Add formulas for calculating diskless SBD timeouts (#370)

* Add formulas for calculating diskless SBD timeouts

bsc#1219972
jsc#DOCTEAM-1289

* Add warning about diskless SBD timeout misconfiguration

xml/ha_storage_protection.xml

@@ -360,8 +360,9 @@
        <para>
         This timeout is set in the CIB as a global cluster property. If not set
         explicitly, it defaults to <literal>0</literal>, which is appropriate for
-        using SBD with one to three devices. For use of SBD in diskless mode, see <xref
-        linkend="pro-ha-storage-protect-confdiskless"/> for more details.</para>
+        using SBD with one to three devices. For SBD in diskless mode, this timeout
+        must <emphasis>not</emphasis> be <literal>0</literal>. For details, see
+        <xref linkend="pro-ha-storage-protect-confdiskless"/>.</para>
       </listitem>
      </varlistentry>
     </variablelist>
@@ -1037,7 +1038,9 @@ SBD_WATCHDOG_TIMEOUT=5</screen>
 <screen>&prompt.crm.conf;<command>property stonith-enabled="true"</command><co
         xml:id="co-ha-sbd-stonith-enabled"/>
 &prompt.crm.conf;<command>property stonith-watchdog-timeout=10</command><co
-        xml:id="co-ha-sbd-diskless-watchdog-timeout"/></screen>
+        xml:id="co-ha-sbd-diskless-watchdog-timeout"/>
+&prompt.crm.conf;<command>property stonith-timeout=15</command><co
+        xml:id="co-ha-sbd-diskless-stonith-timeout"/></screen>
     <calloutlist>
      <callout arearefs="co-ha-sbd-stonith-enabled">
       <para>
@@ -1048,12 +1051,29 @@ SBD_WATCHDOG_TIMEOUT=5</screen>
      <callout arearefs="co-ha-sbd-diskless-watchdog-timeout">
       <para>For diskless SBD, this parameter must not equal zero.
        It defines after how long it is assumed that the fencing target has already
-       self-fenced. Therefore its value needs to be &gt;= the value of
-       <varname>SBD_WATCHDOG_TIMEOUT</varname> in <filename>/etc/sysconfig/sbd</filename>.
-       Starting with &productname; 15, if you set <parameter>stonith-watchdog-timeout</parameter>
-       to a negative value, Pacemaker will automatically calculate this timeout
-       and set it to twice the value of <parameter>SBD_WATCHDOG_TIMEOUT</parameter>.
+       self-fenced. Use the following formula to calculate this timeout:
       </para>
+      <screen>stonith-watchdog-timeout &gt;= (SBD_WATCHDOG_TIMEOUT * 2)</screen>
+      <para>
+        If you set <parameter>stonith-watchdog-timeout</parameter>
+        to a negative value, Pacemaker automatically calculates this timeout
+        and sets it to twice the value of <parameter>SBD_WATCHDOG_TIMEOUT</parameter>.
+      </para>
+     </callout>
+     <callout arearefs="co-ha-sbd-diskless-stonith-timeout">
+       <para>
+         This parameter must allow sufficient time for fencing to complete.
+         For diskless SBD, use the following formula to calculate this timeout:
+       </para>
+       <screen>stonith-timeout &gt;= stonith-watchdog-timeout + 20%</screen>
+       <important>
+        <title>Diskless SBD timeouts</title>
+        <para>
+          With diskless SBD, if the <literal>stonith-timeout</literal> value is smaller than the
+          <literal>stonith-watchdog-timeout</literal> value, failed nodes can become stuck
+          in an <literal>UNCLEAN</literal> state and block failover of active resources.
+        </para>
+       </important>
      </callout>
     </calloutlist>
    </step>