veritas: Panic System On Disk Group Loss

Be wary of the following DiskGroup resource attribute "PanicSystemOnDGLoss"

If you enable I/O fencing and set the DiskGroup attribute PanicSystemOnDGLoss to true, you'll get the desired failover behavior. The behavior you're seeing is by design and is intended to favor data integrity over availability. 

The reason for halting the system is to ensure a failover takes place and there is no data corruption due to 2 hosts wanting to write to the shared storage.

I had experienced panic on one of my production servers and found the following example in the messages file:

 2009/07/23 09:37:57 VCS CRITICAL V-16-10001-1073 (cluster2) DiskGroup:mydg:monitorisk Group: mydg is disabled on system: cluster2. System will panic to migrate all service groups to another VCS node in system list

By default the PanicSystemOnDGLoss attribute is set to 1 (true).

The attribute will cause VCS to panic the system on sudden loss of the diskgroup, when imported by VCS. The resource will also need to be marked as "Critical" for the panic to occur. VCS will not panic the system if the resource is not marked critical.

VCS will then perform an evacuation of the resource and related service group to the next surviving node. If the surviving node is unable to online the resource, no further panics are induced. The clean procedure is called and VCS stops trying to online the resource until the fault is cleared.