ADSM-L

[ADSM-L] MOVE DRM & CHECKIN LIBVOL

2010-07-18 18:34:22
Subject: [ADSM-L] MOVE DRM & CHECKIN LIBVOL
From: Nick Laflamme <dplaflamme AT GMAIL DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Sun, 18 Jul 2010 17:31:27 -0500
It's fairly well documented that MOVE DRMEDIA will execute CHECKOUT LIBVOLUME 
under the covers to make "safe" any copypool volumes and DB backup volumes that 
are allegedly on their way off-site. 

Is it documented anywhere if MOVE DRM will execute CHECKIN LIBVOL commands?

At the root of my question is a discussion several of us are having at work. 
We're painfully aware that in some TSM 5.5 and TSM 6.1 levels, there's a 
deadlock related to DELETE VOLHISTORY fighting with LABEL LIBVOL. As we run 
with our virtual tape libraries set for RELABELSCRATCH=YES, we sometimes see an 
nearly spontaneous LABEL LIBVOL as volumes are released from storage pools or 
are no longer needed as DB backups. 

What's less agreed upon is whether we have to check in the returning volumes 
before executing a MOVE DRM FROMST=VAULTR TOST=ONSITER to trigger the bug or 
not. One colleague is adamant that MOVE DRM will find -- and relabel -- a 
volume that's sitting in the VTL even if it doesn't show up in a QUERY LIBVOL 
before we do the MOVE DRM. My own expectation is that if we do the MOVE DRM 
command before the volumes are checked back in the library, the automatic 
volume relabeling won't go on its own. 

Because the price of failure is a TSM server hang, we're a bit leary about 
playing with different permutations until we map out the various scenarios and 
outcomes. But, if we can convince ourselves that in some cases, MOVE DRM 
doesn't trigger an implicit LABEL LIBVOL if a library is set for 
RELABELSCRATCH=YES, we could return to checking in our "off-site" volumes and 
then, say, running manual LABEL LIBVOL OVERWRITE=YES  to have the same effect.

So, does anyone know about safe or unsafe sequences of MOVE DRM with or without 
CHECKIN LIBVOL when a library is set for RELABELSCRATCH=YES? 

Thanks,
Nick
<Prev in Thread] Current Thread [Next in Thread>