Doug,
I have not heard from you, so I am assuming you don't think this is a
problem in the driver. I re-tested on another system with an NCR
chipset, and the SCSI recovered during the failover testing. This
problem appears to be unique to the Adaptec Driver.
:-)
Jeff
"Jeff V. Merkey" wrote:
>
> Doug,
>
> I downloaded the 5.1.24-5.1.29 patches for the Adaptec Driver as you
> directed me to do, and I tested them with a little better luck, but not
> much. I also checked over the POCA and it does appear that we have the
> jumpers on the SCSI backplane all configured correctly for both buses
> (wanted to make certain the buses were terminated correctly for the
> slots we were using).
>
> After I pull disk 0 on bus 0 out of the system during mirrored I/O this
> is what gets printed on the screen. The newer drivers seem to have
> stopped the scrolling problem, so there's more info to report and it
> retries and reprints this info every two or three seconds over and over
> (/dev/sdb is the Linux root volume on this system, not /dev/sda).
>
> SCSI Disk I/O error: dev 08:00 sector XXXXXXXX (this number cycles
> through different LBA offsets)
> scsidisk I/O error: host 0 channel 0 id 8 lun 0 return code = 26030000
> SCSI bus is being reset for host 0 channel 0
> SCSI0: resetting for second half of retries
> ....
> [ ... this message repeats over and over on the system console ... ]
>
> After I reinsert the hard disk drive, I then start getting a different
> error, which is:
>
> scsidisk I/O error: host 0 channel 0 id 8 lun 0 return code = 280000002
> [valid=0] Info fld=0x0 Current sd08:00 sense Key not ready
> Additional sense indicates logical unit not ready, initializing command
> required
> ....
> [ ... this message also repeats over and over on the system console ...
> ]
>
> After about 10 minutes, the system finally completely feezes up, and has
> to be powered down and restarted. I put in a couple of IDE disks on
> this box to make certain this was SCSI specific, mirrored them, then
> jerked out a drive while I/O was in process, the system recovered.
>
> :-)
>
> Jeff
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Apr 30 2000 - 21:00:16 EST