topic Re: Issued START/QUEUE/MANAGER/NEW a second time by mistake in Operating System - OpenVMS

Issued START/QUEUE/MANAGER/NEW a second time by mistake

Scot Newton — Fri, 10 Aug 2012 01:58:36 GMT

I have a 2 node cluster running OpenVMS 8.4.

I issued the command START/QUEUE/MANAGER/NEW a second time by mistake, now when I issue SHOW QUEUE/MANAGER I get the message "SYS$QUEUE_MANAGER, failing over" on both nodes (full command output below). I imagine I am not the first one to ever make this goof. Anyone know how to get the queue manager corrected? Thanks in advance for any and all assistance.

$ SHOW QUEUE/MANAGER/FULL

Master file: SYS$SYSROOT:[SYSEXE]QMAN$MASTER.DAT

Queue manager SYS$QUEUE_MANAGER, failing over

/ON=(*)

Database location: SYS$COMMON:[SYSEXE]

Re: Issued START/QUEUE/MANAGER/NEW a second time by mistake

Steven Schweda — Fri, 10 Aug 2012 04:01:40 GMT

Re: Issued START/QUEUE/MANAGER/NEW a second time by mistake

Richard Brodie_1 — Fri, 10 Aug 2012 08:13:16 GMT

I think this may be one of those times where the useful messages get sent to OPCOM. Worth having a look see.

Re: Issued START/QUEUE/MANAGER/NEW a second time by mistake

Scot Newton — Fri, 10 Aug 2012 16:22:56 GMT

hi Steven,

Yes, shared SAN for the 2 node cluster.

Note that one node seems to have finally failed over, and the queue manager started again. I rebooted the node that was still failing over, and then both nodes queue managers went into "failing over" state. They have been that way for about 70 minutes and counting, so something is not good.

Re: Issued START/QUEUE/MANAGER/NEW a second time by mistake

Volker Halle — Fri, 10 Aug 2012 17:20:45 GMT

Scot,

did you check for error messages in SYS$MANAGER:OPERATOR.LOG ?

Volker.

Re: Issued START/QUEUE/MANAGER/NEW a second time by mistake

Scot Newton — Sat, 11 Aug 2012 02:08:05 GMT

Essentially had to delete the QMAN$MASTER.DAT file and recreate the queue database as though it were a new system. I tried renaming the file I created accidentally and rebooting, hoping that the original database file would resume and all would be well. Unfortunately, not so easy. Somewhere during the process, I think I also corrupted the original, as I kept returning to he state where both nodes would have batch queues "failing over". Thanks everyone for helping!