Drive mount cleanup after server crash (?)

[DanGiles]

Every TSM administrator has been through this: tapes are mounted in the drives then the server crashes. When TSM comes back up it no longer knows about the mounts so when backups or migrations try to start up again, all types of drive and tape errors happen, tapes get marked read-only or unavailable, and drives get taken off line. Then you spend an hour or two cleaning up the mess. This gets compounded when the crashed server is a media manager! You can't just start dismounting volumes because they'd probably still be in use by other servers or LAN-FREE clients.

What do others do? Any way of automating the cleanup? (TSM hasn't gained the abililty of cleaning itself up, has it?)

 

[moon-buddy]

Ah, the agony of a crash!

In the past where I managed a big tape backup system, I gather a snapshot of the libray, libvol and other configuration every 15 minutes. I store this on a managemnt server and delete the logs every 3 hours.

I did this because of the pain I have to go through every time the server crashed especially on a Windows platform. Not much of this excitement on an AIX TSM server!

 

[DanGiles]

Well then, M-B, I have to ask why there's less of a problem with AIX? (I am not talking about drive path consistency here, but tapes being left in the drives.)

 

[moon-buddy]

Well then, M-B, I have to ask why there's less of a problem with AIX? (I am not talking about drive path consistency here, but tapes being left in the drives.)

I am talking about crashes - the issue about tapes left in the drives is when the system crashes!