<?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>24.1. Routine Vacuuming</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="maintenance.html" title="Chapter 24. Routine Database Maintenance Tasks" /><link rel="next" href="routine-reindex.html" title="24.2. Routine Reindexing" /></head><body><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">24.1. Routine Vacuuming</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="maintenance.html" title="Chapter 24. Routine Database Maintenance Tasks">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="maintenance.html" title="Chapter 24. Routine Database Maintenance Tasks">Up</a></td><th width="60%" align="center">Chapter 24. Routine Database Maintenance Tasks</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 11.5 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="routine-reindex.html" title="24.2. Routine Reindexing">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="ROUTINE-VACUUMING"><div class="titlepage"><div><div><h2 class="title" style="clear: both">24.1. Routine Vacuuming</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="routine-vacuuming.html#VACUUM-BASICS">24.1.1. Vacuuming Basics</a></span></dt><dt><span class="sect2"><a href="routine-vacuuming.html#VACUUM-FOR-SPACE-RECOVERY">24.1.2. Recovering Disk Space</a></span></dt><dt><span class="sect2"><a href="routine-vacuuming.html#VACUUM-FOR-STATISTICS">24.1.3. Updating Planner Statistics</a></span></dt><dt><span class="sect2"><a href="routine-vacuuming.html#VACUUM-FOR-VISIBILITY-MAP">24.1.4. Updating The Visibility Map</a></span></dt><dt><span class="sect2"><a href="routine-vacuuming.html#VACUUM-FOR-WRAPAROUND">24.1.5. Preventing Transaction ID Wraparound Failures</a></span></dt><dt><span class="sect2"><a href="routine-vacuuming.html#AUTOVACUUM">24.1.6. The Autovacuum Daemon</a></span></dt></dl></div><a id="id-1.6.11.10.2" class="indexterm"></a><p>
<span class="productname">PostgreSQL</span> databases require periodic
maintenance known as <em class="firstterm">vacuuming</em>. For many installations, it
is sufficient to let vacuuming be performed by the <em class="firstterm">autovacuum
daemon</em>, which is described in <a class="xref" href="routine-vacuuming.html#AUTOVACUUM" title="24.1.6. The Autovacuum Daemon">Section 24.1.6</a>. You might
need to adjust the autovacuuming parameters described there to obtain best
results for your situation. Some database administrators will want to
supplement or replace the daemon's activities with manually-managed
<code class="command">VACUUM</code> commands, which typically are executed according to a
schedule by <span class="application">cron</span> or <span class="application">Task
Scheduler</span> scripts. To set up manually-managed vacuuming properly,
it is essential to understand the issues discussed in the next few
subsections. Administrators who rely on autovacuuming may still wish
to skim this material to help them understand and adjust autovacuuming.
</p><div class="sect2" id="VACUUM-BASICS"><div class="titlepage"><div><div><h3 class="title">24.1.1. Vacuuming Basics</h3></div></div></div><p>
<span class="productname">PostgreSQL</span>'s
<a class="xref" href="sql-vacuum.html" title="VACUUM"><span class="refentrytitle">VACUUM</span></a> command has to
process each table on a regular basis for several reasons:
</p><div class="orderedlist"><ol class="orderedlist" type="1"><li class="listitem">To recover or reuse disk space occupied by updated or deleted
rows.</li><li class="listitem">To update data statistics used by the
<span class="productname">PostgreSQL</span> query planner.</li><li class="listitem">To update the visibility map, which speeds
up <a class="link" href="indexes-index-only-scans.html" title="11.9. Index-Only Scans and Covering Indexes">index-only
scans</a>.</li><li class="listitem">To protect against loss of very old data due to
<em class="firstterm">transaction ID wraparound</em> or
<em class="firstterm">multixact ID wraparound</em>.</li></ol></div><p>
Each of these reasons dictates performing <code class="command">VACUUM</code> operations
of varying frequency and scope, as explained in the following subsections.
</p><p>
There are two variants of <code class="command">VACUUM</code>: standard <code class="command">VACUUM</code>
and <code class="command">VACUUM FULL</code>. <code class="command">VACUUM FULL</code> can reclaim more
disk space but runs much more slowly. Also,
the standard form of <code class="command">VACUUM</code> can run in parallel with production
database operations. (Commands such as <code class="command">SELECT</code>,
<code class="command">INSERT</code>, <code class="command">UPDATE</code>, and
<code class="command">DELETE</code> will continue to function normally, though you
will not be able to modify the definition of a table with commands such as
<code class="command">ALTER TABLE</code> while it is being vacuumed.)
<code class="command">VACUUM FULL</code> requires exclusive lock on the table it is
working on, and therefore cannot be done in parallel with other use
of the table. Generally, therefore,
administrators should strive to use standard <code class="command">VACUUM</code> and
avoid <code class="command">VACUUM FULL</code>.
