<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Write Ahead Log</TITLE
><META
NAME="GENERATOR"
CONTENT="Modular DocBook HTML Stylesheet Version 1.79"><LINK
REV="MADE"
HREF="mailto:pgsql-docs@postgresql.org"><LINK
REL="HOME"
TITLE="PostgreSQL 9.6.22 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="Server Configuration"
HREF="runtime-config.html"><LINK
REL="PREVIOUS"
TITLE="Resource Consumption"
HREF="runtime-config-resource.html"><LINK
REL="NEXT"
TITLE="Replication"
HREF="runtime-config-replication.html"><LINK
REL="STYLESHEET"
TYPE="text/css"
HREF="stylesheet.css"><META
HTTP-EQUIV="Content-Type"
CONTENT="text/html; charset=ISO-8859-1"><META
NAME="creation"
CONTENT="2021-05-18T09:16:10"></HEAD
><BODY
CLASS="SECT1"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="4"
ALIGN="center"
VALIGN="bottom"
><A
HREF="index.html"
>PostgreSQL 9.6.22 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="Resource Consumption"
HREF="runtime-config-resource.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="runtime-config.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
>Chapter 19. Server Configuration</TD
><TD
WIDTH="20%"
ALIGN="right"
VALIGN="top"
><A
TITLE="Replication"
HREF="runtime-config-replication.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="RUNTIME-CONFIG-WAL"
>19.5. Write Ahead Log</A
></H1
><P
> For additional information on tuning these settings,
see <A
HREF="wal-configuration.html"
>Section 30.4</A
>.
</P
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-WAL-SETTINGS"
>19.5.1. Settings</A
></H2
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-WAL-LEVEL"
></A
><TT
CLASS="VARNAME"
>wal_level</TT
> (<TT
CLASS="TYPE"
>enum</TT
>)
</DT
><DD
><P
> <TT
CLASS="VARNAME"
>wal_level</TT
> determines how much information is written
to the WAL. The default value is <TT
CLASS="LITERAL"
>minimal</TT
>, which writes
only the information needed to recover from a crash or immediate
shutdown. <TT
CLASS="LITERAL"
>replica</TT
> adds logging required for WAL
archiving as well as information required to run
read-only queries on a standby server. Finally,
<TT
CLASS="LITERAL"
>logical</TT
> adds information necessary to support logical
decoding. Each level includes the information logged at all lower
levels. This parameter can only be set at server start.
</P
><P
> In <TT
CLASS="LITERAL"
>minimal</TT
> level, WAL-logging of some bulk
operations can be safely skipped, which can make those
operations much faster (see <A
HREF="populate.html#POPULATE-PITR"
>Section 14.4.7</A
>).
Operations in which this optimization can be applied include:
<P
></P
><TABLE
BORDER="0"
><TBODY
><TR
><TD
><TT
CLASS="COMMAND"
>CREATE TABLE AS</TT
></TD
></TR
><TR
><TD
><TT
CLASS="COMMAND"
>CREATE INDEX</TT
></TD
></TR
><TR
><TD
><TT
CLASS="COMMAND"
>CLUSTER</TT
></TD
></TR
><TR
><TD
><TT
CLASS="COMMAND"
>COPY</TT
> into tables that were created or truncated in the same
transaction</TD
></TR
></TBODY
></TABLE
><P
></P
>
But minimal WAL does not contain enough information to reconstruct the
data from a base backup and the WAL logs, so <TT
CLASS="LITERAL"
>replica</TT
> or
higher must be used to enable WAL archiving
(<A
HREF="runtime-config-wal.html#GUC-ARCHIVE-MODE"
>archive_mode</A
>) and streaming replication.
</P
><P
> In <TT
CLASS="LITERAL"
>logical</TT
> level, the same information is logged as
with <TT
CLASS="LITERAL"
>replica</TT
>, plus information needed to allow
extracting logical change sets from the WAL. Using a level of
<TT
CLASS="LITERAL"
>logical</TT
> will increase the WAL volume, particularly if many
tables are configured for <TT
CLASS="LITERAL"
>REPLICA IDENTITY FULL</TT
> and
many <TT
CLASS="COMMAND"
>UPDATE</TT
> and <TT
CLASS="COMMAND"
>DELETE</TT
> statements are
executed.
