diff options
Diffstat (limited to 'doc/src/sgml/html/runtime-config-developer.html')
| -rw-r--r-- | doc/src/sgml/html/runtime-config-developer.html | 464 |
1 files changed, 464 insertions, 0 deletions
diff --git a/doc/src/sgml/html/runtime-config-developer.html b/doc/src/sgml/html/runtime-config-developer.html new file mode 100644 index 0000000..1a4f239 --- /dev/null +++ b/doc/src/sgml/html/runtime-config-developer.html @@ -0,0 +1,464 @@ +<?xml version="1.0" encoding="UTF-8" standalone="no"?> +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>20.17. Developer Options</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="runtime-config-custom.html" title="20.16. Customized Options" /><link rel="next" href="runtime-config-short.html" title="20.18. Short Options" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">20.17. Developer Options</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="runtime-config-custom.html" title="20.16. Customized Options">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="runtime-config.html" title="Chapter 20. Server Configuration">Up</a></td><th width="60%" align="center">Chapter 20. Server Configuration</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 16.2 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="runtime-config-short.html" title="20.18. Short Options">Next</a></td></tr></table><hr /></div><div class="sect1" id="RUNTIME-CONFIG-DEVELOPER"><div class="titlepage"><div><div><h2 class="title" style="clear: both">20.17. Developer Options <a href="#RUNTIME-CONFIG-DEVELOPER" class="id_link">#</a></h2></div></div></div><p> + The following parameters are intended for developer testing, and + should never be used on a production database. However, some of + them can be used to assist with the recovery of severely damaged + databases. As such, they have been excluded from the sample + <code class="filename">postgresql.conf</code> file. Note that many of these + parameters require special source compilation flags to work at all. + </p><div class="variablelist"><dl class="variablelist"><dt id="GUC-ALLOW-IN-PLACE-TABLESPACES"><span class="term"><code class="varname">allow_in_place_tablespaces</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.1.1.3" class="indexterm"></a> + </span> <a href="#GUC-ALLOW-IN-PLACE-TABLESPACES" class="id_link">#</a></dt><dd><p> + Allows tablespaces to be created as directories inside + <code class="filename">pg_tblspc</code>, when an empty location string + is provided to the <code class="command">CREATE TABLESPACE</code> command. This + is intended to allow testing replication scenarios where primary and + standby servers are running on the same machine. Such directories + are likely to confuse backup tools that expect to find only symbolic + links in that location. + Only superusers and users with the appropriate <code class="literal">SET</code> + privilege can change this setting. + </p></dd><dt id="GUC-ALLOW-SYSTEM-TABLE-MODS"><span class="term"><code class="varname">allow_system_table_mods</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.2.1.3" class="indexterm"></a> + </span> <a href="#GUC-ALLOW-SYSTEM-TABLE-MODS" class="id_link">#</a></dt><dd><p> + Allows modification of the structure of system tables as well as + certain other risky actions on system tables. This is otherwise not + allowed even for superusers. Ill-advised use of this setting can + cause irretrievable data loss or seriously corrupt the database + system. + Only superusers and users with the appropriate <code class="literal">SET</code> + privilege can change this setting. + </p></dd><dt id="GUC-BACKTRACE-FUNCTIONS"><span class="term"><code class="varname">backtrace_functions</code> (<code class="type">string</code>) + <a id="id-1.6.7.20.3.3.1.3" class="indexterm"></a> + </span> <a href="#GUC-BACKTRACE-FUNCTIONS" class="id_link">#</a></dt><dd><p> + This parameter contains a comma-separated list of C function names. + If an error is raised and the name of the internal C function where + the error happens matches a value in the list, then a backtrace is + written to the server log together with the error message. This can + be used to debug specific areas of the source code. + </p><p> + Backtrace support is not available on all platforms, and the quality + of the backtraces depends on compilation options. + </p><p> + Only superusers and users with the appropriate <code class="literal">SET</code> + privilege can change this setting. + </p></dd><dt id="GUC-DEBUG-DISCARD-CACHES"><span class="term"><code class="varname">debug_discard_caches</code> (<code class="type">integer</code>) + <a id="id-1.6.7.20.3.4.1.3" class="indexterm"></a> + </span> <a href="#GUC-DEBUG-DISCARD-CACHES" class="id_link">#</a></dt><dd><p> + When set to <code class="literal">1</code>, each