[5018] in linux-scsi channel archive
aic7xxx-5.x broke multi-lun detect?
daemon@ATHENA.MIT.EDU (Evan Willett)
Wed Oct 28 08:48:12 1998
From: Evan Willett <willett@maniac.pet.navo.hpc.mil>
To: linux-scsi@vger.rutgers.edu
Date: Wed, 28 Oct 1998 07:17:48 -0600 (CST)
I have a 2940uw with a CMD attached raid with 3 volumes lun 0-2.
Under 2.0.29 with aic7xxx-4.0 driver I am able to detect all 3 luns upon
bootup. Under 2.0.36-pre15 with aic7xxx-5.1.2 and aic7xxx-5.0.19, my
SMP=1 goliath mb system cannot auto detect ANY luns from the CMD when it
boots.
If I comment out the drives from fstab and boot anyway, I can
'echo "scsi add-single-device 0 1 2 3" >/proc/scsi/scsi' to get access to
lun 0 and lun1, but I cannot get lun 2. When I add lun 2, it shows up in
/proc/scsi/scsi and dmesg, but not in /proc/scsi/aic7xxx/0 and I cannot
mount it. I do have append="max_scsi_luns=8" in lilo.conf and
CONFIG_SCSI_MULTI_LUN=y.
Is this aic7xxx specific or 2.0.36-pre15?
Note everything works under 2.0.29 except that the system goes south sometimes
under very heavy network/disk loads.
I would like to use pre15 because it seems more stable under SMP. I can
pingflood the machine from 11 other systems while it is compiling kernel,
compiling dosemu and doing GB file transfer over 100TX vortex card.
I had tried some 2.1.x kernels, but I still had aic7xxx/vortex/SMP problems.
I guess I could try 2.1.127pre2 the next time I can take the machine down
or should I just wait for 2.2?
Evan
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu