Tim --

This should be driven by your D/R requirements.

We have two MS-Exchange servers and a fax server that need to come up fast,
as well as a 600 GB SAP/Oracle instance on R/S 6000. We do *not* do bare
metal restores. As a result, I have:

1) A copy pool for SAP database backups (two tapes per backup)
2) A co-located copy pool for MS-Exchange
3) Two copy pools for SAP/Oracle redo logs (If one tape is unreadable, we've
got the other)
4) A copy pool for AIX/NT incrementals (backups, in other words)
5) A copy pool for long-term SAP Archives from commonstore (to keep them out
of the way)

On a 'normal' D/R (tests, so far), we get mksysb restores running on all the
R/S 6000 systems (six) and start installing NT from CD on the three NT
servers. About the time the TSM server has been recovered the NT systems are
ready to start restoring. About 45 minutes later, we're ready to start
recovering the non-database SAP files from one of the redo log tapes (I
archive this every week); three filesystems use the archive as a base for
point-in-time restores from the incremental tape.

As soon as we have the code down, we start the SAP database restore. While
it's running, we recover all the off-line redo logs to roll forward.
Somewhere in here we recover two oracle databases used by our in-house
labor/payroll and edi processing -- these archive nightly and end up on the
redo log tapes as well.

Our published requirements are to recover to a state that existed not more
than 12 hours before an outage, and to do it within 48 hours of the
occurrance of the outage. And we're trying to cut that 12-hour window down
to 3 hours.

So - determine what you need to recover, how it is to be recovered, and how
long you have to do it. From there you can work out the copy pool (and
original storage pool) requirements and the D/R site hardware requirements.
Don't forget that you also need to get the TSM database off-site, as well as
the device config and volume history files (I put these on floppy and strap
it to the TSM database backup tape).

Tom Kauffman
NIBCO, Inc

> -----Original Message-----
> From: Tim Melly [mailto:tim.melly.b AT BAYER DOT COM]
> Sent: Tuesday, November 13, 2001 3:48 PM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Archives, Storage Pools, and DR
>
>
> I currently have 2 primary storage pools, one for archives and one for
> backups.
> From a DR standpoint, should I have both primary storage
> pools copy to a
> single copy storage pool or should I split them into two copy storage
> pools.
>
> Any thoughts......
>
> Regards, Tim
> NAFTA IS Technical Operations
> (203) 812-3469
> tim.melly.b AT bayer DOT com
>