github/repmgr
1
0
Fork 0
mirror of https://github.com/EnterpriseDB/repmgr.git synced 2026-03-24 07:36:30 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity

repmgrd: monitor standbys attached to primary

Browse Source 
This functionality enables repmgrd (when running on the primary) to
monitor connected child nodes. It will log connections and disconnections
and generate events.

Additionally, repmgrd can execute a custom script if the number of connected
child nodes falls below a configurable threshold. This script can be used
e.g. to "fence" the primary following a failover situation where a new primary
has been promoted and all standbys are now child nodes of that primary.
This commit is contained in:
Ian Barwick
2019-04-22 16:16:59 +09:00
parent 64c4cb81d5
commit 5a90513878
9 changed files with 917 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
doc/appendix-release-notes.sgml
View File

@@ -15,6 +15,11 @@
See also: <xref linkend="upgrading-repmgr">
</para>
<sect1 id="release-4.4">
<title>Release 4.4</title>
<para><emphasis>???, 2019</emphasis></para>
</sect1>
<sect1 id="release-4.3.1">
<title>Release 4.3.1</title>
<para><emphasis>???, 2019</emphasis></para>

105
doc/repmgrd-automatic-failover.sgml
View File

@@ -165,6 +165,111 @@
</sect1>
<sect1 id="repmgrd-primary-standby-disconnection" xreflabel="Monitoring standby disconnections on the primary">
<indexterm>
<primary>repmgrd</primary>
<secondary>standby disconnection</secondary>
</indexterm>
<indexterm>
<primary>repmgrd</primary>
<secondary>child node disconnection</secondary>
</indexterm>
<title>Monitoring standby disconnections on the primary node</title>
<note>
<para>
This functionality is available in <link linkend="release-4.4">&repmgr 4.4</link> and later.
</para>
</note>
<para>
When running on the primary node, <application>repmgrd</application> can
monitor connections and in particular disconnections by its attached
child nodes (standbys), and optionally execute a custom command
if certain criteria are met (such as the number of attached nodes falling to
zero following a failover to a new primary); this command can be used for
example to &quot;fence&quot; the node and ensure it is isolated from any
applications attempting to access the replication cluster.
</para>
<para>
<itemizedlist>
<listitem>
<para>
Every few seconds (defined by the configuration parameter <varname>child_nodes_check_interval</varname>
(a value of <literal>0</literal> disables this altogether), <application>repmgrd</application> queries
the <literal>pg_stat_replication</literal> system view and compares
the nodes present there against the list of nodes registered with &repmgr; which
should be attached to the primary.
</para>
</listitem>
<listitem>
<para>
If a child node (standby) is no longer present in <literal>pg_stat_replication</literal>,
<application>repmgrd</application> notes the time it detected the node's absence, and additionally generates a
<literal>child_node_disconnect</literal> event.
</para>
</listitem>
<listitem>
<para>
If a chile node (standby) which was absent from <literal>pg_stat_replication</literal> reappears,
<application>repmgrd</application> clears the time it detected the node's absence, and additionally generates a
<literal>child_node_reconnect</literal> event.
</para>
</listitem>
<listitem>
<para>
If an entirely new child node (standby) is detected, <application>repmgrd</application> adds it to its internal list
and additionally generates a <literal>child_node_new_connect</literal> event.
</para>
</listitem>
<listitem>
<para>
If the <varname>child_nodes_disconnect_command</varname> parameter is set in
<filename>repmgr.conf</filename>, <application>repmgrd</application> will then loop through all child nodes.
If it determines that insufficient child nodes are connected, and a
minimum of <varname>child_nodes_disconnect_timeout</varname> seconds (default: <literal>30</literal>
has elapsed since the last node became disconnected, <application>repmgrd</application> will then execute the
<varname>child_nodes_disconnect_command</varname> script.
</para>
<para>
By default, the <varname>child_nodes_disconnect_command</varname> will only be executed
if all child nodes are disconnected. If <varname>child_nodes_connected_min_count</varname>
is set, the <varname>child_nodes_disconnect_command</varname> script will be triggered
if the number of connected child nodes falls below the specified value (e.g.
if set to <literal>2</literal>, the script will be triggered if only one child node
is connected). Alternatively, if <varname>child_nodes_disconnect_min_count</varname>
and more than that number of child nodes disconnects, the script will be triggered.
</para>
<para>
The <varname>child_nodes_disconnect_command</varname> script will only be executed once
while the criteria for its execution are met. If the criteria for its execution are no longer
met (i.e. some child nodes have reconnected), it will be executed again if
the criteria for its execution are met again.
</para>
<para>
The <varname>child_nodes_disconnect_command</varname> script will not be executed if <application>repmgrd</application> is paused.
</para>
</listitem>
<listitem>
<para>
Note that child nodes which are not attached when <application>repmgrd</application>
starts will <emphasis>not</emphasis> be considered as missing, as <application>repmgrd</application>
cannot know why they are not attached.
</para>
</listitem>
</itemizedlist>
</para>
</sect1>
<sect1 id="repmgrd-standby-disconnection-on-failover" xreflabel="Standby disconnection on failover">
<indexterm>
<primary>repmgrd</primary>