</p><p>
<code class="command">VACUUM</code> creates a substantial amount of I/O
traffic, which can cause poor performance for other active sessions.
There are configuration parameters that can be adjusted to reduce the
performance impact of background vacuuming — see
<a class="xref" href="runtime-config-resource.html#RUNTIME-CONFIG-RESOURCE-VACUUM-COST" title="19.4.4. Cost-based Vacuum Delay">Section 19.4.4</a>.
</p></div><div class="sect2" id="VACUUM-FOR-SPACE-RECOVERY"><div class="titlepage"><div><div><h3 class="title">24.1.2. Recovering Disk Space</h3></div></div></div><a id="id-1.6.11.10.5.2" class="indexterm"></a><p>
In <span class="productname">PostgreSQL</span>, an
<code class="command">UPDATE</code> or <code class="command">DELETE</code> of a row does not
immediately remove the old version of the row.
This approach is necessary to gain the benefits of multiversion
concurrency control (<acronym class="acronym">MVCC</acronym>, see <a class="xref" href="mvcc.html" title="Chapter 13. Concurrency Control">Chapter 13</a>): the row version
must not be deleted while it is still potentially visible to other
transactions. But eventually, an outdated or deleted row version is no
longer of interest to any transaction. The space it occupies must then be
reclaimed for reuse by new rows, to avoid unbounded growth of disk
space requirements. This is done by running <code class="command">VACUUM</code>.
</p><p>
The standard form of <code class="command">VACUUM</code> removes dead row
versions in tables and indexes and marks the space available for
future reuse. However, it will not return the space to the operating
system, except in the special case where one or more pages at the
end of a table become entirely free and an exclusive table lock can be
easily obtained. In contrast, <code class="command">VACUUM FULL</code> actively compacts
tables by writing a complete new version of the table file with no dead
space. This minimizes the size of the table, but can take a long time.
It also requires extra disk space for the new copy of the table, until
the operation completes.
</p><p>
The usual goal of routine vacuuming is to do standard <code class="command">VACUUM</code>s
often enough to avoid needing <code class="command">VACUUM FULL</code>. The
autovacuum daemon attempts to work this way, and in fact will
never issue <code class="command">VACUUM FULL</code>. In this approach, the idea
is not to keep tables at their minimum size, but to maintain steady-state
usage of disk space: each table occupies space equivalent to its
minimum size plus however much space gets used up between vacuumings.
Although <code class="command">VACUUM FULL</code> can be used to shrink a table back
to its minimum size and return the disk space to the operating system,
there is not much point in this if the table will just grow again in the
future. Thus, moderately-frequent standard <code class="command">VACUUM</code> runs are a
better approach than infrequent <code class="command">VACUUM FULL</code> runs for
maintaining heavily-updated tables.
</p><p>
Some administrators prefer to schedule vacuuming themselves, for example
doing all the work at night when load is low.
The difficulty with doing vacuuming according to a fixed schedule
is that if a table has an unexpected spike in update activity, it may
get bloated to the point that <code class="command">VACUUM FULL</code> is really necessary
to reclaim space. Using the autovacuum daemon alleviates this problem,
since the daemon schedules vacuuming dynamically in response to update
activity. It is unwise to disable the daemon completely unless you
have an extremely predictable workload. One possible compromise is
to set the daemon's parameters so that it will only react to unusually
heavy update activity, thus keeping things from getting out of hand,
while scheduled <code class="command">VACUUM</code>s are expected to do the bulk of the
work when the load is typical.
</p><p>
For those not using autovacuum, a typical approach is to schedule a
database-wide <code class="command">VACUUM</code> once a day during a low-usage period,
supplemented by more frequent vacuuming of heavily-updated tables as
necessary. (Some installations with extremely high update rates vacuum
their busiest tables as often as once every few minutes.) If you have
multiple databases in a cluster, don't forget to
<code class="command">VACUUM</code> each one; the program <a class="xref" href="app-vacuumdb.html" title="vacuumdb"><span class="refentrytitle"><span class="application">vacuumdb</span></span></a> might be helpful.
</p><div class="tip"><h3 class="title">Tip</h3><p>
Plain <code class="command">VACUUM</code> may not be satisfactory when
a table contains large numbers of dead row versions as a result of
massive update or delete activity. If you have such a table and
you need to reclaim the excess disk space it occupies, you will need
to use <code class="command">VACUUM FULL</code>, or alternatively
<a class="xref" href="sql-cluster.html" title="CLUSTER"><span class="refentrytitle">CLUSTER</span></a>
or one of the table-rewriting variants of
<a class="xref" href="sql-altertable.html" title="ALTER TABLE"><span class="refentrytitle">ALTER TABLE</span></a>.
