diff options
Diffstat (limited to 'doc/src/sgml/html/amcheck.html')
| -rw-r--r-- | doc/src/sgml/html/amcheck.html | 267 |
1 files changed, 267 insertions, 0 deletions
diff --git a/doc/src/sgml/html/amcheck.html b/doc/src/sgml/html/amcheck.html new file mode 100644 index 0000000..ec1e6b2 --- /dev/null +++ b/doc/src/sgml/html/amcheck.html @@ -0,0 +1,267 @@ +<?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>F.2. amcheck</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 V1.79.1" /><link rel="prev" href="adminpack.html" title="F.1. adminpack" /><link rel="next" href="auth-delay.html" title="F.3. auth_delay" /></head><body id="docContent" class="container-fluid col-10"><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">F.2. amcheck</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="adminpack.html" title="F.1. adminpack">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="contrib.html" title="Appendix F. Additional Supplied Modules">Up</a></td><th width="60%" align="center">Appendix F. Additional Supplied Modules</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 13.4 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="auth-delay.html" title="F.3. auth_delay">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="AMCHECK"><div class="titlepage"><div><div><h2 class="title" style="clear: both">F.2. amcheck</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="amcheck.html#id-1.11.7.11.7">F.2.1. Functions</a></span></dt><dt><span class="sect2"><a href="amcheck.html#id-1.11.7.11.8">F.2.2. Optional <em class="parameter"><code>heapallindexed</code></em> Verification</a></span></dt><dt><span class="sect2"><a href="amcheck.html#id-1.11.7.11.9">F.2.3. Using <code class="filename">amcheck</code> Effectively</a></span></dt><dt><span class="sect2"><a href="amcheck.html#id-1.11.7.11.10">F.2.4. Repairing Corruption</a></span></dt></dl></div><a id="id-1.11.7.11.2" class="indexterm"></a><p> + The <code class="filename">amcheck</code> module provides functions that allow you to + verify the logical consistency of the structure of relations. If the + structure appears to be valid, no error is raised. + </p><p> + The functions verify various <span class="emphasis"><em>invariants</em></span> in the + structure of the representation of particular relations. The + correctness of the access method functions behind index scans and + other important operations relies on these invariants always + holding. For example, certain functions verify, among other things, + that all B-Tree pages have items in <span class="quote">“<span class="quote">logical</span>”</span> order (e.g., + for B-Tree indexes on <code class="type">text</code>, index tuples should be in + collated lexical order). If that particular invariant somehow fails + to hold, we can expect binary searches on the affected page to + incorrectly guide index scans, resulting in wrong answers to SQL + queries. + </p><p> + Verification is performed using the same procedures as those used by + index scans themselves, which may be user-defined operator class + code. For example, B-Tree index verification relies on comparisons + made with one or more B-Tree support function 1 routines. See <a class="xref" href="xindex.html#XINDEX-SUPPORT" title="37.16.3. Index Method Support Routines">Section 37.16.3</a> for details of operator class support + functions. + </p><p> + <code class="filename">amcheck</code> functions may only be used by superusers. + </p><div class="sect2" id="id-1.11.7.11.7"><div class="titlepage"><div><div><h3 class="title">F.2.1. Functions</h3></div></div></div><div class="variablelist"><dl class="variablelist"><dt><span class="term"> + <code class="function">bt_index_check(index regclass, heapallindexed boolean) returns void</code> + <a id="id-1.11.7.11.7.2.1.1.2" class="indexterm"></a> + </span></dt><dd><p> + <code class="function">bt_index_check</code> tests that its target, a + B-Tree index, respects a variety of invariants. Example usage: +</p><pre class="screen"> +test=# SELECT bt_index_check(index => c.oid, heapallindexed => i.indisunique), + c.relname, + c.relpages +FROM pg_index i +JOIN pg_opclass op ON i.indclass[0] = op.oid +JOIN pg_am am ON op.opcmethod = am.oid +JOIN pg_class c ON i.indexrelid = c.oid +JOIN pg_namespace n ON c.relnamespace = n.oid +WHERE am.amname = 'btree' AND n.nspname = 'pg_catalog' +-- Don't check temp tables, which may be from another session: +AND c.relpersistence != 't' +-- Function may throw an error when this is omitted: +AND c.relkind = 'i' AND i.indisready AND i.indisvalid +ORDER BY c.relpages DESC LIMIT 10; + bt_index_check | relname | relpages +----------------+---------------------------------+---------- + | pg_depend_reference_index | 43 + | pg_depend_depender_index | 40 + | pg_proc_proname_args_nsp_index | 31 + | pg_description_o_c_o_index | 21 + | pg_attribute_relid_attnam_index | 14 + | pg_proc_oid_index | 10 + | pg_attribute_relid_attnum_index | 9 + | pg_amproc_fam_proc_index | 5 + | pg_amop_opr_fam_index | 5 + | pg_amop_fam_strat_index | 5 +(10 rows) +</pre><p> + This example shows a session that performs verification of the + 10 largest catalog indexes in the database <span class="quote">“<span class="quote">test</span>”</span>. + Verification of the presence of heap tuples as index tuples is + requested for the subset that are unique indexes. Since no + error is raised, all indexes tested appear to be logically + consistent. Naturally, this query could easily be changed to + call <code class="function">bt_index_check</code> for every index in the + database where verification is supported. + </p><p> + <code class="function">bt_index_check</code> acquires an <code class="literal">AccessShareLock</code> + on the target index and the heap relation it belongs to. This lock mode + is the same lock mode acquired on relations by simple + <code class="literal">SELECT</code> statements. + <code class="function">bt_index_check</code> does not verify invariants + that span child/parent relationships, but will verify the + presence of all heap tuples as index tuples within the index + when <em class="parameter"><code>heapallindexed</code></em> is + <code class="literal">true</code>. When a routine, lightweight test for + corruption is required in a live production environment, using + <code class="function">bt_index_check</code> often provides the best + trade-off between thoroughness of verification and limiting the + impact on application performance and availability. + </p></dd><dt><span class="term"> + <code class="function">bt_index_parent_check(index regclass, heapallindexed boolean, rootdescend boolean) returns void</code> + <a id="id-1.11.7.11.7.2.2.1.2" class="indexterm"></a> + </span></dt><dd><p> + <code class="function">bt_index_parent_check</code> tests that its + target, a B-Tree index, respects a variety of invariants. + Optionally, when the <em class="parameter"><code>heapallindexed</code></em> + argument is <code class="literal">true</code>, the function verifies the + presence of all heap tuples that should be found within the + index. When the optional <em class="parameter"><code>rootdescend</code></em> + argument is <code class="literal">true</code>, verification re-finds + tuples on the leaf level by performing a new search from the + root page for each tuple. The checks that can be performed by + <code class="function">bt_index_parent_check</code> are a superset of the + checks that can be performed by <code class="function">bt_index_check</code>. + <code class="function">bt_index_parent_check</code> can be thought of as + a more thorough variant of <code class="function">bt_index_check</code>: + unlike <code class="function">bt_index_check</code>, + <code class="function">bt_index_parent_check</code> also checks + invariants that span parent/child relationships, including checking + that there are no missing downlinks in the index structure. + <code class="function">bt_index_parent_check</code> follows the general + convention of raising an error if it finds a logical + inconsistency or other problem. + </p><p> + A <code class="literal">ShareLock</code> is required on the target index by + <code class="function">bt_index_parent_check</code> (a + <code class="literal">ShareLock</code> is also acquired on the heap relation). + These locks prevent concurrent data modification from + <code class="command">INSERT</code>, <code class="command">UPDATE</code>, and <code class="command">DELETE</code> + commands. The locks also prevent the underlying relation from + being concurrently processed by <code class="command">VACUUM</code>, as well as + all other utility commands. Note that the function holds locks + only while running, not for the entire transaction. + </p><p> + <code class="function">bt_index_parent_check</code>'s additional + verification is more likely to detect various pathological + cases. These cases may involve an incorrectly implemented + B-Tree operator class used by the index that is checked, or, + hypothetically, undiscovered bugs in the underlying B-Tree index + access method code. Note that + <code class="function">bt_index_parent_check</code> cannot be used when + Hot Standby mode is enabled (i.e., on read-only physical + replicas), unlike <code class="function">bt_index_check</code>. + </p></dd></dl></div><div class="tip"><h3 class="title">Tip</h3><p> + <code class="function">bt_index_check</code> and + <code class="function">bt_index_parent_check</code> both output log + messages about the verification process at + <code