Merge pull request #10699 from g-pan/H17846-MultiThor

HPCC-17846 DOCS:Additional Info for Multiple Thors

Reviewed-By: Jake Smith <jake.smith@lexisnexis.com>
Reviewed-By: Richard Chapman <rchapman@hpccsystems.com>

docs/HPCCSystemAdmin/HPCCSystemAdministratorsGuide.xml

@@ -1665,6 +1665,99 @@ dfsSSLPrivateKeyFile=/keyfilepath/keyfile</programlisting>Set the <emphasis
         the Thors. This needs to be configured per Thor cluster
         definition.</para>
 
+        <para>Multiple Thors on the same cluster require them to share the
+        same build and installation. The environment defines each Thor
+        cluster, which can share the same machine set. There are slave and
+        master port settings that need to be set to avoid clashing. There are
+        also memory sharing/splitting considerations and settings that need to
+        be made. The table below indicates settings in the environment to
+        consider.</para>
+
+        <para><informaltable border="all" colsep="1" rowsep="1">
+            <tgroup cols="2">
+              <colspec colwidth="94.50pt" />
+
+              <tbody>
+                <row>
+                  <entry><emphasis role="bold">Setting</emphasis></entry>
+
+                  <entry><emphasis role="bold">Description</emphasis></entry>
+                </row>
+
+                <row>
+                  <entry><emphasis
+                  role="bold">globalMemorySize</emphasis></entry>
+
+                  <entry>The maximum memory a slave process can use. Typically
+                  85 percent of the memory on the system divided by the total
+                  number of slaves running on the hardware across all
+                  Thors.</entry>
+                </row>
+
+                <row>
+                  <entry><emphasis
+                  role="bold">localThorPortInc</emphasis></entry>
+
+                  <entry>This value is the increment from the base slave
+                  port.</entry>
+                </row>
+
+                <row>
+                  <entry><emphasis
+                  role="bold">masterMemorySize</emphasis></entry>
+
+                  <entry>The maximum memory a Thor master can use. If left
+                  blank it will use the <emphasis>globalMemorySize</emphasis>
+                  value.</entry>
+                </row>
+
+                <row>
+                  <entry><emphasis role="bold">masterport</emphasis></entry>
+
+                  <entry>This value must be unique between Thor instances
+                  running on the same hardware.</entry>
+                </row>
+
+                <row>
+                  <entry><emphasis role="bold">name</emphasis></entry>
+
+                  <entry>The name of each Thor instance must be
+                  unique.</entry>
+                </row>
+
+                <row>
+                  <entry><emphasis role="bold">nodeGroup</emphasis></entry>
+
+                  <entry>This value is associated with files published by this
+                  Thor instance. Normally it is left blank and defaults to the
+                  same as the <emphasis>name</emphasis> attribute. In
+                  environments with multiple Thors sharing the same group of
+                  nodes, the <emphasis>name</emphasis> value of each Thor must
+                  be different. However, the <emphasis>nodeGroup</emphasis>
+                  value of all the Thors sharing the same physical nodes
+                  should be set to the same name. It is very important to make
+                  the <emphasis>nodeGroup</emphasis> value equal to one of the
+                  Thor instance name values.</entry>
+                </row>
+
+                <row>
+                  <entry><emphasis role="bold">slaveport</emphasis></entry>
+
+                  <entry>This value must be unique between Thor instances
+                  running on the same hardware.</entry>
+                </row>
+
+                <row>
+                  <entry><emphasis
+                  role="bold">SlavesPerNode</emphasis></entry>
+
+                  <entry>The number of slaves per node per Thor
+                  instance.</entry>
+                </row>
+              </tbody>
+            </tgroup>
+          </informaltable></para>
+
         <para>You must not place multiple Thors on hardware which does not
         have enough CPU cores to support it. You should not have more Thors
         than number of cores. One good rule is to use a formula where the