These commands rewrite an entire new copy of the table and build
new indexes for it. All these options require exclusive lock. Note that
they also temporarily use extra disk space approximately equal to the size
of the table, since the old copies of the table and indexes can't be
released until the new ones are complete.
</p></div><div class="tip"><h3 class="title">Tip</h3><p>
If you have a table whose entire contents are deleted on a periodic
basis, consider doing it with
<a class="xref" href="sql-truncate.html" title="TRUNCATE"><span class="refentrytitle">TRUNCATE</span></a> rather
than using <code class="command">DELETE</code> followed by
<code class="command">VACUUM</code>. <code class="command">TRUNCATE</code> removes the
entire content of the table immediately, without requiring a
subsequent <code class="command">VACUUM</code> or <code class="command">VACUUM
FULL</code> to reclaim the now-unused disk space.
The disadvantage is that strict MVCC semantics are violated.
</p></div></div><div class="sect2" id="VACUUM-FOR-STATISTICS"><div class="titlepage"><div><div><h3 class="title">24.1.3. Updating Planner Statistics</h3></div></div></div><a id="id-1.6.11.10.6.2" class="indexterm"></a><a id="id-1.6.11.10.6.3" class="indexterm"></a><p>
The <span class="productname">PostgreSQL</span> query planner relies on
statistical information about the contents of tables in order to
generate good plans for queries. These statistics are gathered by
the <a class="xref" href="sql-analyze.html" title="ANALYZE"><span class="refentrytitle">ANALYZE</span></a> command,
which can be invoked by itself or
as an optional step in <code class="command">VACUUM</code>. It is important to have
reasonably accurate statistics, otherwise poor choices of plans might
degrade database performance.
</p><p>
The autovacuum daemon, if enabled, will automatically issue
<code class="command">ANALYZE</code> commands whenever the content of a table has
changed sufficiently. However, administrators might prefer to rely
on manually-scheduled <code class="command">ANALYZE</code> operations, particularly
if it is known that update activity on a table will not affect the
statistics of <span class="quote">“<span class="quote">interesting</span>”</span> columns. The daemon schedules
<code class="command">ANALYZE</code> strictly as a function of the number of rows
inserted or updated; it has no knowledge of whether that will lead
to meaningful statistical changes.
</p><p>
As with vacuuming for space recovery, frequent updates of statistics
are more useful for heavily-updated tables than for seldom-updated
ones. But even for a heavily-updated table, there might be no need for
statistics updates if the statistical distribution of the data is
not changing much. A simple rule of thumb is to think about how much
the minimum and maximum values of the columns in the table change.
For example, a <code class="type">timestamp</code> column that contains the time
of row update will have a constantly-increasing maximum value as
rows are added and updated; such a column will probably need more
frequent statistics updates than, say, a column containing URLs for
pages accessed on a website. The URL column might receive changes just
as often, but the statistical distribution of its values probably
changes relatively slowly.
</p><p>
It is possible to run <code class="command">ANALYZE</code> on specific tables and even
just specific columns of a table, so the flexibility exists to update some
statistics more frequently than others if your application requires it.
In practice, however, it is usually best to just analyze the entire
database, because it is a fast operation. <code class="command">ANALYZE</code> uses a
statistically random sampling of the rows of a table rather than reading
every single row.
</p><div class="tip"><h3 class="title">Tip</h3><p>
Although per-column tweaking of <code class="command">ANALYZE</code> frequency might not be
very productive, you might find it worthwhile to do per-column
adjustment of the level of detail of the statistics collected by
<code class="command">ANALYZE</code>. Columns that are heavily used in <code class="literal">WHERE</code>
clauses and have highly irregular data distributions might require a
finer-grain data histogram than other columns. See <code class="command">ALTER TABLE
SET STATISTICS</code>, or change the database-wide default using the <a class="xref" href="runtime-config-query.html#GUC-DEFAULT-STATISTICS-TARGET">default_statistics_target</a> configuration parameter.
</p><p>
Also, by default there is limited information available about
the selectivity of functions. However, if you create an expression
index that uses a function call, useful statistics will be
gathered about the function, which can greatly improve query
plans that use the expression index.
</p></div><div class="tip"><h3 class="title">Tip</h3><p>
The autovacuum daemon does not issue <code class="command">ANALYZE</code> commands for
foreign tables, since it has no means of determining how often that
might be useful. If your queries require statistics on foreign tables
for proper planning, it's a good idea to run manually-managed
<code class="command">ANALYZE</code> commands on those tables on a suitable schedule.
