ADSM-L

Re: AW: Designing a solution for fast manual restore

2000-02-06 19:24:40
Subject: Re: AW: Designing a solution for fast manual restore
From: Eric Winters <ewinters AT AU1.IBM DOT COM>
Date: Mon, 7 Feb 2000 11:24:40 +1100
Juraj,

Thanks for the comprehensive reply. TSM 3.7 sounds like the best way ahead
to give us what we need.

Concerning your following question....

*************************************************************************************************************
P.S. I have a question about your process:
You create tapes in an storage pool in an tape library,
and during disaster recovery you restore
from this tapes using singe tape drives.
How do you manage this?
As far as I now one cannot move tapes among storage pools
and one cannot mix libraries and single drives in one
device class => storage pool?
One even cannot combine two identical libraries together!
*************************************************************************************************************
What we actually do is we have a primary storage pool of cartridges which
stay in the library.
We also have a collocated copypool for the most critical data kept offsite
(there is also a non collocated copypool kept offsite for the not so
critical data). Reclamation of the offsite copypools is done daily.

The disaster we are trying to prepare for is one in which we lose all
hardware (server, library and all cartridges within the library). We would
take our copypool cartridges to another location where we have a system
with the four external drives on standby. The database would be restored,
the primary copypools would be marked as destroyed and the client data
directly restored from the copypool as quickly as possible. This has been
tested.

This approach gives us highly collocated data but is very heavy on the
recovery log and as others have suggested, the reclamation activity is
intense too.

Another ADSMer raised the question if we had ever tried restoring from
pools with collocation off. No, this has not been done. The disaster
recovery testing was done prior to my arrival in the picture. At that time
the system had space to spare. But the backup requirements have increased
rapidly  in the meantime and it is now that the high 'cost' (rec log
filling up, time taken to reclaim volumes) is becoming apparent. The fear
is that if collocation were not used, the data for a single filesystem
could be on tens of cartridges. There could potentially be a great deal of
manual mounting and dismounting, and more importantly time would be wasted.
But this is all supposition.

I suspect a future with 3590 E drives (we have B drives today), TSM 3.7,
incremental backups only and non collocated offsite copypools lies ahead.

Thank you all for your suggestions. Alternative points of view are
extremely helpful.

Cheers,

Eric Winters
<Prev in Thread] Current Thread [Next in Thread>