Re: [SLUG] error loading kernel

From: John Danielson, II (jdii1215@comcast.net)
Date: Mon Jun 03 2002 - 16:03:03 EDT


Mikes work account wrote:

>It did not indicate bad magic at all, just unable to mount fs for the
>kernel. I think I will redo the initrd but don't know how that will help.
>I have the old rpm for that kernel, maybe I should just remove the old rpm
>and redo it again...
>
>
>Michael C. Rock
>
>
>-----Original Message-----
>From: slug@lists.nks.net [mailto:slug@lists.nks.net]On Behalf Of
>awyatt@fewt.com
>Sent: Monday, June 03, 2002 10:03 AM
>To: slug@nks.net
>Subject: Re: [SLUG] error loading kernel
>
>
>
>If it says "bad magic" boot your cd, go into recovery mode and fsck you
>/dev's. Then try making an initrd for that kernel. Sounds like a problem I
>had a few weeks ago.
>
>
>
>You Wrote:
>
>
>Date: Tue, 4 Jun 2002 10:04:02 -0400
>
>From: "Mikes work account" <mrock@stewartsigns.com>
>
>To: <slug@nks.net>
>
>Subject: [SLUG] error loading kernel
>
>
>
>I have a few kernels loaded that I can boot into and for some reason the
>
>2.4.18smp kernel is giving me the following error message when attempting to
>
>load:
>
>
>Kernel Panic: Unable to mount root fs on fs 08:07
>
>
>This kernel has been running the system for several weeks until now.
>
>
>I have checked the /boot partition and all looks ok to me and I have checked
>
>lilo.conf and nothing looks bad their either. We had an extended power
>
>failure last week during my absence and my supervisor rebooted into the
>
>default kernel which is very old and did not notice it. I came in and tried
>
>to reboot into the proper kernel and found this.
>
>
>I have booted the system into an older kernel without any problems and that
>
>is where we are now running, but that does not solve the problem.
>
>
>What else should I check?
>
>
>Michael C. Rock
>
>
>
>
>
>
>
Given that, try checking and seeing if your old kernel also uses a
different fstab. If so, please see if somehow th eold kernel boot did
something like addressing EXT3s as EXT2s and then stripping the
journalling as a result. A Parted or Ranish run might tell you this
faster. Prefer fixing with Parted, however, as the Ranish that can
handle EXT3 right is still beta-- given that this appears to be a
production box, would back up if possible first also....

As for the multiple kernels, you might find some versions of running
software need some of the software loaded(called dependency h*** by
some)when the older kernels are used, would be careful with just syncing
unless you can do it when you expect at least a few hours of inactivity
and have a FULL backup in hand. An unhooked RAID or soft created backup
image\mirror snapped just before you experiment, with the image taken
offline, might be nice during the experimentation phase. Actually, if
were an option, would image to another box entirely, experiment and
annotate the changes there, then copy the process on a server.

John.

-- 
_______________________________________________________________________________
Registered Linux User #276212, Machine #158988
"Use what works best for your needs, at minimum total cost of ownership."



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 20:18:32 EDT