305305 When the node gets synchronized up to the minimum recovery
306306 lag, all the cluster nodes get locked for write transactions to allow the
307307 recovery process to finish. By default, the minimum recovery
308- lag is 10MB . You can change this value in the
308+ lag is 10KB . You can change this value in the
309309 <varname>multimaster.min_recovery_lag</varname> variable.
310310 </para>
311311 </listitem>
363363 <application>pg_basebackup</application> on behalf of this user.
364364 Run the following command on each node you are going to add:
365365 <programlisting>
366- pg_basebackup -D <replaceable>datadir</> -h node1 -U myuser mydb
366+ pg_basebackup -D <replaceable>datadir</> -h node1 -U myuser
367367</programlisting>
368368where <replaceable>datadir</> is the directory containing the database cluster. This directory is specified at the cluster initialization stage, or set in the <envar>PGDATA</envar> environment variable.
369369You can also use any other DBMS user with superuser rights instead of <literal>myuser</literal> to perform this task.
@@ -465,10 +465,11 @@ max_worker_processes = 250
465465multimaster.max_nodes = 3 # cluster size
466466multimaster.node_id = 1 # the 1-based index of this node
467467 # in the cluster
468- multimaster.conn_strings = 'dbname=mydb user=myuser host=node1,dbname=mydb user=myuser host=node2,dbname=mydb user=myuser host=node3'
468+ multimaster.conn_strings = 'dbname=mydb user=myuser host=node1 port=5432 arbiter_port=5433 ,dbname=mydb user=myuser host=node2 port=5432 arbiter_port=5433 ,dbname=mydb user=myuser host=node3 port=5432 arbiter_port=5433 '
469469 # comma-separated list
470470 # of connection strings
471471 # to neighbor nodes
472+ multimaster.arbiter_port = 5433
472473</programlisting>
473474<para>The <literal>multimaster.max_nodes</literal> variable defines the maximum cluster size. If you plan to add new nodes to your cluster, the <literal>multimaster.max_nodes</literal> value should exceed the initial number of nodes. In this case, you can add new nodes without restarting <productname>&productname;</productname> until the specified number of nodes is reached.
474475</para>
@@ -487,24 +488,27 @@ multimaster.conn_strings = 'dbname=mydb user=myuser host=node1,dbname=mydb user=
487488 You can customize connection parameters by adding other
488489 <link linkend="libpq-paramkeywords">libpq connection options</link>
489490 to connection strings in the <varname>multimaster.conn_strings</varname>
490- variable. By default, <filename>multimaster</filename> uses ports 5432
491- and 5433 to establish connections between the nodes and for the arbiter
492- process to listen for connections, respectively . If these ports are
493- already in use, you must specify other ports by adding
491+ variable. By default, <filename>multimaster</filename> uses port 5432
492+ to establish connections between the nodes and port 5433 for the arbiter
493+ process to listen for connections. If these ports are
494+ already in use, you must specify other ports in
494495 <literal>port</literal> and <literal>arbiter_port</literal> options
495- to each connection string in the <varname>multimaster.conn_strings</varname>
496- variable. To check whether the default ports are available, you can run with the following command:
497- <programlisting>
498- netstat -ln | grep -E '5432|5433'
499- </programlisting>
500- Make sure these ports are not blocked by firewall.
496+ in each connection string in the <varname>multimaster.conn_strings</varname>
497+ variable.
501498 </para>
502499 <para>
503500 If you change the <literal>arbiter_port</literal>
504501 option, you must also specify this port in the
505502 <literal>multimaster.arbiter_port</literal> variable. For details,
506503 see <xref linkend="multimaster-arbiter-port"> and
507504 <xref linkend="multimaster-conn-strings">.
505+ </para>
506+ <para>
507+ To check whether the default ports are available, you can run the following command:
508+ <programlisting>
509+ netstat -ln | grep -E '5432|5433'
510+ </programlisting>
511+ Make sure these ports are not blocked by firewall.
508512 </para>
509513 </listitem>
510514 </itemizedlist>
@@ -565,7 +569,7 @@ SELECT mtm.get_cluster_state();
565569 <para>
566570 The <literal>multimaster.heartbeat_send_timeout</literal>
567571 variable defines the time interval between the
568- heartbeats. By default, this variable is set to 1000ms .
572+ heartbeats. By default, this variable is set to 200ms .
569573 </para>
570574 </listitem>
571575 <listitem>
@@ -574,7 +578,7 @@ SELECT mtm.get_cluster_state();
574578 variable sets the timeout for the response. If no heartbeats are
575579 received during this time, the node is assumed to be
576580 disconnected and is excluded from the cluster. By default,
577- this variable is set to 10000ms .
581+ this variable is set to 1000ms .
578582 </para>
579583 </listitem>
580584 </itemizedlist>
@@ -601,7 +605,7 @@ SELECT mtm.get_cluster_state();
601605 <para>
602606 <varname>multimaster.min_recovery_lag</varname> — sets the
603607 minimal WAL lag between the node to be restored and the current cluster state.
604- By default, <varname>multimaster.min_recovery_lag</varname> is set to 10MB .
608+ By default, <varname>multimaster.min_recovery_lag</varname> is set to 10KB .
605609 When the disconnected node is fast-forwarded up to the
606610 <varname>multimaster.min_recovery_lag</varname> threshold,
607611 <filename>multimaster</filename> stops all new commits to the
@@ -1048,20 +1052,20 @@ pg_ctl -D <replaceable>datadir</replaceable> -l <replaceable>pg.log</replaceable
10481052 Time interval
10491053 between heartbeat messages, in milliseconds. An arbiter process
10501054 broadcasts heartbeat messages to all nodes to detect connection
1051- problems. </para><para>Default: 1000
1055+ problems. </para><para>Default: 200
10521056 </para></listitem></varlistentry>
10531057 <varlistentry><term><varname>multimaster.heartbeat_recv_timeout</varname><indexterm><primary><varname>multimaster.heartbeat_recv_timeout</varname></primary></indexterm></term><listitem><para>
10541058 Timeout, in
10551059 milliseconds. If no heartbeat message is received from the node
10561060 within this timeframe, the node is excluded from the cluster.
1057- </para><para>Default: 10000
1061+ </para><para>Default: 1000
10581062 </para></listitem></varlistentry>
10591063 <varlistentry><term><varname>multimaster.min_recovery_lag</varname><indexterm><primary><varname>multimaster.min_recovery_lag</varname></primary></indexterm></term><listitem><para>
10601064 Minimal WAL lag
10611065 between the node to be restored and the current cluster state, in
10621066 kB. When this threshold is reached during node recovery, the
10631067 cluster is locked for write transactions until the recovery is
1064- complete. </para><para>Default: 10MB
1068+ complete. </para><para>Default: 10KB
10651069 </para></listitem></varlistentry>
10661070 <varlistentry><term><varname>multimaster.max_recovery_lag</varname><indexterm><primary><varname>multimaster.max_recovery_lag</varname></primary></indexterm></term><listitem><para>
10671071 Maximal WAL lag
0 commit comments