Check manual page of heartbeat_crm
General status of the CRM for heartbeat nodes using crm=yes
This check uses the output of the crm_mon command as sent by the Check_MK
linux agent. The check uses general information from that output like:
The last update time of the output, the DC, the number of nodes and the number
of resources in the cluster.
The check reports CRITICAL in the following situations: When the "last update"
time is larger than the threshold (normally 60 seconds), when the number
of nodes or resources change or - optionally - when the expected DC of the cluster changes.
Till version 1.2.5i6 the check nailed down the DC to the node which was DC during
inventory. This has been changed to not nailing down the node. You can change this back
to the old default by changing your configuration.
One service will be created on each node of the cluster. If naildown_dc is
set to True in inventory_heartbeat_crm_rules, then the node which is currently
DC will be required to stay DC in future.