class="literal">DEBUG1</code> and <code class="literal">DEBUG2</code> severity + levels. These messages provide detailed information about the + verification process that may be of interest to + <span class="productname">PostgreSQL</span> developers. Advanced users + may also find this information helpful, since it provides + additional context should verification actually detect an + inconsistency. Running: +</p><pre class="programlisting"> +SET client_min_messages = DEBUG1; +</pre><p> + in an interactive <span class="application">psql</span> session before + running a verification query will display messages about the + progress of verification with a manageable level of detail. + </p></div></div><div class="sect2" id="id-1.11.7.11.8"><div class="titlepage"><div><div><h3 class="title">F.2.2. Optional <em class="parameter"><code>heapallindexed</code></em> Verification</h3></div></div></div><p> + When the <em class="parameter"><code>heapallindexed</code></em> argument to + verification functions is <code class="literal">true</code>, an additional + phase of verification is performed against the table associated with + the target index relation. This consists of a <span class="quote">“<span class="quote">dummy</span>”</span> + <code class="command">CREATE INDEX</code> operation, which checks for the + presence of all hypothetical new index tuples against a temporary, + in-memory summarizing structure (this is built when needed during + the basic first phase of verification). The summarizing structure + <span class="quote">“<span class="quote">fingerprints</span>”</span> every tuple found within the target + index. The high level principle behind + <em class="parameter"><code>heapallindexed</code></em> verification is that a new + index that is equivalent to the existing, target index must only + have entries that can be found in the existing structure. + </p><p> + The additional <em class="parameter"><code>heapallindexed</code></em> phase adds + significant overhead: verification will typically take several times + longer. However, there is no change to the relation-level locks + acquired when <em class="parameter"><code>heapallindexed</code></em> verification is + performed. + </p><p> + The summarizing structure is bound in size by + <code class="varname">maintenance_work_mem</code>. In order to ensure that + there is no more than a 2% probability of failure to detect an + inconsistency for each heap tuple that should be represented in the + index, approximately 2 bytes of memory are needed per tuple. As + less memory is made available per tuple, the probability of missing + an inconsistency slowly increases. This approach limits the + overhead of verification significantly, while only slightly reducing + the probability of detecting a problem, especially for installations + where verification is treated as a routine maintenance task. Any + single absent or malformed tuple has a new opportunity to be + detected with each new verification attempt. + </p></div><div class="sect2" id="id-1.11.7.11.9"><div class="titlepage"><div><div><h3 class="title">F.2.3. Using <code class="filename">amcheck</code> Effectively</h3></div></div></div><p> + <code class="filename">amcheck</code> can be effective at detecting various types of + failure modes that <a class="link" href="app-initdb.html#APP-INITDB-DATA-CHECKSUMS"><span class="application">data page + checksums</span></a> will always fail to catch. These include: + + </p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p> + Structural inconsistencies caused by incorrect operator class + implementations. + </p><p> + This includes issues caused by the comparison rules of operating + system collations changing. Comparisons of datums of a collatable + type like <code class="type">text</code> must be immutable (just as all + comparisons used for B-Tree index scans must be immutable), which + implies that operating system collation rules must never change. + Though rare, updates to operating system collation rules can + cause these issues. More commonly, an inconsistency in the + collation order between a master server and a standby server is + implicated, possibly because the <span class="emphasis"><em>major</em></span> operating + system version in use is inconsistent. Such inconsistencies will + generally only arise on standby servers, and so can generally + only be detected on standby servers. + </p><p> + If a problem like this arises, it may not affect each individual + index that is ordered using an affected collation, simply because + <span class="emphasis"><em>indexed</em></span> values might happen to have the same + absolute ordering regardless of the behavioral inconsistency. See + <a class="xref" href="locale.html" title="23.1. Locale