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>26.3. Failover</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="warm-standby.html" title="26.2. Log-Shipping Standby Servers" /><link rel="next" href="hot-standby.html" title="26.4. Hot Standby" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">26.3. Failover</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="warm-standby.html" title="26.2. Log-Shipping Standby Servers">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="high-availability.html" title="Chapter 26. High Availability, Load Balancing, and Replication">Up</a></td><th width="60%" align="center">Chapter 26. High Availability, Load Balancing, and Replication</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="hot-standby.html" title="26.4. Hot Standby">Next</a></td></tr></table><hr /></div><div class="sect1" id="WARM-STANDBY-FAILOVER"><div class="titlepage"><div><div><h2 class="title" style="clear: both">26.3. Failover <a href="#WARM-STANDBY-FAILOVER" class="id_link">#</a></h2></div></div></div><p>
3 If the primary server fails then the standby server should begin
6 If the standby server fails then no failover need take place. If the
7 standby server can be restarted, even some time later, then the recovery
8 process can also be restarted immediately, taking advantage of
9 restartable recovery. If the standby server cannot be restarted, then a
10 full new standby server instance should be created.
12 If the primary server fails and the standby server becomes the
13 new primary, and then the old primary restarts, you must have
14 a mechanism for informing the old primary that it is no longer the primary. This is
15 sometimes known as <acronym class="acronym">STONITH</acronym> (Shoot The Other Node In The Head), which is
16 necessary to avoid situations where both systems think they are the
17 primary, which will lead to confusion and ultimately data loss.
19 Many failover systems use just two systems, the primary and the standby,
20 connected by some kind of heartbeat mechanism to continually verify the
21 connectivity between the two and the viability of the primary. It is
22 also possible to use a third system (called a witness server) to prevent
23 some cases of inappropriate failover, but the additional complexity
24 might not be worthwhile unless it is set up with sufficient care and
27 <span class="productname">PostgreSQL</span> does not provide the system
28 software required to identify a failure on the primary and notify
29 the standby database server. Many such tools exist and are well
30 integrated with the operating system facilities required for
31 successful failover, such as IP address migration.
33 Once failover to the standby occurs, there is only a
34 single server in operation. This is known as a degenerate state.
35 The former standby is now the primary, but the former primary is down
36 and might stay down. To return to normal operation, a standby server
38 either on the former primary system when it comes up, or on a third,
39 possibly new, system. The <a class="xref" href="app-pgrewind.html" title="pg_rewind"><span class="refentrytitle"><span class="application">pg_rewind</span></span></a> utility can be
40 used to speed up this process on large clusters.
41 Once complete, the primary and standby can be
42 considered to have switched roles. Some people choose to use a third
43 server to provide backup for the new primary until the new standby
45 though clearly this complicates the system configuration and
46 operational processes.
48 So, switching from primary to standby server can be fast but requires
49 some time to re-prepare the failover cluster. Regular switching from
50 primary to standby is useful, since it allows regular downtime on
51 each system for maintenance. This also serves as a test of the
52 failover mechanism to ensure that it will really work when you need it.
53 Written administration procedures are advised.
55 If you have opted for logical replication slot synchronization (see
56 <a class="xref" href="logicaldecoding-explanation.html#LOGICALDECODING-REPLICATION-SLOTS-SYNCHRONIZATION" title="47.2.3. Replication Slot Synchronization">Section 47.2.3</a>),
57 then before switching to the standby server, it is recommended to check
58 if the logical slots synchronized on the standby server are ready
59 for failover. This can be done by following the steps described in
60 <a class="xref" href="logical-replication-failover.html" title="29.3. Logical Replication Failover">Section 29.3</a>.
62 To trigger failover of a log-shipping standby server, run
63 <code class="command">pg_ctl promote</code> or call <code class="function">pg_promote()</code>.
64 If you're setting up reporting servers that are only used to offload
65 read-only queries from the primary, not for high availability purposes,
66 you don't need to promote.
67 </p></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="warm-standby.html" title="26.2. Log-Shipping Standby Servers">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="high-availability.html" title="Chapter 26. High Availability, Load Balancing, and Replication">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="hot-standby.html" title="26.4. Hot Standby">Next</a></td></tr><tr><td width="40%" align="left" valign="top">26.2. Log-Shipping Standby Servers </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"> 26.4. Hot Standby</td></tr></table></div></body></html>