CentOS

Hi all,

I'm having some problems with a Adaptec 29320A under CentOS 5, i386.

Setup:

HP ML110 G4 (Xeon 3040, 1.5GB RAM), Adaptec 29320A, 1 x SCSI seagate disk, 2 x sata disks.


I'm getting lots of these:

[quote]
scsi2: SCSI offset overrun detected. Resetting bus.
scsi2: device overrun (status a) on 0:0:0
[/quote]

I've changed the cable (Both cables are advertised as being U320) - this made no difference, next I knocked the speed from 320 to 160 on the controller card and it seems to stopped doing the above.

I haven't had chance to test the machine under windows so I can't rule out a driver issue (yet). I'm using the driver provided with CentOS, I've downloaded the one from the Adaptec website but not sure on the installation steps (had a quick look but couldn't find anything - need another look).

Anyone had this problem and solved it?

Cheers, Rich

I've found in U320 mode turning off QAS in the BIOS of the SCSI card seems to be a work around.

Here's what Adaptec says about QAS:

[quote]
QAS—(Default:Yes) Quick Arbitration and Selection reduces
the overhead of control release on the SCSI bus from one device
to another. This improvement reduces command overhead and
maximizes bus utilization. When enabled, this option is
available if it is implemented in the device drivers. However,
this option is not supported by the BIOS.
[/quote]

Which could suggest a driver issue.

I did find something via Google about a *BSD user having similar problems and for them dropping down to U160 or turning of QAS 'worked'.

I've also asked Seagate about firmware updates for the hard disk.

Does anyone know if I will loose much from having QAS disabled?

edit: I've also seen reports suggesting the use of LVM can cause problems and a normal ext3 partition would be ok...need to check this out.