system catalog cache entry is + invalidated at the first possible opportunity, whether or not + anything that would render it invalid really occurred. Caching of + system catalogs is effectively disabled as a result, so the server + will run extremely slowly. Higher values run the cache invalidation + recursively, which is even slower and only useful for testing + the caching logic itself. The default value of <code class="literal">0</code> + selects normal catalog caching behavior. + </p><p> + This parameter can be very helpful when trying to trigger + hard-to-reproduce bugs involving concurrent catalog changes, but it + is otherwise rarely needed. See the source code files + <code class="filename">inval.c</code> and + <code class="filename">pg_config_manual.h</code> for details. + </p><p> + This parameter is supported when + <code class="symbol">DISCARD_CACHES_ENABLED</code> was defined at compile time + (which happens automatically when using the + <span class="application">configure</span> option + <code class="option">--enable-cassert</code>). In production builds, its value + will always be <code class="literal">0</code> and attempts to set it to another + value will raise an error. + </p></dd><dt id="GUC-DEBUG-IO-DIRECT"><span class="term"><code class="varname">debug_io_direct</code> (<code class="type">string</code>) + <a id="id-1.6.7.20.3.5.1.3" class="indexterm"></a> + </span> <a href="#GUC-DEBUG-IO-DIRECT" class="id_link">#</a></dt><dd><p> + Ask the kernel to minimize caching effects for relation data and WAL + files using <code class="literal">O_DIRECT</code> (most Unix-like systems), + <code class="literal">F_NOCACHE</code> (macOS) or + <code class="literal">FILE_FLAG_NO_BUFFERING</code> (Windows). + </p><p> + May be set to an empty string (the default) to disable use of direct + I/O, or a comma-separated list of operations that should use direct I/O. + The valid options are <code class="literal">data</code> for + main data files, <code class="literal">wal</code> for WAL files, and + <code class="literal">wal_init</code> for WAL files when being initially + allocated. + </p><p> + Some operating systems and file systems do not support direct I/O, so + non-default settings may be rejected at startup or cause errors. + </p><p> + Currently this feature reduces performance, and is intended for + developer testing only. + </p></dd><dt id="GUC-DEBUG-PARALLEL-QUERY"><span class="term"><code class="varname">debug_parallel_query</code> (<code class="type">enum</code>) + <a id="id-1.6.7.20.3.6.1.3" class="indexterm"></a> + </span> <a href="#GUC-DEBUG-PARALLEL-QUERY" class="id_link">#</a></dt><dd><p> + Allows the use of parallel queries for testing purposes even in cases + where no performance benefit is expected. + The allowed values of <code class="varname">debug_parallel_query</code> are + <code class="literal">off</code> (use parallel mode only when it is expected to improve + performance), <code class="literal">on</code> (force parallel query for all queries + for which it is thought to be safe), and <code class="literal">regress</code> (like + <code class="literal">on</code>, but with additional behavior changes as explained + below). + </p><p> + More specifically, setting this value to <code class="literal">on</code> will add + a <code class="literal">Gather</code> node to the top of any query plan for which this + appears to be safe, so that the query runs inside of a parallel worker. + Even when a parallel worker is not available or cannot be used, + operations such as starting a subtransaction that would be prohibited + in a parallel query context will be prohibited unless the planner + believes that this will cause the query to fail. If failures or + unexpected results occur when this option is set, some functions used + by the query may need to be marked <code class="literal">PARALLEL UNSAFE</code> + (or, possibly, <code class="literal">PARALLEL RESTRICTED</code>). + </p><p> + Setting this value to <code class="literal">regress</code> has all of the same effects + as setting it to <code class="literal">on</code> plus some additional effects that are + intended to facilitate automated regression testing. Normally, + messages from a parallel worker include a context line indicating that, + but a setting of <code class="literal">regress</code> suppresses this line so that the + output is the same as in non-parallel execution. Also, + the <code class="literal">Gather</code> nodes added to plans by this setting are hidden + in <code class="literal">EXPLAIN</code> output so that the output matches what + would be obtained if this setting were turned <code class="literal">off</code>. + </p></dd><dt id="GUC-IGNORE-SYSTEM-INDEXES"><span class="term"><code