</P
><P
> In releases prior to 9.6, this parameter also allowed the
values <TT
CLASS="LITERAL"
>archive</TT
> and <TT
CLASS="LITERAL"
>hot_standby</TT
>.
These are still accepted but mapped to <TT
CLASS="LITERAL"
>replica</TT
>.
</P
></DD
><DT
><A
NAME="GUC-FSYNC"
></A
><TT
CLASS="VARNAME"
>fsync</TT
> (<TT
CLASS="TYPE"
>boolean</TT
>)
</DT
><DD
><P
> If this parameter is on, the <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> server
will try to make sure that updates are physically written to
disk, by issuing <CODE
CLASS="FUNCTION"
>fsync()</CODE
> system calls or various
equivalent methods (see <A
HREF="runtime-config-wal.html#GUC-WAL-SYNC-METHOD"
>wal_sync_method</A
>).
This ensures that the database cluster can recover to a
consistent state after an operating system or hardware crash.
</P
><P
> While turning off <TT
CLASS="VARNAME"
>fsync</TT
> is often a performance
benefit, this can result in unrecoverable data corruption in
the event of a power failure or system crash. Thus it
is only advisable to turn off <TT
CLASS="VARNAME"
>fsync</TT
> if
you can easily recreate your entire database from external
data.
</P
><P
> Examples of safe circumstances for turning off
<TT
CLASS="VARNAME"
>fsync</TT
> include the initial loading of a new
database cluster from a backup file, using a database cluster
for processing a batch of data after which the database
will be thrown away and recreated,
or for a read-only database clone which
gets recreated frequently and is not used for failover. High
quality hardware alone is not a sufficient justification for
turning off <TT
CLASS="VARNAME"
>fsync</TT
>.
</P
><P
> For reliable recovery when changing <TT
CLASS="VARNAME"
>fsync</TT
>
off to on, it is necessary to force all modified buffers in the
kernel to durable storage. This can be done while the cluster
is shutdown or while fsync is on by running <TT
CLASS="COMMAND"
>initdb
--sync-only</TT
>, running <TT
CLASS="COMMAND"
>sync</TT
>, unmounting the
file system, or rebooting the server.
</P
><P
> In many situations, turning off <A
HREF="runtime-config-wal.html#GUC-SYNCHRONOUS-COMMIT"
>synchronous_commit</A
>
for noncritical transactions can provide much of the potential
performance benefit of turning off <TT
CLASS="VARNAME"
>fsync</TT
>, without
the attendant risks of data corruption.
</P
><P
> <TT
CLASS="VARNAME"
>fsync</TT
> can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
If you turn this parameter off, also consider turning off
<A
HREF="runtime-config-wal.html#GUC-FULL-PAGE-WRITES"
>full_page_writes</A
>.
</P
></DD
><DT
><A
NAME="GUC-SYNCHRONOUS-COMMIT"
></A
><TT
CLASS="VARNAME"
>synchronous_commit</TT
> (<TT
CLASS="TYPE"
>enum</TT
>)
</DT
><DD
><P
> Specifies how much WAL processing must complete before
the database server returns a <SPAN
CLASS="QUOTE"
>"success"</SPAN
>
indication to the client. Valid values are
<TT
CLASS="LITERAL"
>remote_apply</TT
>, <TT
CLASS="LITERAL"
>on</TT
>
(the default), <TT
CLASS="LITERAL"
>remote_write</TT
>,
<TT
CLASS="LITERAL"
>local</TT
>, and <TT
CLASS="LITERAL"
>off</TT
>.
</P
><P
> If <TT
CLASS="VARNAME"
>synchronous_standby_names</TT
> is empty,
the only meaningful settings are <TT
CLASS="LITERAL"
>on</TT
> and
<TT
CLASS="LITERAL"
>off</TT
>; <TT
CLASS="LITERAL"
>remote_apply</TT
>,
<TT
CLASS="LITERAL"
>remote_write</TT
> and <TT
CLASS="LITERAL"
>local</TT
>
all provide the same local synchronization level
as <TT
CLASS="LITERAL"
>on</TT
>. The local behavior of all
non-<TT
CLASS="LITERAL"
>off</TT
> modes is to wait for local flush of WAL
to disk. In <TT
CLASS="LITERAL"
>off</TT
> mode, there is no waiting,
so there can be a delay between when success is reported to the
client and when the transaction is later guaranteed to be safe
against a server crash. (The maximum
delay is three times <A
HREF="runtime-config-wal.html#GUC-WAL-WRITER-DELAY"
>wal_writer_delay</A
>.) Unlike
<A
HREF="runtime-config-wal.html#GUC-FSYNC"
>fsync</A
>, setting this parameter to <TT
CLASS="LITERAL"
>off</TT
>
does not create any risk of database inconsistency: an operating
system or database crash might
result in some recent allegedly-committed transactions being lost, but
the database state will be just the same as if those transactions had
been aborted cleanly. So, turning <TT
CLASS="VARNAME"
>synchronous_commit</TT
> off
can be a useful alternative when performance is more important than
exact certainty about the durability of a transaction. For more
discussion see <A
HREF="wal-async-commit.html"
>Section 30.3</A
>.
</P
><P
> If <A
HREF="runtime-config-replication.html#GUC-SYNCHRONOUS-STANDBY-NAMES"
>synchronous_standby_names</A
> is non-empty,
<TT
CLASS="VARNAME"
>synchronous_commit</TT
> also controls whether
transaction commits will wait for their WAL records to be
processed on the standby server(s).
</P
><P
> When set to <TT
CLASS="LITERAL"
>remote_apply</TT
>, commits will wait
until replies from the current synchronous standby(s) indicate they
have received the commit record of the transaction and applied
it, so that it has become visible to queries on the standby(s),
and also written to durable storage on the standbys. This will
cause much larger commit delays than previous settings since
it waits for WAL replay. When set to <TT
CLASS="LITERAL"
>on</TT
>,
commits wait until replies
from the current synchronous standby(s) indicate they have received
the commit record of the transaction and flushed it to durable storage. This
ensures the transaction will not be lost unless both the primary and
all synchronous standbys suffer corruption of their database storage.
When set to <TT
CLASS="LITERAL"
>remote_write</TT
>, commits will wait until replies
from the current synchronous standby(s) indicate they have
received the commit record of the transaction and written it to
their file systems. This setting ensures data preservation if a standby instance of
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> crashes, but not if the standby
suffers an operating-system-level crash because the data has not
necessarily reached durable storage on the standby.
The setting <TT
CLASS="LITERAL"
>local</TT
> causes commits to wait for
local flush to disk, but not for replication. This is usually not
desirable when synchronous replication is in use, but is provided for
completeness.
</P
><P
> This parameter can be changed at any time; the behavior for any
one transaction is determined by the setting in effect when it
commits. It is therefore possible, and useful, to have some
transactions commit synchronously and others asynchronously.
For example, to make a single multistatement transaction commit
asynchronously when the default is the opposite, issue <TT
CLASS="COMMAND"
>SET
LOCAL synchronous_commit TO OFF</TT
> within the transaction.
</P
><P
> <A
HREF="runtime-config-wal.html#SYNCHRONOUS-COMMIT-MATRIX"
>Table 19-1</A
> summarizes the
capabilities of the <TT
CLASS="VARNAME"
>synchronous_commit</TT
> settings.
