Got the pleasure of a 4am call last week from our support team to say issues being experienced launching an array of published applications in our xenapp 6.5 environment (usual generic “an error occurred making the requested connection” error).
As the vast majority of servers are using PVS 6.1 vDisks assumed PVS was at play here.
So RDP’ed to one of the PVS servers in question(v.slow to get in as was connecting to any of the affected provisioned VM’s) and was greeted with a ton of similar messages to the below in the application event log:
Going to the command prompt and doing a “mvadm status” could see the p01 move server was showing active
Subsequently rdp’ed to the move 01 server above and again was sluggish, key thing for me it was unresponsive coming back with information on a “mvadm status” so took the punt that it was hosed. A restart of the 01 server immediately resolved the issue.
Key takeaway here was the move server was restarted and not the PVS servers which was done in the past. Main issue of course with the PVS servers being rebooted – in our instance we have a 200 target device limit on one node so it would of knocked out a lot of target devices (about 100) which would in turn of created a boot storm subsequent bringing them back up.
Issue with move (currently on version 2.6.2) was it didn’t failover to the secondary node, believe to be hopefully addressed in new release but we’ll see.