class="varname">ignore_system_indexes</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.7.1.3" class="indexterm"></a> + </span> <a href="#GUC-IGNORE-SYSTEM-INDEXES" class="id_link">#</a></dt><dd><p> + Ignore system indexes when reading system tables (but still + update the indexes when modifying the tables). This is useful + when recovering from damaged system indexes. + This parameter cannot be changed after session start. + </p></dd><dt id="GUC-POST-AUTH-DELAY"><span class="term"><code class="varname">post_auth_delay</code> (<code class="type">integer</code>) + <a id="id-1.6.7.20.3.8.1.3" class="indexterm"></a> + </span> <a href="#GUC-POST-AUTH-DELAY" class="id_link">#</a></dt><dd><p> + The amount of time to delay when a new + server process is started, after it conducts the + authentication procedure. This is intended to give developers an + opportunity to attach to the server process with a debugger. + If this value is specified without units, it is taken as seconds. + A value of zero (the default) disables the delay. + This parameter cannot be changed after session start. + </p></dd><dt id="GUC-PRE-AUTH-DELAY"><span class="term"><code class="varname">pre_auth_delay</code> (<code class="type">integer</code>) + <a id="id-1.6.7.20.3.9.1.3" class="indexterm"></a> + </span> <a href="#GUC-PRE-AUTH-DELAY" class="id_link">#</a></dt><dd><p> + The amount of time to delay just after a + new server process is forked, before it conducts the + authentication procedure. This is intended to give developers an + opportunity to attach to the server process with a debugger to + trace down misbehavior in authentication. + If this value is specified without units, it is taken as seconds. + A value of zero (the default) disables the delay. + This parameter can only be set in the <code class="filename">postgresql.conf</code> + file or on the server command line. + </p></dd><dt id="GUC-TRACE-NOTIFY"><span class="term"><code class="varname">trace_notify</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.10.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-NOTIFY" class="id_link">#</a></dt><dd><p> + Generates a great amount of debugging output for the + <code class="command">LISTEN</code> and <code class="command">NOTIFY</code> + commands. <a class="xref" href="runtime-config-client.html#GUC-CLIENT-MIN-MESSAGES">client_min_messages</a> or + <a class="xref" href="runtime-config-logging.html#GUC-LOG-MIN-MESSAGES">log_min_messages</a> must be + <code class="literal">DEBUG1</code> or lower to send this output to the + client or server logs, respectively. + </p></dd><dt id="GUC-TRACE-RECOVERY-MESSAGES"><span class="term"><code class="varname">trace_recovery_messages</code> (<code class="type">enum</code>) + <a id="id-1.6.7.20.3.11.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-RECOVERY-MESSAGES" class="id_link">#</a></dt><dd><p> + Enables logging of recovery-related debugging output that otherwise + would not be logged. This parameter allows the user to override the + normal setting of <a class="xref" href="runtime-config-logging.html#GUC-LOG-MIN-MESSAGES">log_min_messages</a>, but only for + specific messages. This is intended for use in debugging hot standby. + Valid values are <code class="literal">DEBUG5</code>, <code class="literal">DEBUG4</code>, + <code class="literal">DEBUG3</code>, <code class="literal">DEBUG2</code>, <code class="literal">DEBUG1</code>, and + <code class="literal">LOG</code>. The default, <code class="literal">LOG</code>, does not affect + logging decisions at all. The other values cause recovery-related + debug messages of that priority or higher to be logged as though they + had <code class="literal">LOG</code> priority; for common settings of + <code class="varname">log_min_messages</code> this results in unconditionally sending + them to the server log. + This parameter can only be set in the <code class="filename">postgresql.conf</code> + file or on the server command line. + </p></dd><dt id="GUC-TRACE-SORT"><span class="term"><code class="varname">trace_sort</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.12.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-SORT" class="id_link">#</a></dt><dd><p> + If on, emit information about resource usage during sort operations. + This parameter is only available if the <code class="symbol">TRACE_SORT</code> macro + was defined when <span class="productname">PostgreSQL</span> was compiled. + (However, <code class="symbol">TRACE_SORT</code> is currently defined by default.) + </p></dd><dt id="GUC-TRACE-LOCKS"><span class="term"><code class="varname">trace_locks</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.13.