ADSM-L

Recovering primary storage pool

2006-09-08 13:58:28
Subject: Recovering primary storage pool
From: "Mueller, Ken" <KMueller AT MCARTA DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Fri, 8 Sep 2006 13:57:52 -0400
Following on the heels of the postings earlier this week about restoring
primary storage pools from copy pools, I was curious about the process used
to perform the restore:

If I have 'large objects' (Oracle databases, Exchange Databases, image
Backups, etc) in one or more copy pools, client restores directly from the
copy pool would make sense - you would likely only need 1 or 2 tapes to
restore the client data regardless of colocation.

In the case where I have millions of 'small objects' (document imaging file
system), spread across many, many copypool tapes, it seems like a restore to
primary stgpool is the way to go.   However, is the restore stgpool going to
deal with each individual file, or does it work with aggregates, or some
other 'block' to speed things along?   In other words, if I did a restore
stgpool for a pool containing 10 100G objects vs 20,000,000 50k objects both
occupying the same number of tapes - would the restore time be roughly the
same, or dramatically different.  (I'm ready for the "why don't you test it
and tell us" comment somebody will lob my way!)

Wouldn't it also make sense to have a 'restore stgpool xxx inactive=no' to
only restore the active versions of files.  In a D/R scenario, you only care
about the most recent version of an object (in most cases).  If you did need
to go back to a prior version, it would automatically hit your copypool.
That would make restoring the primary pools a much quicker process,
especially for those that retain many versions.  Once your clients are up
and running, you could then backfill the primaries with a 'restore stgpool
xxx inactive=yes'.

-Ken

<Prev in Thread] Current Thread [Next in Thread>
  • Recovering primary storage pool, Mueller, Ken <=