DOC: sync doc for scheduler - multimaster support added

Author: Liudmila Mantrova

doc/src/sgml/pgpro-scheduler.sgml

@@ -64,13 +64,6 @@
     </listitem>
   </itemizedlist>
 
-  <note>
-    <para>
-     <filename>pgpro_scheduler</filename> is not supported for clusters
-     configured with <xref linkend="multimaster">.
-    </para>
-  </note>
-
   <sect2 id="pgpro-scheduler-installation">
     <title>Installation and Setup</title>
       <para>
@@ -634,6 +627,33 @@ schedule.cancel_job(<replaceable>job_id</replaceable> <type>bigint</type>);
       </itemizedlist></para>
           <para>To learn more about each function, see <xref linkend="pgpro-scheduler-functions">.</para>
           </sect3>
+          <sect3>
+            <title>Scheduling Jobs on a Multi-Master Cluster</title>
+            <para>
+            Using <filename>pgpro_scheduler</filename>, you can manage
+            scheduled and one-time jobs on a cluster configured
+            with <xref linkend="multimaster">.
+            <filename>pgpro_scheduler</filename> can only manage jobs
+            on the node on which it is installed. Thus, you must
+            install and enable <filename>pgpro_scheduler</filename>
+            on all nodes on which you would like to schedule jobs.
+            <filename>pgpro_scheduler</filename> instances will manage
+            jobs on different nodes independently, but the executed jobs
+            will be replicated to other nodes.
+            </para>
+
+            <para>
+             Even if you are planning to schedule jobs on a single node only,
+             it is recommended to enable <filename>pgpro_scheduler</filename>
+             on several nodes. In this case, if a node with
+             scheduled jobs fails, another <filename>pgpro_scheduler</filename>
+             instance picks up these jobs.
+             If <filename>pgpro_scheduler</filename> is running on more than
+             one node, the node with the smallest node ID is selected.
+             The naming pattern of node IDs is defined by the
+             <xref linkend="pgpro-scheduler-schedule-nodename"> GUC variable.
+            </para>
+          </sect3>
    </sect2>
    <sect2 id="pgpro-scheduler-reference">
    <title>Reference</title>