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-LOCKS" class="id_link">#</a></dt><dd><p> + If on, emit information about lock usage. Information dumped + includes the type of lock operation, the type of lock and the unique + identifier of the object being locked or unlocked. Also included + are bit masks for the lock types already granted on this object as + well as for the lock types awaited on this object. For each lock + type a count of the number of granted locks and waiting locks is + also dumped as well as the totals. An example of the log file output + is shown here: +</p><pre class="screen"> +LOG: LockAcquire: new: lock(0xb7acd844) id(24688,24696,0,0,0,1) + grantMask(0) req(0,0,0,0,0,0,0)=0 grant(0,0,0,0,0,0,0)=0 + wait(0) type(AccessShareLock) +LOG: GrantLock: lock(0xb7acd844) id(24688,24696,0,0,0,1) + grantMask(2) req(1,0,0,0,0,0,0)=1 grant(1,0,0,0,0,0,0)=1 + wait(0) type(AccessShareLock) +LOG: UnGrantLock: updated: lock(0xb7acd844) id(24688,24696,0,0,0,1) + grantMask(0) req(0,0,0,0,0,0,0)=0 grant(0,0,0,0,0,0,0)=0 + wait(0) type(AccessShareLock) +LOG: CleanUpLock: deleting: lock(0xb7acd844) id(24688,24696,0,0,0,1) + grantMask(0) req(0,0,0,0,0,0,0)=0 grant(0,0,0,0,0,0,0)=0 + wait(0) type(INVALID) +</pre><p> + Details of the structure being dumped may be found in + <code class="filename">src/include/storage/lock.h</code>. + </p><p> + This parameter is only available if the <code class="symbol">LOCK_DEBUG</code> + macro was defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-TRACE-LWLOCKS"><span class="term"><code class="varname">trace_lwlocks</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.14.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-LWLOCKS" class="id_link">#</a></dt><dd><p> + If on, emit information about lightweight lock usage. Lightweight + locks are intended primarily to provide mutual exclusion of access + to shared-memory data structures. + </p><p> + This parameter is only available if the <code class="symbol">LOCK_DEBUG</code> + macro was defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-TRACE-USERLOCKS"><span class="term"><code class="varname">trace_userlocks</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.15.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-USERLOCKS" class="id_link">#</a></dt><dd><p> + If on, emit information about user lock usage. Output is the same + as for <code class="symbol">trace_locks</code>, only for advisory locks. + </p><p> + This parameter is only available if the <code class="symbol">LOCK_DEBUG</code> + macro was defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-TRACE-LOCK-OIDMIN"><span class="term"><code class="varname">trace_lock_oidmin</code> (<code class="type">integer</code>) + <a id="id-1.6.7.20.3.16.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-LOCK-OIDMIN" class="id_link">#</a></dt><dd><p> + If set, do not trace locks for tables below this OID (used to avoid + output on system tables). + </p><p> + This parameter is only available if the <code class="symbol">LOCK_DEBUG</code> + macro was defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-TRACE-LOCK-TABLE"><span class="term"><code class="varname">trace_lock_table</code> (<code class="type">integer</code>) + <a id="id-1.6.7.20.3.17.1.3" class="indexterm"></a> + </span> <a href="#GUC-TRACE-LOCK-TABLE" class="id_link">#</a></dt><dd><p> + Unconditionally trace locks on this table (OID). + </p><p> + This parameter is only available if the <code class="symbol">LOCK_DEBUG</code> + macro was defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-DEBUG-DEADLOCKS"><span class="term"><code class="varname">debug_deadlocks</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.18.1.3" class="indexterm"></a> + </span> <a href="#GUC-DEBUG-DEADLOCKS" class="id_link">#</a></dt><dd><p> + If set, dumps information about all current locks when a + deadlock timeout occurs. + </p><p> + This parameter is only available if the <code class="symbol">LOCK_DEBUG</code> + macro was defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-LOG-BTREE-BUILD-STATS"><span class="term"><code class="varname">log_btree_build_stats</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.19.1.3" class="indexterm"></a> + </span> <a href="#GUC-LOG-BTREE-BUILD-STATS" class="id_link">#</a></dt><dd><p> + If set, logs system resource usage statistics (memory and CPU) on + various B-tree operations. + </p><p> + This parameter is only available if the <code class="symbol">BTREE_BUILD_STATS</code> + macro was defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-WAL-CONSISTENCY-CHECKING"><span class="term"><code class="varname">wal_consistency_checking</code> (<code class="type">string</code>) + <a id="id-1.6.7.20.3.20.1.3" class="indexterm"></a> + </span> <a href="#GUC-WAL-CONSISTENCY-CHECKING" class="id_link">#</a></dt><dd><p> + This parameter is intended to be used to check for bugs in the WAL + redo routines. When enabled, full-page images of any buffers