ISSUE :
In case one of the devices which was working as ping node was removed from the network, the cluster continue to run in production, but in order to remove an unwanted ping node you need to stop the cluster. This will require to stop production and re-configure ping nodes and start the cluster.
SOLUTION :
There are two options :
1.
As such procedure is often not possible and customer would like to avoid receiving error messages about missing ping node, the quickest workaround is to create new ping node in the network. You can add an IP alias or add vm-kernel in any ESXi hosts.
2.
You just need to add into Round-Robin multipath settings at all ESXi hosts, new path to physical IP next to currently configured virtual IP paths.
Once ready, you can stop the cluster remove the ping node and start the cluster again.
Once cluster is back up and running you can remove the physical IP paths.
Here is the link to webinar:
http://www.open-e.com/service-and-support/library/webinars-and-videos/accessing-iscsi-volumes-without-downtime-when-failover-is-disabled/
