The local time of the cluster member machines are used to update the heartbeat timestamp in the database. It’s crucial that the cluster member machines have a synchronized time.
By default the Central instances update the heartbeat timestamp in the database every 10 seconds, as set in the
<HeartbeatInterval value=“10” />
This can be tweaked between values of 5 and 15.
When a Central instance stops to write the heartbeat timestamp to the DB the HeartbeatMaxAllowedDowntime value is used to determine if a failover is done to the next valid Central in the list. (Valid meaning, in this case, that the next Central application is a HA cluster member and is writing its heartbeat timestamp to the database.)
The default maximum downtime is set for 30 seconds. When the current time [UTC – heartbeat timestamp UTC] > [HeartbeatMaxAllowedDowntime] an automatic failover is triggered.
The default maximum downtime can be overwritten in the
/var/mmsuite/mmcentral/preferences.cfg for each Central instance:
<HeartbeatMaxAllowedDowntime value=“30” />
Values between 20 and 120 are allowed, setting the maximum downtime to 20-120 seconds.