 4 +.IX Item "DEADLOCK_LOCKS" +This data is extracted from the transaction list in the \s-1LATEST DETECTED DEADLOCK\s0 +section of \s-1SHOW INNODB STATUS. \s0 It is nested two levels deep: transactions, then +locks. +.IP "\s-1DEADLOCK_TRANSACTIONS\s0" 4 +.IX Item "DEADLOCK_TRANSACTIONS" +This data is from the transaction list in the \s-1LATEST DETECTED DEADLOCK\s0 +section of \s-1SHOW INNODB STATUS. \s0 It is nested one level deep. +.IP "\s-1EXPLAIN\s0" 4 +.IX Item "EXPLAIN" +This data is from the result set returned by \s-1EXPLAIN.\s0 +.IP "\s-1INNODB_BLOCKED_BLOCKER\s0" 4 +.IX Item "INNODB_BLOCKED_BLOCKER" +This data is from the \s-1INFORMATION_SCHEMA\s0 tables related to InnoDB locks and +the processlist. +.IP "\s-1INNODB_TRANSACTIONS\s0" 4 +.IX Item "INNODB_TRANSACTIONS" +This data is from the \s-1TRANSACTIONS\s0 section of \s-1SHOW INNODB STATUS.\s0 +.IP "\s-1IO_THREADS\s0" 4 +.IX Item "IO_THREADS" +This data is from the list of threads in the \s-1FILE I/O\s0 section of \s-1SHOW INNODB +STATUS.\s0 +.IP "\s-1INNODB_LOCKS\s0" 4 +.IX Item "INNODB_LOCKS" +This data is from the \s-1TRANSACTIONS\s0 section of \s-1SHOW INNODB STATUS\s0 and is nested +two levels deep. +.IP "\s-1MASTER_SLAVE\s0" 4 +.IX Item "MASTER_SLAVE" +This data is from the combination of \s-1SHOW MASTER STATUS\s0 and \s-1SHOW SLAVE STATUS.\s0 +.IP "\s-1OPEN_TABLES\s0" 4 +.IX Item "OPEN_TABLES" +This data is from \s-1SHOW OPEN TABLES.\s0 +.IP "\s-1PROCESSLIST\s0" 4 +.IX Item "PROCESSLIST" +This data is from \s-1SHOW FULL PROCESSLIST.\s0 +.IP "\s-1PROCESSLIST_STATS\s0" 4 +.IX Item "PROCESSLIST_STATS" +This data is from \s-1SHOW FULL PROCESSLIST\s0 and computes stats such as the maximum time +a user query has been running, and how many user queries are running. A \*(L"user +query\*(R" excludes replication threads. +.IP "\s-1OS_WAIT_ARRAY\s0" 4 +.IX Item "OS_WAIT_ARRAY" +This data is from the \s-1SEMAPHORES\s0 section of \s-1SHOW INNODB STATUS\s0 and is nested one +level deep. It comes from the lines that look like this: +.Sp +.Vb 1 +\& \-\-Thread 1568861104 has waited at btr0cur.c line 424 .... +.Ve +.SH "MYSQL PRIVILEGES" +.IX Header "MYSQL PRIVILEGES" +.IP "\(bu" 4 +You must connect to MySQL as a user who has the \s-1SUPER\s0 privilege for many of the +functions. +.IP "\(bu" 4 +If you don't have the \s-1SUPER\s0 privilege, you can still run some functions, but you +won't necessarily see all the same data. +.IP "\(bu" 4 +You need the \s-1PROCESS\s0 privilege to see the list of currently running queries in Q +mode. +.IP "\(bu" 4 +You need special privileges to start and stop slave servers. +.IP "\(bu" 4 +You need appropriate privileges to create and drop the deadlock tables if needed +(see \*(L"\s-1SERVER CONNECTIONS\*(R"\s0). +.SH "SYSTEM REQUIREMENTS" +.IX Header "SYSTEM REQUIREMENTS" +You need Perl to run innotop, of course. You also need a few Perl modules: \s-1DBI,\s0 +DBD::MariaDB, Term::ReadKey, and Time::HiRes. These should be included with most +Perl distributions, but in case they are not, I recommend using versions +distributed with your operating system or Perl distribution, not from \s-1CPAN.\s0 +Term::ReadKey in particular has been known to cause problems if installed from +\&\s-1CPAN.