SCSI timeouts with aacraid on Dell PowerEdge 2400

[Christopher Arndt]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I get SCSI timeouts with the aacraid driver on Ubuntu Breezy on my Dell
Server with the result that access to the /home partition regularly
hangs until after reboot.

I wanted to replace the Windows 2000 Server installation on our old Dell
PowerEdge 2400 Server with a Linux/Samba setup because it was acting
only as a fileserver. I decided to install Ubuntu Breezy. First the
installer was unable to format the /home partition. After I updated the
BIOS and the RAID firmware to the latest version, the installation was
succesful but now I get the timeout problems.

The setup is as follows:

- - Dell PowerEdge 2400 PIII 667 Mhz with Adaptec PERC2/Si - single RAID
channel, Ultra2 SCSI

- - One RAID mirror array with 2x 8.5G SCSI disks
  - partition 1: Dell Utilities
  - partition 2: NTFS (Windows Server)
  - partition 3: ext3 (/boot)
  - partion 4 (extended): LVM (contains /root and swap partition)

- - One RAID 5 array with 4x 33G SCSI disks => ~107G
  - partition 1: ext3 (/home partition)

- - Updated the BIOS to version A09 and the RAID firmware to version
2.8.1.6098, A14

- - Normal installation with X + some servers, Kernel "linux-386".

When the /home partition crashes, I get the errors shown in the attached
excerpts from /var/log/messages.

Further observations:

- - After the crash 'mount' still shows the partition as mounted but when
I do an 'ls' it just shows nothing.
- - When I unmount the partition and try to mount it again, mount
complains it can not find a valid superblock
- - after a reboot I can mount the partition again, but it will crash
later again.
- - I first had a logical volume accross the whole partition, but reverted
to a normal partition after the error occured first. But the error
persisted.
- - the RAID firmware control utility shows that the array is in "Optimal"
condition and I even low-level formatted all the disk from there before
creating the RAID array.
- - the server ran fine under Windows 2000 Server
- - the first RAID array (mirroring) seems to run fine

Has anybody experience with running Ubuntu on a PowerEdge 2400? Should I
try to compile a vanilla kernel? Are there any known issues with the
aacraid module?

I would really appreciate any hints on this, as I want to get the server
up and running again as quickly as possible.

Thanks, Chris
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Puzzled by the all these strange characters?
Comment: See http://smurl.name/lz5 for an introduction
Comment: to public key encryption and signing.

iD8DBQFDfHV3yacnDOuTzPwRAhqoAJ9pAkpxD6O3H+RdHXFMRmRqRJtM7ACcC2g+
/V6NeeRhYatfrfFlj9ZkrmM=
=HlaR
-----END PGP SIGNATURE-----
-------------- next part --------------
A non-text attachment was scrubbed...
Name: raidfailure.log
Type: text/x-log
Size: 4766 bytes
Desc: not available
URL: <https://lists.ubuntu.com/archives/ubuntu-users/attachments/20051117/2a0a0bca/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: chris.vcf
Type: text/x-vcard
Size: 373 bytes
Desc: not available
URL: <https://lists.ubuntu.com/archives/ubuntu-users/attachments/20051117/2a0a0bca/attachment.vcf>