Support">Section 23.1</a> and <a class="xref" href="collation.html" title="23.2. Collation Support">Section 23.2</a> for + further details about how <span class="productname">PostgreSQL</span> uses + operating system locales and collations. + </p></li><li class="listitem"><p> + Structural inconsistencies between indexes and the heap relations + that are indexed (when <em class="parameter"><code>heapallindexed</code></em> + verification is performed). + </p><p> + There is no cross-checking of indexes against their heap relation + during normal operation. Symptoms of heap corruption can be subtle. + </p></li><li class="listitem"><p> + Corruption caused by hypothetical undiscovered bugs in the + underlying <span class="productname">PostgreSQL</span> access method + code, sort code, or transaction management code. + </p><p> + Automatic verification of the structural integrity of indexes + plays a role in the general testing of new or proposed + <span class="productname">PostgreSQL</span> features that could plausibly allow a + logical inconsistency to be introduced. Verification of table + structure and associated visibility and transaction status + information plays a similar role. One obvious testing strategy + is to call <code class="filename">amcheck</code> functions continuously + when running the standard regression tests. See <a class="xref" href="regress-run.html" title="32.1. Running the Tests">Section 32.1</a> for details on running the tests. + </p></li><li class="listitem"><p> + File system or storage subsystem faults where checksums happen to + simply not be enabled. + </p><p> + Note that <code class="filename">amcheck</code> examines a page as represented in some + shared memory buffer at the time of verification if there is only a + shared buffer hit when accessing the block. Consequently, + <code class="filename">amcheck</code> does not necessarily examine data read from the + file system at the time of verification. Note that when checksums are + enabled, <code class="filename">amcheck</code> may raise an error due to a checksum + failure when a corrupt block is read into a buffer. + </p></li><li class="listitem"><p> + Corruption caused by faulty RAM, or the broader memory subsystem. + </p><p> + <span class="productname">PostgreSQL</span> does not protect against correctable + memory errors and it is assumed you will operate using RAM that + uses industry standard Error Correcting Codes (ECC) or better + protection. However, ECC memory is typically only immune to + single-bit errors, and should not be assumed to provide + <span class="emphasis"><em>absolute</em></span> protection against failures that + result in memory corruption. + </p><p> + When <em class="parameter"><code>heapallindexed</code></em> verification is + performed, there is generally a greatly increased chance of + detecting single-bit errors, since strict binary equality is + tested, and the indexed attributes within the heap are tested. + </p></li></ul></div><p> + In general, <code class="filename">amcheck</code> can only prove the presence of + corruption; it cannot prove its absence. + </p></div><div class="sect2" id="id-1.11.7.11.10"><div class="titlepage"><div><div><h3 class="title">F.2.4. Repairing Corruption</h3></div></div></div><p> + No error concerning corruption raised by <code class="filename">amcheck</code> should + ever be a false positive. <code class="filename">amcheck</code> raises + errors in the event of conditions that, by definition, should never + happen, and so careful analysis of <code class="filename">amcheck</code> + errors is often required. + </p><p> + There is no general method of repairing problems that + <code class="filename">amcheck</code> detects. An explanation for the root cause of + an invariant violation should be sought. <a class="xref" href="pageinspect.html" title="F.22. pageinspect">pageinspect</a> may play a useful role in diagnosing + corruption that <code class="filename">amcheck</code> detects. A <code class="command">REINDEX</code> + may not be effective in repairing corruption. + </p></div></div><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navfooter"><hr></hr><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="adminpack.html" title="F.1. adminpack">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="contrib.html" title="Appendix F. Additional Supplied Modules">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="auth-delay.html" title="F.3. auth_delay">Next</a></td></tr><tr><td width="40%" align="left" valign="top">F.1. adminpack </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 13.4 Documentation">Home</a></td><td width="40%" align="right" valign="top"> F.3. auth_delay</td></tr></table></div></body></html>
\ No newline at end of file |