</P
><DIV
CLASS="TABLE"
><A
NAME="SYNCHRONOUS-COMMIT-MATRIX"
></A
><P
><B
>Table 19-1. synchronous_commit Modes</B
></P
><TABLE
BORDER="1"
CLASS="CALSTABLE"
><COL
WIDTH="1*"
TITLE="col1"><COL
WIDTH="1*"
TITLE="col2"><COL
WIDTH="1*"
TITLE="col3"><COL
WIDTH="1*"
TITLE="col4"><COL
WIDTH="1*"
TITLE="col5"><THEAD
><TR
><TH
>synchronous_commit setting</TH
><TH
>local durable commit</TH
><TH
>standby durable commit after PG crash</TH
><TH
>standby durable commit after OS crash</TH
><TH
>standby query consistency</TH
></TR
></THEAD
><TBODY
><TR
><TD
>remote_apply</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
>•</TD
></TR
><TR
><TD
>on</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
> </TD
></TR
><TR
><TD
>remote_write</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
> </TD
><TD
ALIGN="CENTER"
> </TD
></TR
><TR
><TD
>local</TD
><TD
ALIGN="CENTER"
>•</TD
><TD
ALIGN="CENTER"
> </TD
><TD
ALIGN="CENTER"
> </TD
><TD
ALIGN="CENTER"
> </TD
></TR
><TR
><TD
>off</TD
><TD
ALIGN="CENTER"
> </TD
><TD
ALIGN="CENTER"
> </TD
><TD
ALIGN="CENTER"
> </TD
><TD
ALIGN="CENTER"
> </TD
></TR
></TBODY
></TABLE
></DIV
></DD
><DT
><A
NAME="GUC-WAL-SYNC-METHOD"
></A
><TT
CLASS="VARNAME"
>wal_sync_method</TT
> (<TT
CLASS="TYPE"
>enum</TT
>)
</DT
><DD
><P
> Method used for forcing WAL updates out to disk.
If <TT
CLASS="VARNAME"
>fsync</TT
> is off then this setting is irrelevant,
since WAL file updates will not be forced out at all.
Possible values are:
</P
><P
></P
><UL
><LI
><P
> <TT
CLASS="LITERAL"
>open_datasync</TT
> (write WAL files with <CODE
CLASS="FUNCTION"
>open()</CODE
> option <TT
CLASS="SYMBOL"
>O_DSYNC</TT
>)
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>fdatasync</TT
> (call <CODE
CLASS="FUNCTION"
>fdatasync()</CODE
> at each commit)
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>fsync</TT
> (call <CODE
CLASS="FUNCTION"
>fsync()</CODE
> at each commit)
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>fsync_writethrough</TT
> (call <CODE
CLASS="FUNCTION"
>fsync()</CODE
> at each commit, forcing write-through of any disk write cache)
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>open_sync</TT
> (write WAL files with <CODE
CLASS="FUNCTION"
>open()</CODE
> option <TT
CLASS="SYMBOL"
>O_SYNC</TT
>)
</P
></LI
></UL
><P
> The <TT
CLASS="LITERAL"
>open_</TT
>* options also use <TT
CLASS="LITERAL"
>O_DIRECT</TT
> if available.
Not all of these choices are available on all platforms.
The default is the first method in the above list that is supported
by the platform, except that <TT
CLASS="LITERAL"
>fdatasync</TT
> is the default on
Linux and FreeBSD. The default is not necessarily ideal; it might be
necessary to change this setting or other aspects of your system
configuration in order to create a crash-safe configuration or
achieve optimal performance.
These aspects are discussed in <A
HREF="wal-reliability.html"
>Section 30.1</A
>.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-FULL-PAGE-WRITES"
></A
><TT
CLASS="VARNAME"
>full_page_writes</TT
> (<TT
CLASS="TYPE"
>boolean</TT
>)
</DT
><DD
><P
> When this parameter is on, the <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> server
writes the entire content of each disk page to WAL during the
first modification of that page after a checkpoint.
This is needed because
a page write that is in process during an operating system crash might
be only partially completed, leading to an on-disk page
that contains a mix of old and new data. The row-level change data
normally stored in WAL will not be enough to completely restore
such a page during post-crash recovery. Storing the full page image
guarantees that the page can be correctly restored, but at the price
of increasing the amount of data that must be written to WAL.
(Because WAL replay always starts from a checkpoint, it is sufficient
to do this during the first change of each page after a checkpoint.
Therefore, one way to reduce the cost of full-page writes is to
increase the checkpoint interval parameters.)
</P
><P
> Turning this parameter off speeds normal operation, but
might lead to either unrecoverable data corruption, or silent
data corruption, after a system failure. The risks are similar to turning off
<TT
CLASS="VARNAME"
>fsync</TT
>, though smaller, and it should be turned off
only based on the same circumstances recommended for that parameter.
