lvm over raid confusion

Dan grinnz at gmail.com
Mon Mar 27 19:29:47 UTC 2006


Jack Tanner wrote:
> Paul Howarth <paul <at> city-fan.org> writes:
>
>   
>> Try it again. I rebuilt the initrd using the --noresume option of mkinitrd.
>>     
>
> OK, chapter 2.
> - Boot from rescue CD.
> - After it asks for keyboard layout and network, but before it scans for Linux
> partitions, Ctrl-Alt-F2 and do # mdadm -AR /dev/md0 /dev/hdb /dev/hdc. It runs
> /dev/md0.
> - Switch back to rescue installer, and have it scan for Linux partitions. It
> finds the install!
> - I can now chroot /mnt/sysimage and build my own initrd. It builds in the RAID
> drivers, and device mapper!
> - Reboot with the new initrd. It loads the RAID module, and seems to correctly
> detect /dev/md0, but it still does NOT find the volume group!!! Same output:
> Couldn't find device with uuid ...
> Couldn't find all physical volumes for volume group VolGroup00
>
> I tried again and generated an initrd with --noresume, but that doesn't change
> things at all.
>
> Argh!
>
> Maybe I can separate the VGs so that the RAID array isn't part of VolGroup00.
>
>   
For future reference, you can also ignore its request to find your 
installation, mkdir /mnt/sysimage, and mount your root partition there 
yourself, and your /boot partition and whatever else you need to work 
with. I did this all the time on my FC4 installation, since it was on 
NVRAID and i needed dmraid loaded before I could access it; I'd mount a 
partition simply containing the binary /sbin/dmraid (or in my case, my 
backup of my root partition) in /mnt/sysimage, then run dmraid -ay, 
unmount it, and then mount whatever NVRAID partitions I wanted to work on.
-Dan




More information about the fedora-list mailing list