Find the answer to your Linux question:
Results 1 to 5 of 5
Hello All, Background Debian linux 2.0 with Kernel 2.2.18pre IDE. Root is on the IDE and everything else is on the SCSI disks. I installed the new kernel 2.4.18 and ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1
    Just Joined!
    Join Date
    Dec 2002
    Location
    Boston
    Posts
    8

    failed Kernel upgrade, system totally down


    Hello All,

    Background
    Debian linux 2.0 with Kernel 2.2.18pre IDE. Root is on the IDE and everything else is on the SCSI disks. I installed the new kernel 2.4.18 and made the changes to the lilo.conf file and ran liloconf, everything seemed to go fine and as normal. Reboot and the system cant or wont load the SCSI mods so I get no swap or any other file system.

    I get error Modprobe:modprobe can'r locate module block-major-8

    I tried booting for a rescue disk but I wont load, So somehow I have now totally hosed the system.

    Any ideas on how to get out of this?

    I have tried repointing back to the old kernel img with no luck

    Hartman

  2. #2
    Linux Guru
    Join Date
    Oct 2001
    Location
    Täby, Sweden
    Posts
    7,578
    Did you install the kernel from a package or did you compile it yourself? It would seem that block-major-8 is missing from /etc/modules.conf, or that the module that it refers to isn't there. Maybe SCSI was compiled into your previous kernel?
    When does the error occur? Does it make the system hang? Why doesn't the boot disk load (what happens instead)?
    You should have saved your old kernel in case something went wrong, though, but I guess it's too late for that now. Maybe you did, I'm not sure what you mean by "repointing" back to the old kernel. What do you mean when you say that you don't had any luck with it? What happens?

  3. #3
    Just Joined!
    Join Date
    Dec 2002
    Location
    Boston
    Posts
    8
    Well,

    I first build the system with only IDE drives. I added SCSiI later and the system worked fine. I installed the new kernel with apt-get on debian. The thing must have rebuilt the new kernel without SCSI built in. So when I reboot and only "root" is on the IDE the system fails becuase swap,/var, /etc,/home are all on the SCSI disks.

    I have rebuilt the box onto a new IDE drive and am trying to mount the old SCSI to restore the system but I cant mount them

    I get the following error
    mount: /dev/sda1 is not a valid block device

    I can see the partitions in fdisk but that is all

    Hartman

  4. $spacer_open
    $spacer_close
  5. #4
    Linux Guru
    Join Date
    Oct 2001
    Location
    Täby, Sweden
    Posts
    7,578
    It seems strange that you could have /etc on a SCSI disk, since init should have to have /etc/inittab before anything else, unless you've done that in some alternative way that mounts /etc before /etc/inittab is used, but I know of no such (at the moment). How had you done that?
    Can't you get the SCSI modules for your kernel from somewhere and use them to set things right?

  6. #5
    Just Joined!
    Join Date
    Dec 2002
    Location
    Boston
    Posts
    8
    Here is the order.

    1) Built the system with IDE only, Debian 2.2.18pre
    2) ran out of disk space
    3) Installed a SCSI controller with 2 9GB 10K drives
    4) Moved /etc /var /home from the IDE to the SCSI disks plus swap
    5) System worked sweet no issues, rebooted fine

    6) Have 2 other newer servers running Debian 3.0 with kernel 2.4.18 wanted to get the kernels all the same

    7) Installed the new kernel on the system, made the initrd img added the line to lilo.conf, and ran liloconf. Everythings looked fine.

    reboot and now scsi, no scsi means no swap /etc etc...

    9) Booted off a new fb2.4 disk and tried to rescue root=/dev/hda1
    10) SCSI would not come up at all
    11) Rebuild system on the IDE and tried to get the scsi drives mounted so to get the data off, wont mount bad block device error

    12) Replaced old SCSI drives with new one mounts fine

    So basiclly the part table on the old SCSI are hosed. I can see them vi fdisk but for the life of me they will not mount

    Hartman

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •