Mike A. Lewis, CNE wrote:
>
> New RH5.0 install. AST Manhattan with Adaptec 2940UW controller (same
> controller working with RH5.0 in another machine).
>
> Earlier today, installed RH5.0 via FTP. Absolutely no problems during
> install. After finishing installation, when rebooting, I get the
> following:
>
> Scsi: 0 hosts
> Scsi: detected total
> Partition check
So far so good, this is right for an initrd based boot setup.
> Ramdisk: Compressed image found at block 0
> Crc errorVFS: Cannot open root device 08:01
^^^^^^^^^
This is the problem. It's getting a crc error reading the initrd file from
the boot directory. The file is responsible for loading the aic7xxx driver
so that you can read/write to your scsi disk once the system is booted up.
> Kernel panic: VFS: Unable to mount root fs on 08:01
Just what would be expected if you can't load that aic7xxx driver and it
needs to access a root partition on that scsi controller.
The answer you ask? Boot up the install disks in rescue mode, load the
aic7xxx driver, mount the root partition of the hard disk, go into the
root-mnt/boot directory and remake the initrd file (something like mkinitrd
should work, or just gunzip the initrd file and see if you can find the
problem). After you've remade the initrd file, re-run lilo to re-init the
boot setup on the hard drive. Things should work fine after that.
NOTE: I'm not sure the exact options to make the mkinitrd package rebuild
from a specific mount point, which is what you need, but to make lilo
re-init the hard disk, do the following. Assuming you mount your root
partition on the /mnt directory, the lilo command to use would be "lilo -r
/mnt" and it would do the right thing. You may have to use /mnt/sbin/lilo
if lilo isn't in the current search patch. Additionally, the initrd file
may not be corrupted, lilo may just have a bad map of the file. If that's
the case, the lilo command should be enough to fix the problem without
having to remake the initrd file. You can find out if this is the case by
going into the /mnt/boot directory and issuing the following commands:
mv initrd initrd.gz
gunzip initrd.gz
^ If this commands succeeds, then you shouldn't need to remake the
initrd file, at this point you would just put it back like it was
by issuing the next two commands
gzip initrd
mv initrd.gz initrd
Then issue the following command before re-running lilo to make sure all of
the disk blocks are allocated and properly written:
sync
Then re-run the lilo command as I mentioned above. If the gunzip command
does fail, then you need to remake the initrd file and I'm sure someone here
is up on the mkinitrd command more that I am.
--
Doug Ledford <[EMAIL PROTECTED]>
Opinions expressed are my own, but
they should be everybody's.
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
