adaptec 2810sa ide raid problem

["Grunloh, Robert" <grunlohr u library arizona edu>]

Hi everyone,

we have a new dual AMD64 server with an Adaptec 2810sa ide raid card. I have
the card seemingly working, but the system locks up overnight (nothing in
the logs or on screen) and I suspect driver issues. Has anyone had
experience with this card and rhel3 for amd64 (fully patched with
2.4.21-15.0.3.ELsmp)?

The salesperson told the developer who bought this that the card was
supported, but Adaptec's site seems to only support rhel2.1; Redhat doesn't
list this card as compatible on their site. Adaptec does have a beta rhel3
driver disk from december but it's for outdated kernels. And I see that
rhel3 update3 beta updates the aacraid driver from 1.1.2 to 1.1.5-2339, but
I'm a bit reluctant to use that for production; I would be interested to
know if it works for people with this card. I could download the iso's and
try it.

Longer version:

The card firmware is currently at build 7211.
The server has a separate ide boot disk with the OS, so I can boot and run
without using the card, which helps with troubleshooting.

The vendor (Western Scientific) shipped the system with rhel3update2
installed and everything worked. However they partitioned the system oddly,
so I decided to reinstall. If I had known to expect driver problems I would
have recorded more info about the firmware/software state as they shipped
it. I did contact them later but got no useful info.

I downloaded the current adaptec firmware and flashed it (ver7244),
reinstalled rhel3 and patched it, and discovered the card could not be seen.
I downloaded the current adaptec drivers (one page says they support rhel3,
all other pages only say rhel2.1) and tried to install them, the rpm scripts
failed because the kernel version wasn't expected. Adaptec's kernel support
seems to end around 2.4.21-4.

I then reflashed the card firmware with the version on the factory cd, which
is 7211 (no idea, unfortunately, if that was what was shipped or not). Now a
modprobe aacraid worked, I could use fdisk, mount them ok, etc., but
occasionally the system would stall for a minute or 2, even when no traffic
was on the raid array; and now it's hung overnight, with nothing I can find
in the logs. It did respond to pings but nothing else over the network. My
remote sessions left open overnight (ssh, normally fine with all servers in
the building) appeared ok, no errors or disconnections, but did not respond
at all to any keystrokes.

Adaptec's cli tools didn't work either (not essential).

Thanks for any info. If this isn't reliable, how are people's experiences
with 3ware cards?

Robert Grunloh
University of Arizona Library