you only can put services on your cluster which are designed to function on a cluster. Only services are existing on the other node after a failover should be put on the cluster host, anything else is specific to the physical host (especially the hardware). I can’t imagine a physical disk bay would walk from one host to another on it’s own
This isn’t that easy i guess
Never thought about where to find the cluster aware services, just experience.
From looking at the check description i would say, none of the Isilon checks are cluster ready, because they are all SNMP based.
You should tell the cluster logic to do a service migration to the other node. After that you will see which monitored services are vanished on the first node and are newly discovered on the second node. These services then need to be assigned to the cluster host.
In my environment I have “Nodes” (emc_isilon_nodes) and “Cluster Health” (emc_isilon_clusterhealth) as Clustered Services. Those 2 change state on every node, when something goes bad on any of the single nodes.
If a node goes down, the host status of the node that goes down will be DOWN (obviously).
As for the Cluster Host, I would use “No IP” in the “IP address family” settings, so that the (virtual) Cluster Host will always be shown as UP.
The status of the clustered Service “Nodes” will go to CRIT (showing which node is down) and global “Cluster Health” Service will also go to CRIT, if a node is down.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact an admin if you think this should be re-opened.