Remember that the source during the rebuild of the offsite tapes is the
primary tapes.  So, if the primary pool is reclaimed it would be no
different than a backup to copy stg pool.    My understanding was that what
you are saying is only true if the aggregates are in the same pool.  The
other thing is you can do a RECONSTRUCT=YES and it will remove the deleted
portion of the aggregates within storage pool as it is reconstructed.

-----Original Message-----
From: Steve Roder [mailto:[EMAIL PROTECTED]]
Sent: Saturday, January 26, 2002 7:22 PM
To: [EMAIL PROTECTED]
Subject: Re: Mitigating Risk with TSM's incremental backups


> That tells me what is coming back soon.  You can get elaborate and select
> only the volume name and pipe the stuff into the commands necessary to
issue
> the commands dynamically, but I just do them manually right now.  We are
> heading down the automated path soon.
>
> The command that I use to move the data is:
>
> move data [volume_name] stg=[same_copy_pool]
>
> By doing this I am not exposed and I my operations department was able to
> continue a tape rotation they are used do.  The side effect is reclamation
> is automatically done by doing this.  I love 2 bird solutions.

Note that the move data command differs from true reclaimation in that
move data does not rebuild aggregates, so the reclaimable space on a tape
that is within an aggregate will not free up.  In other words, if you had
an aggregate of 10MB on one of these tapes, and all of the data within it
was expired, except for a single 1 byte file, move data will copy the
entire 10MB aggregate to the new output volumes.  Depending on the nature
of your data, this could actually add up to a lot of wasted tapes, as that
10MB aggregate will be copied around until that last 1 byte file is
expired.

A nit?  Perhaps...and perhaps not.


Steve Roder, University at Buffalo
HOD Service Coordinator
VM Systems Programmer
UNIX Systems Administrator (Solaris and AIX)
TSM/ADSM Administrator
([EMAIL PROTECTED] | (716)645-3564 |
http://ubvm.cc.buffalo.edu/~tkssteve)

Reply via email to