You are correct.
Another 2 cents:
long-term archive does not only consist of the sole problem to keep the
data safe for a long period of time.
Some other points are often overlooked, here only some of them:
- availabilty of system which the data can be restored on
(if you backed up your unix file 6 year ago with TSM and moved later
to NT,
you will not be able to retrieve your data unless you keep a working
unix machine alive,
even worse if you moved to veritas or Tivoli stopped TSM development
in the meantime..)
- availability of programs which can display / operate on data
or are you able today with the documentation you wrote 10 years ago
with an XY editor?
Evene with text files - moved from ebdic to ascii?
How about databases and so-called databases?
- availability of personal who knows the data and how to use it.
Even if this had been documented - will you find the documentation?
- tracking the archive
Even if you still have the data, would you be able the correct data
to be restored 10 years ago?
- access rights: will you know in ten years who has the acces to the
restore?
If protected by passord - will this be available?
etc. etc.
To summarize:
you have to think not only about archiving the data,
but about archiving of the whole working environment.
regards
Juraj Salak
-----Original Message-----
From: Wayne T. Smith [mailto:[EMAIL PROTECTED]]
Sent: Friday, February 15, 2002 3:00 PM
To: [EMAIL PROTECTED]
Subject: Re: Backup Sets for Long Term Storage
On 12 Feb 2002 at 21:14, Kelly Lipp wrote, in part:
> I believe the key to long term storage is the notion of data
> refreshment on the tapes. With reclamation, we get that. If archive
> data is mixed with backup data we get reclamation due to backup
> retention policies being much less (typically) than archive. Some will
> argue that moving this data around isn't efficient, but if ensuring that
> data can be read is the goal, moving it around occasionally is
> important.
I agree that long term storage is an interesting problem, but I don't
see reclamation (alone) as solving it. It may not be efficient, but it
isn't sufficient! ;-) For example, I believe I put my current drives
into production about 2 years ago. 3% (about 60) of my tapes are
"full" and haven't been mounted since 2000! Reclamation has caused
most data to move, but not all.
If one believes long term storage is important and moving data around
important to achieving success, then reclamation is not sufficient.
*SM doesn't automate mandated movement of data on it's storage, but it
enables it, since it is trivial to write SQL to find "old" tapes (so
one might run MOVE DATA on them). Moving data from "old" tapes allows
one to rebuild from another (hopefully still "good") copy, should a
failure occur.
cheers, wayne
Wayne T. Smith [EMAIL PROTECTED]
ADSM Technical Coordinator - UNET University of Maine System
