Does pbx run on inactive node of clustered master

[PaulStevens]

Nbu 7.7.3 on 2 node Linux cluster using vcs

We experienced all backups failing with statuses 26 and 47.

I shutdown Nbu on all media servers and cleaned up processes. Then I failed cluster over to secondary node then tried to fail back to primary. Bp.kill-all was hanging and cluster faulted. We got fault cleared and tried again. Same result.

We are up and running on secondary node but need to fix problem with primary.

Question...
does pbx fail over with the cluster? Or should pbx be running on the inactive node? It is not right now and wondering if that’s why it can’t fail over. I can’t test right now because don’t want to disturb the running backup and duplication jobs.

[Mouse]

PBX is an infrastructure service and required by other non-server daemons such as NBU client, or if you have VCS and InfoScale/Storage Foundation it is used by them as well, therefore you don't need to failover the PBX service

[Marianne]

@Mouse is right about PBX.

You will find normal Client services/daemons (PBX, bpcd, vnetd) running on inactive node so that it can be backed up as regular client.

I am curious about your reason for wanting to fail over and why you are using bp.kill_all instead of cluster commands to switch service group.

Maybe good idea to troubleshoot status 26 and 47 separately.

[Nicolai]

Agree, bp.kill_all and the upstart scripts in /etc/init.d is a no-go when running Netbackup in a VCS configuration.