When installing CentOS - sometimes the RAID-1 /boot partition, usually
/dev/md0 fails to boot.
The bug is known and exists on the bugtracker for CentOS as well as RedHat.
The fix is to re-install GRUB on on each partition of the RAID-1 array.
I think you could use the same method to answer your question.
1. boot server w/ any 'Disk-1' of an installation set
at boot prompt type 'linux rescue'
Continue boot.
2. First double check your disks with the mdadm utility
# mdadm -QD /dev/md0
Make a note of every disk (you can ignore the spare)
example: /dev/{sda1, sdb1, sdc1, sdd1, sde1}
3. drop into the GRUB prompt
# grub
grub> device (hd0) /dev/sda
grub> root (hd0,0)
grub> setup (hd0)
<repeat for each drive in the array>
grub> device (hd0) /dev/sdb
grub> root (hd0,0)
grub> setup (hd0)
Rinse - Repeat, etc.
4. exit from GRUB and reboot.
If you then want to test it, disconnect one of your drives - or just
drop into grub at boot
and tell it to boot from the partition of another drive.
-Peter
2008/6/9 Bowie Bailey <Bowie_Bailey at buc.com>:
> I have two drives in a software mirror. Other than setting the bios to
> boot from the second drive, is there any way to confirm that grub is
> installed properly on the second drive?
>
> --
> Bowie
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> http://lists.centos.org/mailman/listinfo/centos
>