</p></div></div><div class="sect2" id="VACUUM-FOR-VISIBILITY-MAP"><div class="titlepage"><div><div><h3 class="title">24.1.4. Updating The Visibility Map</h3></div></div></div><p>
Vacuum maintains a <a class="link" href="storage-vm.html" title="68.4. Visibility Map">visibility map</a> for each
table to keep track of which pages contain only tuples that are known to be
visible to all active transactions (and all future transactions, until the
page is again modified). This has two purposes. First, vacuum
itself can skip such pages on the next run, since there is nothing to
clean up.
</p><p>
Second, it allows <span class="productname">PostgreSQL</span> to answer some
queries using only the index, without reference to the underlying table.
Since <span class="productname">PostgreSQL</span> indexes don't contain tuple
visibility information, a normal index scan fetches the heap tuple for each
matching index entry, to check whether it should be seen by the current
transaction.
An <a class="link" href="indexes-index-only-scans.html" title="11.9. Index-Only Scans and Covering Indexes"><em class="firstterm">index-only
scan</em></a>, on the other hand, checks the visibility map first.
If it's known that all tuples on the page are
visible, the heap fetch can be skipped. This is most useful on
large data sets where the visibility map can prevent disk accesses.
The visibility map is vastly smaller than the heap, so it can easily be
cached even when the heap is very large.
</p></div><div class="sect2" id="VACUUM-FOR-WRAPAROUND"><div class="titlepage"><div><div><h3 class="title">24.1.5. Preventing Transaction ID Wraparound Failures</h3></div></div></div><a id="id-1.6.11.10.8.2" class="indexterm"></a><a id="id-1.6.11.10.8.3" class="indexterm"></a><p>
<span class="productname">PostgreSQL</span>'s
<a class="link" href="mvcc-intro.html" title="13.1. Introduction">MVCC</a> transaction semantics
depend on being able to compare transaction ID (<acronym class="acronym">XID</acronym>)
numbers: a row version with an insertion XID greater than the current
transaction's XID is <span class="quote">“<span class="quote">in the future</span>”</span> and should not be visible
to the current transaction. But since transaction IDs have limited size
(32 bits) a cluster that runs for a long time (more
than 4 billion transactions) would suffer <em class="firstterm">transaction ID
wraparound</em>: the XID counter wraps around to zero, and all of a sudden
transactions that were in the past appear to be in the future — which
means their output become invisible. In short, catastrophic data loss.
(Actually the data is still there, but that's cold comfort if you cannot
get at it.) To avoid this, it is necessary to vacuum every table
in every database at least once every two billion transactions.
</p><p>
The reason that periodic vacuuming solves the problem is that
<code class="command">VACUUM</code> will mark rows as <span class="emphasis"><em>frozen</em></span>, indicating that
they were inserted by a transaction that committed sufficiently far in
the past that the effects of the inserting transaction are certain to be
visible to all current and future transactions.
Normal XIDs are
compared using modulo-2<sup>32</sup> arithmetic. This means
that for every normal XID, there are two billion XIDs that are
<span class="quote">“<span class="quote">older</span>”</span> and two billion that are <span class="quote">“<span class="quote">newer</span>”</span>; another
way to say it is that the normal XID space is circular with no
endpoint. Therefore, once a row version has been created with a particular
normal XID, the row version will appear to be <span class="quote">“<span class="quote">in the past</span>”</span> for
the next two billion transactions, no matter which normal XID we are
talking about. If the row version still exists after more than two billion
transactions, it will suddenly appear to be in the future. To
prevent this, <span class="productname">PostgreSQL</span> reserves a special XID,
<code class="literal">FrozenTransactionId</code>, which does not follow the normal XID
comparison rules and is always considered older
than every normal XID.
Frozen row versions are treated as if the inserting XID were
<code class="literal">FrozenTransactionId</code>, so that they will appear to be
<span class="quote">“<span class="quote">in the past</span>”</span> to all normal transactions regardless of wraparound
issues, and so such row versions will be valid until deleted, no matter
how long that is.
</p><div class="note"><h3 class="title">Note</h3><p>
In <span class="productname">PostgreSQL</span> versions before 9.4, freezing was
implemented by actually replacing a row's insertion XID
with <code class="literal">FrozenTransactionId</code>, which was visible in the
row's <code class="structname">xmin</code> system column. Newer versions just set a flag
bit, preserving the row's original <code class="structname">xmin</code> for possible
forensic use. However, rows with <code class="structname">xmin</code> equal
to <code class="literal">FrozenTransactionId</code> (2) may still be found
in databases <span class="application">pg_upgrade</span>'d from pre-9.4 versions.
</p><p>
Also, system catalogs may contain rows with <code class="structname">xmin</code> equal
to <code class="literal">BootstrapTransactionId</code> (1), indicating that they were
inserted during the first phase of <span class="application">initdb</span>.
Like <code class="literal">FrozenTransactionId</code>, this special XID is treated as
older than every normal XID.
