1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
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>
|