modified + in conjunction with the WAL record are added to the record. + If the record is subsequently replayed, the system will first apply + each record and then test whether the buffers modified by the record + match the stored images. In certain cases (such as hint bits), minor + variations are acceptable, and will be ignored. Any unexpected + differences will result in a fatal error, terminating recovery. + </p><p> + The default value of this setting is the empty string, which disables + the feature. It can be set to <code class="literal">all</code> to check all + records, or to a comma-separated list of resource managers to check + only records originating from those resource managers. Currently, + the supported resource managers are <code class="literal">heap</code>, + <code class="literal">heap2</code>, <code class="literal">btree</code>, <code class="literal">hash</code>, + <code class="literal">gin</code>, <code class="literal">gist</code>, <code class="literal">sequence</code>, + <code class="literal">spgist</code>, <code class="literal">brin</code>, and <code class="literal">generic</code>. + Extensions may define additional resource managers. Only superusers and users with + the appropriate <code class="literal">SET</code> privilege can change this setting. + </p></dd><dt id="GUC-WAL-DEBUG"><span class="term"><code class="varname">wal_debug</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.21.1.3" class="indexterm"></a> + </span> <a href="#GUC-WAL-DEBUG" class="id_link">#</a></dt><dd><p> + If on, emit WAL-related debugging output. This parameter is + only available if the <code class="symbol">WAL_DEBUG</code> macro was + defined when <span class="productname">PostgreSQL</span> was + compiled. + </p></dd><dt id="GUC-IGNORE-CHECKSUM-FAILURE"><span class="term"><code class="varname">ignore_checksum_failure</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.22.1.3" class="indexterm"></a> + </span> <a href="#GUC-IGNORE-CHECKSUM-FAILURE" class="id_link">#</a></dt><dd><p> + Only has effect if <a class="xref" href="app-initdb.html#APP-INITDB-DATA-CHECKSUMS">data checksums</a> are enabled. + </p><p> + Detection of a checksum failure during a read normally causes + <span class="productname">PostgreSQL</span> to report an error, aborting the current + transaction. Setting <code class="varname">ignore_checksum_failure</code> to on causes + the system to ignore the failure (but still report a warning), and + continue processing. This behavior may <span class="emphasis"><em>cause crashes, propagate + or hide corruption, or other serious problems</em></span>. However, it may allow + you to get past the error and retrieve undamaged tuples that might still be + present in the table if the block header is still sane. If the header is + corrupt an error will be reported even if this option is enabled. The + default setting is <code class="literal">off</code>. + Only superusers and users with the appropriate <code class="literal">SET</code> + privilege can change this setting. + </p></dd><dt id="GUC-ZERO-DAMAGED-PAGES"><span class="term"><code class="varname">zero_damaged_pages</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.23.1.3" class="indexterm"></a> + </span> <a href="#GUC-ZERO-DAMAGED-PAGES" class="id_link">#</a></dt><dd><p> + Detection of a damaged page header normally causes + <span class="productname">PostgreSQL</span> to report an error, aborting the current + transaction. Setting <code class="varname">zero_damaged_pages</code> to on causes + the system to instead report a warning, zero out the damaged + page in memory, and continue processing. This behavior <span class="emphasis"><em>will destroy data</em></span>, + namely all the rows on the damaged page. However, it does allow you to get + past the error and retrieve rows from any undamaged pages that might + be present in the table. It is useful for recovering data if + corruption has occurred due to a hardware or software error. You should + generally not set this on until you have given up hope of recovering + data from the damaged pages of a table. Zeroed-out pages are not + forced to disk so it is recommended to recreate the table or + the index before turning this parameter off again. The + default setting is <code class="literal">off</code>. + Only superusers and users with the appropriate <code class="literal">SET</code> + privilege can change this setting. + </p></dd><dt id="GUC-IGNORE-INVALID-PAGES"><span class="term"><code class="varname">ignore_invalid_pages</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.24.1.3" class="indexterm"></a> + </span> <a href="#GUC-IGNORE-INVALID-PAGES" class="id_link">#</a></dt><dd><p> + If set to <code class="literal">off</code> (the default), detection of + WAL records having references