Install grub on /dev/md0, CentOS 6

[cahalsall]

I have this very same setup (/boot partition in /dev/md0 [raid1]) on a Dell PowerEdge 830. I had to revert back to the (2.6.32-71.29.1.el6.i686) kernel after failing to boot from the new (2.6.32-220.2.1.el6.i686) kernel. The original installed operating system is Centos 6 32 bit.

This bug seems to have been reported already at: http://bugs.centos.org/view.php?id=5381

A recommendation was made to reinstall the kernel as a solution. I think I'm going to wait this one out, for now.

[toracat]

@cahalsall

Bug #5381 is not related to what is being discussed in this forum thread. In #5381, something went wrong when the OP attempted to update the system. It was an isolated incident and the OP was able to fix it by repairing grub.conf + reinstallation of the kernel.

The current problem has been seen by multiple people and it has to do with raid1 and a disabled/non-working device.

[toracat]

CentOS dev [b]Tru[/b] did thorough debugging of the problem [1] and reported it in the upstream bugzilla [2]. He also proposed a workaround:

Add an additionnal "rdshell" to the boot command line and manually activate the detected array.

[1] http://bugs.centos.org/view.php?id=5400
[2] https://bugzilla.redhat.com/show_bug.cgi?id=772926

[mtaa]

Hello,had the bug been fixed ? after i update from 6.0 to 6.1 and unplug the hd,i also get the error...thank you

[pschaff]

[b]mtaa[/b] - Welcome to the CentOS fora. Please see the recommended reading for new users linked in my signature.

After reading those links you should realize why you should not hijack threads as you have done; however, the short answer is [b]no[/b]. If there had been a fix it would be reflected in the bug reports linked above.

[EUGK]

hey, people!
I have the same problem with 6.2 on a virtualbox.
6.1 works on both disks, but 6.2(last kernel) only on sdb.

It's strange, but i think, i solved this problem...
i loaded with 6.1 on sda, and then:

just change in /etc/grub.conf, in that part, where newer kernel is written about:

root=uuid=blabla656585

to

root=/dev/md0

and it works. no kernel panic, no other [d]xxxx[/d] [Moderator edit] stuff.

and it have another string, about 6.1 with root=uuid=...blbla, and it works too
strings looking the same at their another parts

I really don't understand, why it works. but - it works. may be this helps to find a bug.

sorry about bad english.
i will be glad, if somone will teach me, why it works.)

[pschaff]

[b]EUGK[/b] - Welcome to the CentOS fora. Please see the recommended reading for new users linked in my signature.

Thanks for your description of what worked for you, but please omit the curse words, and for your question[s], then rather than hijacking this thread please start a new one to get the attention you need, providing a [url=https://www.centos.org/modules/newbb/viewtopic.php?post_id=155892&topic_id=34988&forum=55#forumpost155892]link[/url] to this one if required for context.

[TrevorH]

Actually the dracut bug was just fixed in the latest update (dracut-004-256.el6_2.1.noarch) so booting from a failed software RAID 1 should now be working again - once you rebuild the initrd affected of course.

[johann8]

yes, I can confirm this.
I rebuilt the initramfs:

mv /boot/initramfs-$(uname -r).img /boot/initramfs-$(uname -r).img.old
dracut --mdadmconf --force /boot/initramfs-$(uname -r).img $(uname -r)

It works!
Thank you

[abfarn]

i had similar problem with 6.3
what i did was :
1 boot to rescue mode
2 chroot /mnt/sysimage/
3 grub-install /dev/md0

then it works... :-D