</p></div><p>
<a class="xref" href="runtime-config-client.html#GUC-VACUUM-FREEZE-MIN-AGE">vacuum_freeze_min_age</a>
controls how old an XID value has to be before rows bearing that XID will be
frozen. Increasing this setting may avoid unnecessary work if the
rows that would otherwise be frozen will soon be modified again,
but decreasing this setting increases
the number of transactions that can elapse before the table must be
vacuumed again.
</p><p>
<code class="command">VACUUM</code> uses the <a class="link" href="storage-vm.html" title="68.4. Visibility Map">visibility map</a>
to determine which pages of a table must be scanned. Normally, it
will skip pages that don't have any dead row versions even if those pages
might still have row versions with old XID values. Therefore, normal
<code class="command">VACUUM</code>s won't always freeze every old row version in the table.
Periodically, <code class="command">VACUUM</code> will perform an <em class="firstterm">aggressive
vacuum</em>, skipping only those pages which contain neither dead rows nor
any unfrozen XID or MXID values.
<a class="xref" href="runtime-config-client.html#GUC-VACUUM-FREEZE-TABLE-AGE">vacuum_freeze_table_age</a>
controls when <code class="command">VACUUM</code> does that: all-visible but not all-frozen
pages are scanned if the number of transactions that have passed since the
last such scan is greater than <code class="varname">vacuum_freeze_table_age</code> minus
<code class="varname">vacuum_freeze_min_age</code>. Setting
<code class="varname">vacuum_freeze_table_age</code> to 0 forces <code class="command">VACUUM</code> to
use this more aggressive strategy for all scans.
</p><p>
The maximum time that a table can go unvacuumed is two billion
transactions minus the <code class="varname">vacuum_freeze_min_age</code> value at
the time of the last aggressive vacuum. If it were to go
unvacuumed for longer than
that, data loss could result. To ensure that this does not happen,
autovacuum is invoked on any table that might contain unfrozen rows with
XIDs older than the age specified by the configuration parameter <a class="xref" href="runtime-config-autovacuum.html#GUC-AUTOVACUUM-FREEZE-MAX-AGE">autovacuum_freeze_max_age</a>. (This will happen even if
autovacuum is disabled.)
</p><p>
This implies that if a table is not otherwise vacuumed,
autovacuum will be invoked on it approximately once every
<code class="varname">autovacuum_freeze_max_age</code> minus
<code class="varname">vacuum_freeze_min_age</code> transactions.
For tables that are regularly vacuumed for space reclamation purposes,
this is of little importance. However, for static tables
(including tables that receive inserts, but no updates or deletes),
there is no need to vacuum for space reclamation, so it can
be useful to try to maximize the interval between forced autovacuums
on very large static tables. Obviously one can do this either by
increasing <code class="varname">autovacuum_freeze_max_age</code> or decreasing
<code class="varname">vacuum_freeze_min_age</code>.
</p><p>
The effective maximum for <code class="varname">vacuum_freeze_table_age</code> is 0.95 *
<code class="varname">autovacuum_freeze_max_age</code>; a setting higher than that will be
capped to the maximum. A value higher than
<code class="varname">autovacuum_freeze_max_age</code> wouldn't make sense because an
anti-wraparound autovacuum would be triggered at that point anyway, and
the 0.95 multiplier leaves some breathing room to run a manual
<code class="command">VACUUM</code> before that happens. As a rule of thumb,
<code class="command">vacuum_freeze_table_age</code> should be set to a value somewhat
below <code class="varname">autovacuum_freeze_max_age</code>, leaving enough gap so that
a regularly scheduled <code class="command">VACUUM</code> or an autovacuum triggered by
normal delete and update activity is run in that window. Setting it too
close could lead to anti-wraparound autovacuums, even though the table
was recently vacuumed to reclaim space, whereas lower values lead to more
frequent aggressive vacuuming.
</p><p>
The sole disadvantage of increasing <code class="varname">autovacuum_freeze_max_age</code>
(and <code class="varname">vacuum_freeze_table_age</code> along with it) is that
the <code class="filename">pg_xact</code> and <code class="filename">pg_commit_ts</code>
subdirectories of the database cluster will take more space, because it
must store the commit status and (if <code class="varname">track_commit_timestamp</code> is
enabled) timestamp of all transactions back to
the <code class="varname">autovacuum_freeze_max_age</code> horizon. The commit status uses
two bits per transaction, so if
<code class="varname">autovacuum_freeze_max_age</code> is set to its maximum allowed value
of two billion, <code class="filename">pg_xact</code> can be expected to grow to about half
a gigabyte and <code class="filename">pg_commit_ts</code> to about 20GB. If this
is trivial compared to your total database size,
setting <code class="varname">autovacuum_freeze_max_age</code> to its maximum allowed value
is recommended. Otherwise, set it depending on what you are willing to
allow for <code class="filename">pg_xact</code> and <code class="filename">pg_commit_ts</code> storage.
(The default, 200 million transactions, translates to about 50MB
of <code class="filename">pg_xact</code> storage and about 2GB of <code class="filename">pg_commit_ts</code>
storage.)
</p><p>
One disadvantage of decreasing <code class="varname">vacuum_freeze_min_age</code> is that
it might cause <code class="command">VACUUM</code> to do useless work: freezing a row
version is a waste of time if the row is modified
soon thereafter (causing it to acquire a new XID). So the setting should
be large enough that rows are not frozen until they are unlikely to change
any more.
</p><p>
To track the age of the oldest unfrozen XIDs in a database,
<code class="command">VACUUM</code> stores XID
statistics in the system tables <code class="structname">pg_class</code> and
<code class="structname">pg_database</code>. In particular,
the <code class="structfield">relfrozenxid</code> column of a table's
<code class="structname">pg_class</code> row contains the freeze cutoff XID that was used
by the last aggressive <code class="command">VACUUM</code> for that table. All rows
inserted by transactions with XIDs older than this cutoff XID are
guaranteed to have been frozen. Similarly,
the <code class="structfield">datfrozenxid</code> column of a database's
<code class="structname">pg_database</code> row is a lower bound on the unfrozen XIDs
appearing in that database — it is just the minimum of the
per-table <code class="structfield">relfrozenxid</code> values within the database.
A convenient way to
examine this information is to execute queries such as:
</p><pre class="programlisting">
SELECT c.oid::regclass as table_name,
greatest(age(c.relfrozenxid),age(t.relfrozenxid)) as age
FROM pg_class c
LEFT JOIN pg_class t ON c.reltoastrelid = t.oid
WHERE c.relkind IN ('r', 'm');
SELECT datname, age(datfrozenxid) FROM pg_database;
</pre><p>
The <code class="literal">age</code> column measures the number of transactions from the
cutoff XID to the current transaction's XID.
</p><p>
<code class="command">VACUUM</code> normally only scans pages that have been modified
since the last vacuum, but <code class="structfield">relfrozenxid</code> can only be
advanced when every page of the table
that might contain unfrozen XIDs is scanned. This happens when
<code class="structfield">relfrozenxid</code> is more than
<code class="varname">vacuum_freeze_table_age</code> transactions old, when
<code class="command">VACUUM</code>'s <code class="literal">FREEZE</code> option is used, or when all
pages that are not already all-frozen happen to
require vacuuming to remove dead row versions. When <code class="command">VACUUM</code>
scans every page in the table that is not already all-frozen, it should
set <code class="literal">age(relfrozenxid)</code> to a value just a little more than the
<code class="varname">vacuum_freeze_min_age</code> setting
that was used (more by the number of transactions started since the
<code class="command">VACUUM</code> started). If no <code class="structfield">relfrozenxid</code>-advancing
<code class="command">VACUUM</code> is issued on the table until
<code class="varname">autovacuum_freeze_max_age</code> is reached, an autovacuum will soon
be forced for the table.
</p><p>
If for some reason autovacuum fails to clear old XIDs from a table,
the system will begin to emit warning messages like this when the
database's oldest XIDs reach ten million transactions from the wraparound
point:
</p><pre class="programlisting">
WARNING: database "mydb" must be vacuumed within 177009986 transactions
HINT: To avoid a database shutdown, execute a database-wide VACUUM in "mydb".
</pre><p>
(A manual <code class="command">VACUUM</code> should fix the problem, as suggested by the
hint; but note that the <code class="command">VACUUM</code> must be performed by a
superuser, else it will fail to process system catalogs and thus not
be able to advance the database's <code class="structfield">datfrozenxid</code>.)
If these warnings are
ignored, the system will shut down and refuse to start any new
transactions once there are fewer than 1 million transactions left
until wraparound:
</p><pre class="programlisting">
ERROR: database is not accepting commands to avoid wraparound data loss in database "mydb"
HINT: Stop the postmaster and vacuum that database in single-user mode.
</pre><p>
The 1-million-transaction safety margin exists to let the
administrator recover without data loss, by manually executing the
required <code class="command">VACUUM</code> commands. However, since the system will not
execute commands once it has gone into the safety shutdown mode,
the only way to do this is to stop the server and start the server in single-user
mode to execute <code class="command">VACUUM</code>. The shutdown mode is not enforced
in single-user mode. See the <a class="xref" href="app-postgres.html" title="postgres"><span class="refentrytitle"><span class="application">postgres</span></span></a> reference
page for details about using single-user mode.
