Link to home
Start Free TrialLog in
Avatar of RickNCN
RickNCNFlag for United States of America

asked on

fatal: open /dev/sdc: no such device or address: WHY?

I'm in the process of upgrading the disks on my server. (Dell PowerEdge 2300 with Adaptec 2110s controller) Old setup: 4 36G RAID 5  + 1 137G for backups of user folders. I used Acronis Enterprise to make a backup of my partitions to another 137G drive. I replaced the old 36's with 4  new 137G's. Created the new array RAID5, again booted from CD to Acronis, restored the main partition of about 100GB and the 2G swap.

The first problem came when, after the upgrade,  the SCSI controller labeled my RAID array as sdb instead of sda as it used to be. This is because in the physical order of disks inserted into the Dell backplane, the Acronis backup 137G drive is first (slot 0,0), the RAID disks are next (slots 0,1 - 0,4) and the users backup drive, 137G, is last (slot 0,5). Since my boot disk USED to be sda, I booted from a CD, mounted sdb1, chroot to /mnt/sdb1, changed /etc/fstab and /etc/lilo.conf to point to sdb instead. I ran /sbin/lilo and it seemed to write the mbr. It never found LILO with that configuration, I believe (it's tough to remember everything when you try so many different things)

So.... I physically removed drive in slot (0,0) the Acronis backup. It is unneeded now since I restored already. Now, SCSI controller sees my array as sda and backup drive as sdb. I again booted from CD, mounted sda1, chroot-ed to /mnt/sda1 and modified fstab and lilo.conf back to pointing to sda. This time, when I ran /sbin/lilo, I get:

-----------------------------
Warning LBA32 addressing assumed
raid_setup returns offset = 00000000 ndisk = 0
BIOS    VolumeID    Device
Reading boot sector from /dev/sda
Fatal: open /dev/sdc: No such device or address
-----------------------------

 I tried  running lilo -v -v, didn't help
ASKER CERTIFIED SOLUTION
Avatar of akirhol
akirhol
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of RickNCN

ASKER

no, that's the weird thing, there are no references to sdc in fstab or lilo.conf!. My friend thinks maybe the system map is corrupt. He suggested reinstalling Slack. But that will entail a kernel upgrade. We've been using slack 9.1 which comes with 2.4.22 and it's been updated to kernel 2.4.23 I think - for multi processor support.  I really don't want to reinstall. Can you think of any other reason for this at all?
Avatar of RickNCN

ASKER

Well, we couldn't figure out what the problem was so we went ahead and saved the etc directory and loaded slack 12.1 and restored etc and the conf  files. So - no resolution, but you gave some good suggestions. Thanks