</P
><P
> Turning off this parameter does not affect use of
WAL archiving for point-in-time recovery (PITR)
(see <A
HREF="continuous-archiving.html"
>Section 25.3</A
>).
</P
><P
> This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
The default is <TT
CLASS="LITERAL"
>on</TT
>.
</P
></DD
><DT
><A
NAME="GUC-WAL-LOG-HINTS"
></A
><TT
CLASS="VARNAME"
>wal_log_hints</TT
> (<TT
CLASS="TYPE"
>boolean</TT
>)
</DT
><DD
><P
> When this parameter is <TT
CLASS="LITERAL"
>on</TT
>, the <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
server writes the entire content of each disk page to WAL during the
first modification of that page after a checkpoint, even for
non-critical modifications of so-called hint bits.
</P
><P
> If data checksums are enabled, hint bit updates are always WAL-logged
and this setting is ignored. You can use this setting to test how much
extra WAL-logging would occur if your database had data checksums
enabled.
</P
><P
> This parameter can only be set at server start. The default value is <TT
CLASS="LITERAL"
>off</TT
>.
</P
></DD
><DT
><A
NAME="GUC-WAL-COMPRESSION"
></A
><TT
CLASS="VARNAME"
>wal_compression</TT
> (<TT
CLASS="TYPE"
>boolean</TT
>)
</DT
><DD
><P
> When this parameter is <TT
CLASS="LITERAL"
>on</TT
>, the <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
server compresses a full page image written to WAL when
<A
HREF="runtime-config-wal.html#GUC-FULL-PAGE-WRITES"
>full_page_writes</A
> is on or during a base backup.
A compressed page image will be decompressed during WAL replay.
The default value is <TT
CLASS="LITERAL"
>off</TT
>.
Only superusers can change this setting.
</P
><P
> Turning this parameter on can reduce the WAL volume without
increasing the risk of unrecoverable data corruption,
but at the cost of some extra CPU spent on the compression during
WAL logging and on the decompression during WAL replay.
</P
></DD
><DT
><A
NAME="GUC-WAL-BUFFERS"
></A
><TT
CLASS="VARNAME"
>wal_buffers</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> The amount of shared memory used for WAL data that has not yet been
written to disk. The default setting of -1 selects a size equal to
1/32nd (about 3%) of <A
HREF="runtime-config-resource.html#GUC-SHARED-BUFFERS"
>shared_buffers</A
>, but not less
than <TT
CLASS="LITERAL"
>64kB</TT
> nor more than the size of one WAL
segment, typically <TT
CLASS="LITERAL"
>16MB</TT
>. This value can be set
manually if the automatic choice is too large or too small,
but any positive value less than <TT
CLASS="LITERAL"
>32kB</TT
> will be
treated as <TT
CLASS="LITERAL"
>32kB</TT
>.
This parameter can only be set at server start.
</P
><P
> The contents of the WAL buffers are written out to disk at every
transaction commit, so extremely large values are unlikely to
provide a significant benefit. However, setting this value to at
least a few megabytes can improve write performance on a busy
server where many clients are committing at once. The auto-tuning
selected by the default setting of -1 should give reasonable
results in most cases.
</P
></DD
><DT
><A
NAME="GUC-WAL-WRITER-DELAY"
></A
><TT
CLASS="VARNAME"
>wal_writer_delay</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies how often the WAL writer flushes WAL. After flushing WAL it
sleeps for <TT
CLASS="VARNAME"
>wal_writer_delay</TT
> milliseconds, unless woken up
by an asynchronously committing transaction. If the last flush
happened less than <TT
CLASS="VARNAME"
>wal_writer_delay</TT
> milliseconds ago and
less than <TT
CLASS="VARNAME"
>wal_writer_flush_after</TT
> bytes of WAL have been
produced since, then WAL is only written to the operating system, not
flushed to disk.
The default value is 200 milliseconds (<TT
CLASS="LITERAL"
>200ms</TT
>). Note that
on many systems, the effective resolution of sleep delays is 10
milliseconds; setting <TT
CLASS="VARNAME"
>wal_writer_delay</TT
> to a value that is
not a multiple of 10 might have the same results as setting it to the
next higher multiple of 10. This parameter can only be set in the
<TT
CLASS="FILENAME"
>postgresql.conf</TT
> file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-WAL-WRITER-FLUSH-AFTER"
></A
><TT
CLASS="VARNAME"
>wal_writer_flush_after</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies how often the WAL writer flushes WAL. If the last flush
happened less than <TT
CLASS="VARNAME"
>wal_writer_delay</TT
> milliseconds ago and
less than <TT
CLASS="VARNAME"
>wal_writer_flush_after</TT
> bytes of WAL have been
produced since, then WAL is only written to the operating system, not
flushed to disk. If <TT
CLASS="VARNAME"
>wal_writer_flush_after</TT
> is set
to <TT
CLASS="LITERAL"
>0</TT
> then WAL data is flushed immediately. The default is
<TT
CLASS="LITERAL"
>1MB</TT
>. This parameter can only be set in the
<TT
CLASS="FILENAME"
>postgresql.conf</TT
> file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-COMMIT-DELAY"
></A
><TT
CLASS="VARNAME"
>commit_delay</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> <TT
CLASS="VARNAME"
>commit_delay</TT
> adds a time delay, measured in
microseconds, before a WAL flush is initiated. This can improve
group commit throughput by allowing a larger number of transactions
to commit via a single WAL flush, if system load is high enough
that additional transactions become ready to commit within the
given interval. However, it also increases latency by up to
<TT
CLASS="VARNAME"
>commit_delay</TT
> microseconds for each WAL
flush. Because the delay is just wasted if no other transactions
become ready to commit, a delay is only performed if at least
<TT
CLASS="VARNAME"
>commit_siblings</TT
> other transactions are active
when a flush is about to be initiated. Also, no delays are
performed if <TT
CLASS="VARNAME"
>fsync</TT
> is disabled.
The default <TT
CLASS="VARNAME"
>commit_delay</TT
> is zero (no delay).
Only superusers can change this setting.
</P
><P
> In <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> releases prior to 9.3,
<TT
CLASS="VARNAME"
>commit_delay</TT
> behaved differently and was much
less effective: it affected only commits, rather than all WAL flushes,
and waited for the entire configured delay even if the WAL flush
was completed sooner. Beginning in <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 9.3,
the first process that becomes ready to flush waits for the configured
interval, while subsequent processes wait only until the leader
completes the flush operation.
</P
></DD
><DT
><A
NAME="GUC-COMMIT-SIBLINGS"
></A
><TT
CLASS="VARNAME"
>commit_siblings</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Minimum number of concurrent open transactions to require
before performing the <TT
CLASS="VARNAME"
>commit_delay</TT
> delay. A larger
value makes it more probable that at least one other
transaction will become ready to commit during the delay
interval. The default is five transactions.
</P
></DD
></DL
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-WAL-CHECKPOINTS"
>19.5.2. Checkpoints</A
></H2
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-CHECKPOINT-TIMEOUT"
></A
><TT
CLASS="VARNAME"
>checkpoint_timeout</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Maximum time between automatic WAL checkpoints, in seconds.
The valid range is between 30 seconds and one day.
The default is five minutes (<TT
CLASS="LITERAL"
>5min</TT
>).
