1 <?xml version="1.0" encoding="UTF-8" standalone="no"?>
2 <!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>28.3. Write-Ahead Logging (WAL)</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="checksums.html" title="28.2. Data Checksums" /><link rel="next" href="wal-async-commit.html" title="28.4. Asynchronous Commit" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">28.3. Write-Ahead Logging (<acronym class="acronym">WAL</acronym>)</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="checksums.html" title="28.2. Data Checksums">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="wal.html" title="Chapter 28. Reliability and the Write-Ahead Log">Up</a></td><th width="60%" align="center">Chapter 28. Reliability and the Write-Ahead Log</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 18.0 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="wal-async-commit.html" title="28.4. Asynchronous Commit">Next</a></td></tr></table><hr /></div><div class="sect1" id="WAL-INTRO"><div class="titlepage"><div><div><h2 class="title" style="clear: both">28.3. Write-Ahead Logging (<acronym class="acronym">WAL</acronym>) <a href="#WAL-INTRO" class="id_link">#</a></h2></div></div></div><a id="id-1.6.15.5.2" class="indexterm"></a><a id="id-1.6.15.5.3" class="indexterm"></a><p>
3 <em class="firstterm">Write-Ahead Logging</em> (<acronym class="acronym">WAL</acronym>)
4 is a standard method for ensuring data integrity. A detailed
5 description can be found in most (if not all) books about
6 transaction processing. Briefly, <acronym class="acronym">WAL</acronym>'s central
7 concept is that changes to data files (where tables and indexes
8 reside) must be written only after those changes have been logged,
9 that is, after WAL records describing the changes have been flushed
10 to permanent storage. If we follow this procedure, we do not need
11 to flush data pages to disk on every transaction commit, because we
12 know that in the event of a crash we will be able to recover the
13 database using the log: any changes that have not been applied to
14 the data pages can be redone from the WAL records. (This is
15 roll-forward recovery, also known as REDO.)
16 </p><div class="tip"><h3 class="title">Tip</h3><p>
17 Because <acronym class="acronym">WAL</acronym> restores database file
18 contents after a crash, journaled file systems are not necessary for
19 reliable storage of the data files or WAL files. In fact, journaling
20 overhead can reduce performance, especially if journaling
21 causes file system <span class="emphasis"><em>data</em></span> to be flushed
22 to disk. Fortunately, data flushing during journaling can
23 often be disabled with a file system mount option, e.g.,
24 <code class="literal">data=writeback</code> on a Linux ext3 file system.
25 Journaled file systems do improve boot speed after a crash.
27 Using <acronym class="acronym">WAL</acronym> results in a
28 significantly reduced number of disk writes, because only the WAL
29 file needs to be flushed to disk to guarantee that a transaction is
30 committed, rather than every data file changed by the transaction.
31 The WAL file is written sequentially,
32 and so the cost of syncing the WAL is much less than the cost of
33 flushing the data pages. This is especially true for servers
34 handling many small transactions touching different parts of the data
35 store. Furthermore, when the server is processing many small concurrent
36 transactions, one <code class="function">fsync</code> of the WAL file may
37 suffice to commit many transactions.
39 <acronym class="acronym">WAL</acronym> also makes it possible to support on-line
40 backup and point-in-time recovery, as described in <a class="xref" href="continuous-archiving.html" title="25.3. Continuous Archiving and Point-in-Time Recovery (PITR)">Section 25.3</a>. By archiving the WAL data we can support
41 reverting to any time instant covered by the available WAL data:
42 we simply install a prior physical backup of the database, and
43 replay the WAL just as far as the desired time. What's more,
44 the physical backup doesn't have to be an instantaneous snapshot
45 of the database state — if it is made over some period of time,
46 then replaying the WAL for that period will fix any internal
48 </p></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="checksums.html" title="28.2. Data Checksums">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="wal.html" title="Chapter 28. Reliability and the Write-Ahead Log">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="wal-async-commit.html" title="28.4. Asynchronous Commit">Next</a></td></tr><tr><td width="40%" align="left" valign="top">28.2. Data Checksums </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 18.0 Documentation">Home</a></td><td width="40%" align="right" valign="top"> 28.4. Asynchronous Commit</td></tr></table></div></body></html>