And I would also factor in how often is this archive going to be accessed?
If, like most archives, never, then I really like the backupset option.  If
there is a likelihood of frequent access then you will want to have the data
more readily accessible so a second TSM instance just for that data (and
other archive data) might be the ticket.  Then one might ask should I do
this in 5.5 or 6.2?  5.5 would require a much smaller database but the
instance would then need to be on a separate server from the 6.2 instance.

Finally, you say millions of files.  Is the number tens or hundreds of
millions?  Tens I don't worry so much.  Hundreds, I start to worry.

Thanks,

Kelly J. Lipp
Expert Archive Solutions
719-531-5574

-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Maurice van 't Loo
Sent: Sunday, November 07, 2010 4:56 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: [ADSM-L] TSM Archiving 50tb data with millions of files

Hi,

Also think about about what and why.

If you need to archive 50TB of data to create space, to archive just
once and move the data to TSM, just do it, it isn't such a problem.

But if you do the archive as an other way of doing for instance
monthly backups, don't use archives, use backupsets instead.

And if you do a lot of archives in a way as it ment to be, and there
is no other way, you can also think about an other instance, only for
archives.
Normaly archives has long retentions, so expire inventory doesn't need
to run daily, but can also run once a month. And if all archives in a
stgpool has the same retention, you don't need to run space reclaim
too.

Good luck,
Maurice van 't Loo

> Need to archive 50tb of data with millions of files, here is the caveat,
tsm db at 220gb, planning on upgrading from 5.5 to 6.2 in the near future,
but currently waiting for newer hardware, how would one accomplish this
without causing too much db bloat? Any ideas would be greatly appreciated.