Schmidt, Florian wrote:
-----Ursprüngliche Nachricht-----
Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Im
Auftrag von John Summerfield
Gesendet: Dienstag, 29. Juli 2008 17:10
An: Red Hat Enterprise Linux 5 (Tikanga) discussion mailing-list
Betreff: Re: AW: AW: AW: [rhelv5-list] RAMDISK :ran outof compressed
datainvalidcompressed format (err=1)
Schmidt, Florian wrote:
-----Urspr�ngliche Nachricht-----
Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Im
Auftrag von John Summerfield
Gesendet: Dienstag, 29. Juli 2008 16:23
An: Red Hat Enterprise Linux 5 (Tikanga) discussion mailing-list
Betreff: Re: AW: AW: [rhelv5-list] RAMDISK : ran outof compressed
datainvalidcompressed format (err=1)
Schmidt, Florian wrote:
So I will now try your advice. What steps are neccesary to check the filesystem
on that device?
That depends:-)
I'll boot into rescue mode and then I have 3 possibilities:
Skip
Readonly
This is good. This will sort out LVM and such. It's not something I do
manually often enough to coach someone else over the 'net.
I don't have LVM enabled :)
When you get to the commandline, several commands will show you what's
mounted.
Normal=readwrite?
Which possibility should I choose?
I always chose readwrite, but with this one fsck went wrong!
Of course. You can't fsck a file that's mounted rw.
Well, he asks for mounting my /-partition. I have /boot on another partition.
So it
should be possible to fsck it.
I do not know exactly what I did different, but now I was able to fsck the /boot
->filesystem was fine.
But I didn' chroot in. May that have been the failure?
I thought I said to fsck before chroot.
Yes, you did, but this were the steps I did before reading your mail :)
The steps are
fsck
mount -o remount,rw ...
chroot
What I said.
I would then mount the system rw (still running the CD), chroot in, then
rebuilt the initrd.
This is what I just did. Before that I tried to extract the old initrd. This
failed with
errors, because of sudden end of file. Sounds as if we slowly come closer to the
problem ;)
That's what the original kernel message said.
So I created a new initrd with mkinitrd and once again tried to extract this.
Same
problem. I'm now trying to find a way to copy the initrd from the good machine
to
the broken one.
Did you follow my advice (below) on how to find the correct command to
use? if not, then what do you expect?:-)
Yes I did, the problem, why it was "broken" was, that I forgot to gunzip it
before extracting the CPIO-archive.
With gunzipping it before, only the original ramdisk was broken and all other
(the one I created via mkinitrd and the one I copied from the other machine)
were OK.
So this was my fault.
fwiw the initrd _can_ be an uncompressed CPIO archive. One time when I
needed to fiddle, I didn't bother recompressing it.
In the meantime I managed to start RHEL on the box (via external boot-loader
super-grub)
I think copying the /boot-partition from the second machine destroyed my MBR,
that’s why only cryptic signs appear instead of booting the system. So I will
now find a way to write a new fine MBR.
If you destroyed the MBR, most likely you wouldn't be able to see any
partitions, or they'd appear unformatted, or be the wrong size.
_How_ did you copy the other /boot?
One more little question:
In the grub.conf I find this line:
kernel /vmlinuz-2.6.18-53.el5 ro root=LABEL=/
If I start the kernel with this parameters via super-grub I run into a
kernel-panic again, but if I start
kernel /vmlinuz-2.6.18-53.el5 ro root=/dev/sda3
it works.
(this is, how I started the machine now)
When Anaconda creates (ext{2,3,4}) filesystems, it labels them. "/" is
the label applied to the root filesystem.
Going forward, I think it's using UUIDs instead, so asto confuse folks
even more.
Wherefrom does the kernel get its information, what LABEL=/ is?
Thanks a lot for your help and I hope that the machine is working in 30 minutes
again. :)
Best greetings.
Florian
OK, i'll even try this one.
Thanks for any advice.
Time is running out :(
Let it. Panic, and it's all over.
Don't overlook the possibility you have a hardware problem.
Yes I already asked the HP guys for a tool to check the raid and the disks, but
no
answer until now.
That's fairly important. Stir them up.
Florian
A command modeled on this will show an adequate command:
rpm -qf /boot/vmlinuz-2.6.9-34.EL --scripts
btw I'm not keen on copying all of /boot from another system. It might
be okay, I just don't know _in your circumstances._
--
Cheers
John
-- spambait
[EMAIL PROTECTED] [EMAIL PROTECTED]
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
