Re: Serious SCSI problem on Lenny networkinstall image
Otavio Salvador wrote:
"letstry@home.nl" <letstry@home.nl> writes:
[...]
Before you guys go ask the whole standard "chain" of questions about
my SCSI setup, it might be nice to say that a debootstrap install of
Debian 5 (via Ubuntu livecd) + apt-get update/upgrade produced a
running system on both machines, without problems.
So the problem might be fixed there.
[...]
An interesting question about it is if you used regular Lenny kernel at
installed system and if it worked well or not.
Hi Otavio, list,
I did some more searching on the SCSI errors I encountered using the Debian 5 network install image.
For now I just focussed on the dual-XEON with Qlogic controller.
Kernel differences:
network install image: 2.6.26-1-486 (Debian 2.6.26-13)
debootstrapped install: 2.6.26-1-686 (Debian 2.6.26-13)
After stripping both (dmesg on running 686 system, and syslog.xeon of the network install image) message files to SCSI-only
messages, I see that until the 'sd 0:0:0:0: [sda] Attached SCSI disk' message, everything is the same.
From there the 686-based kernel goes OK, and the 486-based network install image goes wrong.
I just installed the 2.6.26-1-486 (Debian 2.6.26-13) kernel on the running machine, booted into it, and got errors again on the
SCSI controller. I can't even boot into the root filesystem, so no dmesg or syslog output.
So my conclusion for now is that it's not in the network install image as such, but in the 486 kernel that was used for the
network install image. There are differences in the source tree of 2.6.26-1-486 and 2.6.26-1-686 that produce an impossible boot
situation on the SCSI device when the 486 kernel is used.
I'll invest some time to see if I can find any differences in the source of both kernels.
Regards,
Jacco
Reply to: