This is probably a hardware issue, but since HP is of no help, I thought I'd
post here too.

We have a SQL/Exchange 2-node cluster(Windows 2003, Exchange 2003, & SQL
2000) attached to a SAN(MSA1000). Drives are configured in RAID1+0, with a
single array and 6 logical drives.

Three weeks ago, we expanded the 2nd logical drive(which looked like it
causedall the other logical drives to shift down into the free space area so
the free space could be added to the 2nd logical drive contiguously?). After
waiting for about 3.5hrs, the logical drive expansion was complete. I booted
up the node that owned the drive (SQL), and ran diskpart successfully to
expand the logical drive in Windows.

Everything looked good, no errors or obvious problems. The following week,
we started noticing some errors with our stored procedure backup jobs that
backup or SQL databases to another logical drive on the SAN. The timing was
off on a number of things, causing errors. After looking at the issue, we
determined that the SQL backups were taking 2-4x longer than before due to
slow right performance. (Veritas system backups take the same amount of
time, so read performance looks unaffected)

HP Tech Support is useless, and says that they don't deal with performance
issues, even though it had to do with tweaking the hardware.

Has anyone else every seen this? If so, do you know of a fix? I hope I
don't have to blow-away all the logical drives in the SAN and re-create them
to fix the issue. =(

I can't say why the issues but do have a comment<g>. If you get a chance to
reformat the array you should make it a single logical drive. Having
multiple logical drives on a single physical array has little advantages and
many disadvantages. It provides NO performance advantages and if you only
have a single logical array you would not have to maintain multiple volumes
and worry about one of them running short of space.

--
Andrew J. Kelly SQL MVP


"dathrill" <dathrill (AT) discussions (DOT) microsoft.com> wrote