ADSM-L

Re: [ADSM-L] separation of copypool tapes

2007-07-13 08:04:01
Subject: Re: [ADSM-L] separation of copypool tapes
From: Nick Laflamme <nick.laflamme AT NIST DOT GOV>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Fri, 13 Jul 2007 08:00:31 -0400
Richard Sims wrote:
On Jul 12, 2007, at 3:14 PM, Haberstroh, Debbie (IT) wrote:

Deleting all of my copypool volumes could be a problem since I am
only moving half of the tapepool data and have over 600 copypool
tapes.  I will need to find a better way to remove the data. Thanks
for the answer about the data still being there, I wasn't sure how
that would work.

Keep in mind the basic principle that an object remains in the
copypool ONLY as long as it the original object remains in the primary
pool, a tenet which you can use to advantage.

While I know better than to challenge or contradict someone like Richard
too often, I'm curious about how I can prove this to myself.

We've been doing a lot of MODE NODEDATA commands recently as we move
data into pools that collocate by group. The copy storage pools that
correspond to the original smaller tape storage pools don't seem to be
shrinking and expiring as long as the data is still somewhere on the
managed node. We seem to be ending up with three copies of data: the
primary, now in the collocated pool; a secondary in the copy pool of the
collocated pool, and another secondary in the copy pool of the older,
smaller pool.

We're running TSM 5.3 on AIX, if that matter, but for something that
Richard describes as a "tenet", one would hope this goes back to WDSF on
VM days. :)

    Richard Sims

Nick Laflamme