\s0 +.PP +If you have Term::ANSIColor, innotop will use it to format headers more readably +and compactly. (Under Microsoft Windows, you also need Win32::Console::ANSI for +terminal formatting codes to be honored). If you install Term::ReadLine, +preferably Term::ReadLine::Gnu, you'll get nice auto-completion support. +.PP +I run innotop on Gentoo GNU/Linux, Debian and Ubuntu, and I've had feedback from +people successfully running it on Red Hat, CentOS, Solaris, and Mac \s-1OSX. I\s0 +don't see any reason why it won't work on other UNIX-ish operating systems, but +I don't know for sure. It also runs on Windows under ActivePerl without +problem. +.PP +innotop has been used on MySQL versions 3.23.58, 4.0.27, 4.1.0, 4.1.22, 5.0.26, +5.1.15, and 5.2.3. If it doesn't run correctly for you, that is a bug that +should be reported. +.SH "FILES" +.IX Header "FILES" +\&\f(CW$HOMEDIR\fR/.innotop and/or /etc/innotop are used to store +configuration information. Files include the configuration file innotop.conf, +the core_dump file which contains verbose error messages if \*(L"debug\*(R" is +enabled, and the plugins/ subdirectory. +.SH "GLOSSARY OF TERMS" +.IX Header "GLOSSARY OF TERMS" +.IP "tick" 4 +.IX Item "tick" +A tick is a refresh event, when innotop re-fetches data from connections and +displays it. +.SH "ACKNOWLEDGEMENTS" +.IX Header "ACKNOWLEDGEMENTS" +The following people and organizations are acknowledged for various reasons. +Hopefully no one has been forgotten. +.PP +Aaron Racine, +Allen K. Smith, +Aurimas Mikalauskas, +Bartosz Fenski, +Brian Miezejewski, +Christian Hammers, +Cyril Scetbon, +Dane Miller, +David Multer, +Dr. Frank Ullrich, +Giuseppe Maxia, +Google.com Site Reliability Engineers, +Google Code, +Jan Pieter Kunst, +Jari Aalto, +Jay Pipes, +Jeremy Zawodny, +Johan Idren, +Kristian Kohntopp, +Lenz Grimmer, +Maciej Dobrzanski, +Michiel Betel, +MySQL \s-1AB,\s0 +Paul McCullagh, +Sebastien Estienne, +Sourceforge.net, +Steven Kreuzer, +The Gentoo MySQL Team, +Trevor Price, +Yaar Schnitman, +and probably more people that have not been included. +.PP +(If your name has been misspelled, it's probably out of fear of putting +international characters into this documentation; earlier versions of Perl might +not be able to compile it then). +.SH "COPYRIGHT, LICENSE AND WARRANTY" +.IX Header "COPYRIGHT, LICENSE AND WARRANTY" +This program is copyright (c) 2006 Baron Schwartz. +Feedback and improvements are welcome. +.PP +\&\s-1THIS PROGRAM IS PROVIDED \*(L"AS IS\*(R" AND WITHOUT ANY EXPRESS OR IMPLIED +WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF +MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.\s0 +.PP +This program is free software; you can redistribute it and/or modify it under +the terms of the \s-1GNU\s0 General Public License as published by the Free Software +Foundation, version 2; \s-1OR\s0 the Perl Artistic License. On \s-1UNIX\s0 and similar +systems, you can issue `man perlgpl' or `man perlartistic' to read these +licenses. +.PP +You should have received a copy of the \s-1GNU\s0 General Public License along with +this program; if not, write to the Free Software Foundation, Inc., 51 Franklin +Street, Fifth Floor, Boston, \s-1MA 02110-1335 USA\s0. +.PP +Execute innotop and press '!' to see this information at any time. +.SH "AUTHOR" +.IX Header "AUTHOR" +Originally written by Baron Schwartz; currently maintained by Aaron Racine. +.SH "BUGS" +.IX Header "BUGS" +You can report bugs, ask for improvements, and get other help and support at +<https://github.com/innotop/innotop>. There are mailing lists, a source code +browser, a bug tracker, etc. Please use these instead of contacting the +maintainer or author directly, as it makes our job easier and benefits others if the +discussions are permanent and public. Of course, if you need to contact us in +private, please do. |