to invalid pages during + recovery causes <span class="productname">PostgreSQL</span> to + raise a PANIC-level error, aborting the recovery. Setting + <code class="varname">ignore_invalid_pages</code> to <code class="literal">on</code> + causes the system to ignore invalid page references in WAL records + (but still report a warning), and continue the recovery. + This behavior may <span class="emphasis"><em>cause crashes, data loss, + propagate or hide corruption, or other serious problems</em></span>. + However, it may allow you to get past the PANIC-level error, + to finish the recovery, and to cause the server to start up. + The parameter can only be set at server start. It only has effect + during recovery or in standby mode. + </p></dd><dt id="GUC-JIT-DEBUGGING-SUPPORT"><span class="term"><code class="varname">jit_debugging_support</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.25.1.3" class="indexterm"></a> + </span> <a href="#GUC-JIT-DEBUGGING-SUPPORT" class="id_link">#</a></dt><dd><p> + If LLVM has the required functionality, register generated functions + with <span class="productname">GDB</span>. This makes debugging easier. + The default setting is <code class="literal">off</code>. + This parameter can only be set at server start. + </p></dd><dt id="GUC-JIT-DUMP-BITCODE"><span class="term"><code class="varname">jit_dump_bitcode</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.26.1.3" class="indexterm"></a> + </span> <a href="#GUC-JIT-DUMP-BITCODE" class="id_link">#</a></dt><dd><p> + Writes the generated <span class="productname">LLVM</span> IR out to the + file system, inside <a class="xref" href="runtime-config-file-locations.html#GUC-DATA-DIRECTORY">data_directory</a>. This is only + useful for working on the internals of the JIT implementation. + The default setting is <code class="literal">off</code>. + Only superusers and users with the appropriate <code class="literal">SET</code> + privilege can change this setting. + </p></dd><dt id="GUC-JIT-EXPRESSIONS"><span class="term"><code class="varname">jit_expressions</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.27.1.3" class="indexterm"></a> + </span> <a href="#GUC-JIT-EXPRESSIONS" class="id_link">#</a></dt><dd><p> + Determines whether expressions are JIT compiled, when JIT compilation + is activated (see <a class="xref" href="jit-decision.html" title="32.2. When to JIT?">Section 32.2</a>). The default is + <code class="literal">on</code>. + </p></dd><dt id="GUC-JIT-PROFILING-SUPPORT"><span class="term"><code class="varname">jit_profiling_support</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.28.1.3" class="indexterm"></a> + </span> <a href="#GUC-JIT-PROFILING-SUPPORT" class="id_link">#</a></dt><dd><p> + If LLVM has the required functionality, emit the data needed to allow + <span class="productname">perf</span> to profile functions generated by JIT. + This writes out files to <code class="filename">~/.debug/jit/</code>; the + user is responsible for performing cleanup when desired. + The default setting is <code class="literal">off</code>. + This parameter can only be set at server start. + </p></dd><dt id="GUC-JIT-TUPLE-DEFORMING"><span class="term"><code class="varname">jit_tuple_deforming</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.29.1.3" class="indexterm"></a> + </span> <a href="#GUC-JIT-TUPLE-DEFORMING" class="id_link">#</a></dt><dd><p> + Determines whether tuple deforming is JIT compiled, when JIT + compilation is activated (see <a class="xref" href="jit-decision.html" title="32.2. When to JIT?">Section 32.2</a>). + The default is <code class="literal">on</code>. + </p></dd><dt id="GUC-REMOVE-TEMP-FILES-AFTER-CRASH"><span class="term"><code class="varname">remove_temp_files_after_crash</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.30.1.3" class="indexterm"></a> + </span> <a href="#GUC-REMOVE-TEMP-FILES-AFTER-CRASH" class="id_link">#</a></dt><dd><p> + When set to <code class="literal">on</code>, which is the default, + <span class="productname">PostgreSQL</span> will automatically remove + temporary files after a backend crash. If disabled, the files will be + retained and may be used for debugging, for example. Repeated crashes + may however result in accumulation of useless files. This parameter + can only be set in the <code class="filename">postgresql.conf</code> file or on + the server command line. + </p></dd><dt id="GUC-SEND-ABORT-FOR-CRASH"><span class="term"><code class="varname">send_abort_for_crash</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.31.1.3" class="indexterm"></a> + </span> <a href="#GUC-SEND-ABORT-FOR-CRASH" class="id_link">#</a></dt><dd><p> + By default, after a backend crash the postmaster will stop remaining + child processes by sending them <span