Increasing this parameter can increase the amount of time needed
for crash recovery.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-CHECKPOINT-COMPLETION-TARGET"
></A
><TT
CLASS="VARNAME"
>checkpoint_completion_target</TT
> (<TT
CLASS="TYPE"
>floating point</TT
>)
</DT
><DD
><P
> Specifies the target of checkpoint completion, as a fraction of
total time between checkpoints. The default is 0.5.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-CHECKPOINT-FLUSH-AFTER"
></A
><TT
CLASS="VARNAME"
>checkpoint_flush_after</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Whenever more than <TT
CLASS="VARNAME"
>checkpoint_flush_after</TT
> bytes
have been written while performing a checkpoint, attempt to force the
OS to issue these writes to the underlying storage. Doing so will
limit the amount of dirty data in the kernel's page cache, reducing
the likelihood of stalls when an fsync is issued at the end of the
checkpoint, or when the OS writes data back in larger batches in the
background. Often that will result in greatly reduced transaction
latency, but there also are some cases, especially with workloads
that are bigger than <A
HREF="runtime-config-resource.html#GUC-SHARED-BUFFERS"
>shared_buffers</A
>, but smaller
than the OS's page cache, where performance might degrade. This
setting may have no effect on some platforms. The valid range is
between <TT
CLASS="LITERAL"
>0</TT
>, which disables forced writeback,
and <TT
CLASS="LITERAL"
>2MB</TT
>. The default is <TT
CLASS="LITERAL"
>256kB</TT
> on
Linux, <TT
CLASS="LITERAL"
>0</TT
> elsewhere. (If <TT
CLASS="SYMBOL"
>BLCKSZ</TT
> is not
8kB, the default and maximum values scale proportionally to it.)
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-CHECKPOINT-WARNING"
></A
><TT
CLASS="VARNAME"
>checkpoint_warning</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Write a message to the server log if checkpoints caused by
the filling of checkpoint segment files happen closer together
than this many seconds (which suggests that
<TT
CLASS="VARNAME"
>max_wal_size</TT
> ought to be raised). The default is
30 seconds (<TT
CLASS="LITERAL"
>30s</TT
>). Zero disables the warning.
No warnings will be generated if <TT
CLASS="VARNAME"
>checkpoint_timeout</TT
>
is less than <TT
CLASS="VARNAME"
>checkpoint_warning</TT
>.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-MAX-WAL-SIZE"
></A
><TT
CLASS="VARNAME"
>max_wal_size</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Maximum size to let the WAL grow during automatic
checkpoints. This is a soft limit; WAL size can exceed
<TT
CLASS="VARNAME"
>max_wal_size</TT
> under special circumstances, like
under heavy load, a failing <TT
CLASS="VARNAME"
>archive_command</TT
>, or a high
<TT
CLASS="VARNAME"
>wal_keep_segments</TT
> setting. The default is 1 GB.
Increasing this parameter can increase the amount of time needed for
crash recovery.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-MIN-WAL-SIZE"
></A
><TT
CLASS="VARNAME"
>min_wal_size</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> As long as WAL disk usage stays below this setting, old WAL files are
always recycled for future use at a checkpoint, rather than removed.
This can be used to ensure that enough WAL space is reserved to
handle spikes in WAL usage, for example when running large batch
jobs. The default is 80 MB.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
></DL
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-WAL-ARCHIVING"
>19.5.3. Archiving</A
></H2
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-ARCHIVE-MODE"
></A
><TT
CLASS="VARNAME"
>archive_mode</TT
> (<TT
CLASS="TYPE"
>enum</TT
>)
</DT
><DD
><P
> When <TT
CLASS="VARNAME"
>archive_mode</TT
> is enabled, completed WAL segments
are sent to archive storage by setting
<A
HREF="runtime-config-wal.html#GUC-ARCHIVE-COMMAND"
>archive_command</A
>. In addition to <TT
CLASS="LITERAL"
>off</TT
>,
to disable, there are two modes: <TT
CLASS="LITERAL"
>on</TT
>, and
<TT
CLASS="LITERAL"
>always</TT
>. During normal operation, there is no
difference between the two modes, but when set to <TT
CLASS="LITERAL"
>always</TT
>
the WAL archiver is enabled also during archive recovery or standby
mode. In <TT
CLASS="LITERAL"
>always</TT
> mode, all files restored from the archive
or streamed with streaming replication will be archived (again). See
<A
HREF="warm-standby.html#CONTINUOUS-ARCHIVING-IN-STANDBY"
>Section 26.2.9</A
> for details.
</P
><P
> <TT
CLASS="VARNAME"
>archive_mode</TT
> and <TT
CLASS="VARNAME"
>archive_command</TT
> are
separate variables so that <TT
CLASS="VARNAME"
>archive_command</TT
> can be
changed without leaving archiving mode.
This parameter can only be set at server start.
