Resource group doesn't relocate when storage is disconnected and 0 paths remain active in pacemaker cluster
Issue
When unplugging storage used for cluster resource (Fibre Channel cables from the server) the multipath correctly reports 0 remaining paths however the FS and LVM resources are still reported as Started by pacemaker. It seems that pacemaker doesn't spot the failure.
The pacemaker has FS resource on top of LVM resource which sits on the device from which is being unplugged.
Multipath reports following messages:
Dec 1 11:18:29 node01 multipathd: checker failed path 8:16 in map 36000d31000948c00000000000000011a
Dec 1 11:18:29 node01 multipathd: 36000d31000948c00000000000000011a: remaining active paths: 0
Environment
- Red Hat Enterprise Linux 7
- High Availability cluster (pacemaker)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
