I think you probably need to adopt a process of elimination.
Firstly, you say that it backs up at 8MBps. You don't mention whether
you have 100fdx or GB NIC, but at 100fdx this throughput is OK.

Next, are you restoring back to the same server that you have backed up
from.
If so, this more or less eliminates your target h/w and network settings
(apart from the overhead of writing to disk as opposed to reading)

If you are restoring back to different h/w then as previously mentioned,
check the NIC settings and network. Perhaps perform FTP of a largish
file(s) from target server to TSM server.

So, assuming you are restoring back to the same H/W from whence your
backup was taken, I now assume that your backup was going to diskpool
and not exceeding the maxsize threshold of storagepool. If so, why not
try to move nodedata from the primary pool tape to the primary pool
disk. Make sure you don't have the migration threshold set low. Check
the throughput of the move nodedata and then attempt the restore from
the diskpool.

If thoughput is improved from disk, then try to force the restore from a
different drive to eliminate drive from equation. Also run cleaning tape
through drive. Attempt move data from primary tape to different tape to
eliminate specific tape issue.

Have you noticed issues with your migration to tape throughput lately.
I guess the key thing is to determine whether you have an across the
board throughput issue with your tape operations, or it is indeed
localised to this 1 TDP restore.

Leigh


-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Orin Rehorst
Sent: 03 November 2006 13:56
To: ADSM-L AT VM.MARIST DOT EDU
Subject: [SPAM: 4.000] [ADSM-L] Sloooow Restore

TDP for Exchange backs up at 8 MB per second.

It restores at 8 MB in 56 seconds!

What may be wrong?

Environment: OS Win2003, TSM 5.3.2, IBM tape library 3583 LTO 1.

TIA
Orin Rehorst