<TT
CLASS="VARNAME"
>archive_mode</TT
> cannot be enabled when
<TT
CLASS="VARNAME"
>wal_level</TT
> is set to <TT
CLASS="LITERAL"
>minimal</TT
>.
</P
></DD
><DT
><A
NAME="GUC-ARCHIVE-COMMAND"
></A
><TT
CLASS="VARNAME"
>archive_command</TT
> (<TT
CLASS="TYPE"
>string</TT
>)
</DT
><DD
><P
> The local shell command to execute to archive a completed WAL file
segment. Any <TT
CLASS="LITERAL"
>%p</TT
> in the string is
replaced by the path name of the file to archive, and any
<TT
CLASS="LITERAL"
>%f</TT
> is replaced by only the file name.
(The path name is relative to the working directory of the server,
i.e., the cluster's data directory.)
Use <TT
CLASS="LITERAL"
>%%</TT
> to embed an actual <TT
CLASS="LITERAL"
>%</TT
> character in the
command. It is important for the command to return a zero
exit status only if it succeeds. For more information see
<A
HREF="continuous-archiving.html#BACKUP-ARCHIVING-WAL"
>Section 25.3.1</A
>.
</P
><P
> This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line. It is ignored unless
<TT
CLASS="VARNAME"
>archive_mode</TT
> was enabled at server start.
If <TT
CLASS="VARNAME"
>archive_command</TT
> is an empty string (the default) while
<TT
CLASS="VARNAME"
>archive_mode</TT
> is enabled, WAL archiving is temporarily
disabled, but the server continues to accumulate WAL segment files in
the expectation that a command will soon be provided. Setting
<TT
CLASS="VARNAME"
>archive_command</TT
> to a command that does nothing but
return true, e.g., <TT
CLASS="LITERAL"
>/bin/true</TT
> (<TT
CLASS="LITERAL"
>REM</TT
> on
Windows), effectively disables
archiving, but also breaks the chain of WAL files needed for
archive recovery, so it should only be used in unusual circumstances.
</P
></DD
><DT
><A
NAME="GUC-ARCHIVE-TIMEOUT"
></A
><TT
CLASS="VARNAME"
>archive_timeout</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> The <A
HREF="runtime-config-wal.html#GUC-ARCHIVE-COMMAND"
>archive_command</A
> is only invoked for
completed WAL segments. Hence, if your server generates little WAL
traffic (or has slack periods where it does so), there could be a
long delay between the completion of a transaction and its safe
recording in archive storage. To limit how old unarchived
data can be, you can set <TT
CLASS="VARNAME"
>archive_timeout</TT
> to force the
server to switch to a new WAL segment file periodically. When this
parameter is greater than zero, the server will switch to a new
segment file whenever this many seconds have elapsed since the last
segment file switch, and there has been any database activity,
including a single checkpoint. (Increasing
<TT
CLASS="VARNAME"
>checkpoint_timeout</TT
> will reduce unnecessary
checkpoints on an idle system.)
Note that archived files that are closed early
due to a forced switch are still the same length as completely full
files. Therefore, it is unwise to use a very short
<TT
CLASS="VARNAME"
>archive_timeout</TT
> — it will bloat your archive
storage. <TT
CLASS="VARNAME"
>archive_timeout</TT
> settings of a minute or so are
usually reasonable. You should consider using streaming replication,
instead of archiving, if you want data to be copied off the master
server more quickly than that.
This parameter can only be set in the
<TT
CLASS="FILENAME"
>postgresql.conf</TT
> file or on the server command line.
</P
></DD
></DL
></DIV
></DIV
></DIV
><DIV
CLASS="NAVFOOTER"
><HR
ALIGN="LEFT"
WIDTH="100%"><TABLE
SUMMARY="Footer navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><A
HREF="runtime-config-resource.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="index.html"
ACCESSKEY="H"
>Home</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><A
HREF="runtime-config-replication.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Resource Consumption</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="runtime-config.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>Replication</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
distrib
>
Mageia
>
7
>
armv7hl
>
media
>
core-updates
>
by-pkgid
>
5fea23694c765462b86d6ddf74461eab
>
files
>
677