</p><div class="sect3" id="VACUUM-FOR-MULTIXACT-WRAPAROUND"><div class="titlepage"><div><div><h4 class="title">24.1.5.1. Multixacts and Wraparound</h4></div></div></div><a id="id-1.6.11.10.8.17.2" class="indexterm"></a><a id="id-1.6.11.10.8.17.3" class="indexterm"></a><p>
<em class="firstterm">Multixact IDs</em> are used to support row locking by
multiple transactions. Since there is only limited space in a tuple
header to store lock information, that information is encoded as
a <span class="quote">“<span class="quote">multiple transaction ID</span>”</span>, or multixact ID for short,
whenever there is more than one transaction concurrently locking a
row. Information about which transaction IDs are included in any
particular multixact ID is stored separately in
the <code class="filename">pg_multixact</code> subdirectory, and only the multixact ID
appears in the <code class="structfield">xmax</code> field in the tuple header.
Like transaction IDs, multixact IDs are implemented as a
32-bit counter and corresponding storage, all of which requires
careful aging management, storage cleanup, and wraparound handling.
There is a separate storage area which holds the list of members in
each multixact, which also uses a 32-bit counter and which must also
be managed.
</p><p>
Whenever <code class="command">VACUUM</code> scans any part of a table, it will replace
any multixact ID it encounters which is older than
<a class="xref" href="runtime-config-client.html#GUC-VACUUM-MULTIXACT-FREEZE-MIN-AGE">vacuum_multixact_freeze_min_age</a>
by a different value, which can be the zero value, a single
transaction ID, or a newer multixact ID. For each table,
<code class="structname">pg_class</code>.<code class="structfield">relminmxid</code> stores the oldest
possible multixact ID still appearing in any tuple of that table.
If this value is older than
<a class="xref" href="runtime-config-client.html#GUC-VACUUM-MULTIXACT-FREEZE-TABLE-AGE">vacuum_multixact_freeze_table_age</a>, an aggressive
vacuum is forced. As discussed in the previous section, an aggressive
vacuum means that only those pages which are known to be all-frozen will
be skipped. <code class="function">mxid_age()</code> can be used on
<code class="structname">pg_class</code>.<code class="structfield">relminmxid</code> to find its age.
</p><p>
Aggressive <code class="command">VACUUM</code> scans, regardless of
what causes them, enable advancing the value for that table.
Eventually, as all tables in all databases are scanned and their
oldest multixact values are advanced, on-disk storage for older
multixacts can be removed.
</p><p>
As a safety device, an aggressive vacuum scan will occur for any table
whose multixact-age is greater than
<a class="xref" href="runtime-config-autovacuum.html#GUC-AUTOVACUUM-MULTIXACT-FREEZE-MAX-AGE">autovacuum_multixact_freeze_max_age</a>. Aggressive
vacuum scans will also occur progressively for all tables, starting with
those that have the oldest multixact-age, if the amount of used member
storage space exceeds the amount 50% of the addressable storage space.
Both of these kinds of aggressive scans will occur even if autovacuum is
nominally disabled.
</p></div></div><div class="sect2" id="AUTOVACUUM"><div class="titlepage"><div><div><h3 class="title">24.1.6. The Autovacuum Daemon</h3></div></div></div><a id="id-1.6.11.10.9.2" class="indexterm"></a><p>
<span class="productname">PostgreSQL</span> has an optional but highly
recommended feature called <em class="firstterm">autovacuum</em>,
whose purpose is to automate the execution of
<code class="command">VACUUM</code> and <code class="command">ANALYZE </code> commands.
When enabled, autovacuum checks for
tables that have had a large number of inserted, updated or deleted
tuples. These checks use the statistics collection facility;
therefore, autovacuum cannot be used unless <a class="xref" href="runtime-config-statistics.html#GUC-TRACK-COUNTS">track_counts</a> is set to <code class="literal">true</code>.
In the default configuration, autovacuuming is enabled and the related
configuration parameters are appropriately set.
</p><p>
The <span class="quote">“<span class="quote">autovacuum daemon</span>”</span> actually consists of multiple processes.
There is a persistent daemon process, called the
<em class="firstterm">autovacuum launcher</em>, which is in charge of starting
<em class="firstterm">autovacuum worker</em> processes for all databases. The
launcher will distribute the work across time, attempting to start one
worker within each database every <a class="xref" href="runtime-config-autovacuum.html#GUC-AUTOVACUUM-NAPTIME">autovacuum_naptime</a>
seconds. (Therefore, if the installation has <em class="replaceable"><code>N</code></em> databases,
a new worker will be launched every
<code class="varname">autovacuum_naptime</code>/<em class="replaceable"><code>N</code></em> seconds.)
A maximum of <a class="xref" href="runtime-config-autovacuum.html#GUC-AUTOVACUUM-MAX-WORKERS">autovacuum_max_workers</a> worker processes
are allowed to run at the same time. If there are more than
<code class="varname">autovacuum_max_workers</code> databases to be processed,
the next database will be processed as soon as the first worker finishes.