class="systemitem">SIGQUIT</span> + signals, which permits them to exit more-or-less gracefully. When + this option is set to <code class="literal">on</code>, + <span class="systemitem">SIGABRT</span> is sent instead. That normally + results in production of a core dump file for each such child + process. + This can be handy for investigating the states of other processes + after a crash. It can also consume lots of disk space in the event + of repeated crashes, so do not enable this on systems you are not + monitoring carefully. + Beware that no support exists for cleaning up the core file(s) + automatically. + This parameter can only be set in + the <code class="filename">postgresql.conf</code> file or on the server + command line. + </p></dd><dt id="GUC-SEND-ABORT-FOR-KILL"><span class="term"><code class="varname">send_abort_for_kill</code> (<code class="type">boolean</code>) + <a id="id-1.6.7.20.3.32.1.3" class="indexterm"></a> + </span> <a href="#GUC-SEND-ABORT-FOR-KILL" class="id_link">#</a></dt><dd><p> + By default, after attempting to stop a child process with + <span class="systemitem">SIGQUIT</span>, the postmaster will wait five + seconds and then send <span class="systemitem">SIGKILL</span> to force + immediate termination. When this option is set + to <code class="literal">on</code>, <span class="systemitem">SIGABRT</span> is sent + instead of <span class="systemitem">SIGKILL</span>. That normally results + in production of a core dump file for each such child process. + This can be handy for investigating the states + of <span class="quote">“<span class="quote">stuck</span>”</span> child processes. It can also consume lots + of disk space in the event of repeated crashes, so do not enable + this on systems you are not monitoring carefully. + Beware that no support exists for cleaning up the core file(s) + automatically. + This parameter can only be set in + the <code class="filename">postgresql.conf</code> file or on the server + command line. + </p></dd><dt id="GUC-DEBUG-LOGICAL-REPLICATION-STREAMING"><span class="term"><code class="varname">debug_logical_replication_streaming</code> (<code class="type">enum</code>) + <a id="id-1.6.7.20.3.33.1.3" class="indexterm"></a> + </span> <a href="#GUC-DEBUG-LOGICAL-REPLICATION-STREAMING" class="id_link">#</a></dt><dd><p> + The allowed values are <code class="literal">buffered</code> and + <code class="literal">immediate</code>. The default is <code class="literal">buffered</code>. + This parameter is intended to be used to test logical decoding and + replication of large transactions. The effect of + <code class="varname">debug_logical_replication_streaming</code> is different for the + publisher and subscriber: + </p><p> + On the publisher side, <code class="varname">debug_logical_replication_streaming</code> + allows streaming or serializing changes immediately in logical decoding. + When set to <code class="literal">immediate</code>, stream each change if the + <a class="link" href="sql-createsubscription.html#SQL-CREATESUBSCRIPTION-WITH-STREAMING"><code class="literal">streaming</code></a> + option of + <a class="link" href="sql-createsubscription.html" title="CREATE SUBSCRIPTION"><code class="command">CREATE SUBSCRIPTION</code></a> + is enabled, otherwise, serialize each change. When set to + <code class="literal">buffered</code>, the decoding will stream or serialize + changes when <code class="varname">logical_decoding_work_mem</code> is reached. + </p><p> + On the subscriber side, if the <code class="literal">streaming</code> option is set to + <code class="literal">parallel</code>, <code class="varname">debug_logical_replication_streaming</code> + can be used to direct the leader apply worker to send changes to the + shared memory queue or to serialize all changes to the file. When set to + <code class="literal">buffered</code>, the leader sends changes to parallel apply + workers via a shared memory queue. When set to + <code class="literal">immediate</code>, the leader serializes all changes to files + and notifies the parallel apply workers to read and apply them at the + end of the transaction. + </p></dd></dl></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="runtime-config-custom.html" title="20.16. Customized Options">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="runtime-config.html" title="Chapter 20. Server Configuration">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="runtime-config-short.html" title="20.18. Short Options">Next</a></td></tr><tr><td width="40%" align="left" valign="top">20.16. Customized Options </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 16.2 Documentation">Home</a></td><td width="40%" align="right" valign="top"> 20.18. Short Options</td></tr></table></div></body></html>
\ No newline at end of file |