Each worker process will check each table within its database and
execute <code class="command">VACUUM</code> and/or <code class="command">ANALYZE</code> as needed.
<a class="xref" href="runtime-config-autovacuum.html#GUC-LOG-AUTOVACUUM-MIN-DURATION">log_autovacuum_min_duration</a> can be set to monitor
autovacuum workers' activity.
</p><p>
If several large tables all become eligible for vacuuming in a short
amount of time, all autovacuum workers might become occupied with
vacuuming those tables for a long period. This would result
in other tables and databases not being vacuumed until a worker becomes
available. There is no limit on how many workers might be in a
single database, but workers do try to avoid repeating work that has
already been done by other workers. Note that the number of running
workers does not count towards <a class="xref" href="runtime-config-connection.html#GUC-MAX-CONNECTIONS">max_connections</a> or
<a class="xref" href="runtime-config-connection.html#GUC-SUPERUSER-RESERVED-CONNECTIONS">superuser_reserved_connections</a> limits.
</p><p>
Tables whose <code class="structfield">relfrozenxid</code> value is more than
<a class="xref" href="runtime-config-autovacuum.html#GUC-AUTOVACUUM-FREEZE-MAX-AGE">autovacuum_freeze_max_age</a> transactions old are always
vacuumed (this also applies to those tables whose freeze max age has
been modified via storage parameters; see below). Otherwise, if the
number of tuples obsoleted since the last
<code class="command">VACUUM</code> exceeds the <span class="quote">“<span class="quote">vacuum threshold</span>”</span>, the
table is vacuumed. The vacuum threshold is defined as:
</p><pre class="programlisting">
vacuum threshold = vacuum base threshold + vacuum scale factor * number of tuples
</pre><p>
where the vacuum base threshold is
<a class="xref" href="runtime-config-autovacuum.html#GUC-AUTOVACUUM-VACUUM-THRESHOLD">autovacuum_vacuum_threshold</a>,
the vacuum scale factor is
<a class="xref" href="runtime-config-autovacuum.html#GUC-AUTOVACUUM-VACUUM-SCALE-FACTOR">autovacuum_vacuum_scale_factor</a>,
and the number of tuples is
<code class="structname">pg_class</code>.<code class="structfield">reltuples</code>.
The number of obsolete tuples is obtained from the statistics
collector; it is a semi-accurate count updated by each
<code class="command">UPDATE</code> and <code class="command">DELETE</code> operation. (It
is only semi-accurate because some information might be lost under heavy
load.) If the <code class="structfield">relfrozenxid</code> value of the table is more
than <code class="varname">vacuum_freeze_table_age</code> transactions old, an aggressive
vacuum is performed to freeze old tuples and advance
<code class="structfield">relfrozenxid</code>; otherwise, only pages that have been modified
since the last vacuum are scanned.
</p><p>
For analyze, a similar condition is used: the threshold, defined as:
</p><pre class="programlisting">
analyze threshold = analyze base threshold + analyze scale factor * number of tuples
</pre><p>
is compared to the total number of tuples inserted, updated, or deleted
since the last <code class="command">ANALYZE</code>.
</p><p>
Temporary tables cannot be accessed by autovacuum. Therefore,
appropriate vacuum and analyze operations should be performed via
session SQL commands.
</p><p>
The default thresholds and scale factors are taken from
<code class="filename">postgresql.conf</code>, but it is possible to override them
(and many other autovacuum control parameters) on a per-table basis; see
<a class="xref" href="sql-createtable.html#SQL-CREATETABLE-STORAGE-PARAMETERS" title="Storage Parameters">Storage Parameters</a> for more information.
If a setting has been changed via a table's storage parameters, that value
is used when processing that table; otherwise the global settings are
used. See <a class="xref" href="runtime-config-autovacuum.html" title="19.10. Automatic Vacuuming">Section 19.10</a> for more details on
the global settings.
</p><p>
When multiple workers are running, the autovacuum cost delay parameters
(see <a class="xref" href="runtime-config-resource.html#RUNTIME-CONFIG-RESOURCE-VACUUM-COST" title="19.4.4. Cost-based Vacuum Delay">Section 19.4.4</a>) are
<span class="quote">“<span class="quote">balanced</span>”</span> among all the running workers, so that the
total I/O impact on the system is the same regardless of the number
of workers actually running. However, any workers processing tables whose
per-table <code class="literal">autovacuum_vacuum_cost_delay</code> or
<code class="literal">autovacuum_vacuum_cost_limit</code> storage parameters have been set
are not considered in the balancing algorithm.
</p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="maintenance.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="maintenance.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="routine-reindex.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Chapter 24. Routine Database Maintenance Tasks </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> 24.2. Routine Reindexing</td></tr></table></div></body></html>
