On 06/03/14 18:47, Les Mikesell wrote:
> On Tue, Jun 3, 2014 at 7:09 AM, mark <m.roth@xxxxxxxxx> wrote:
>> I had an old server die; we had another, same model, sitting around lightly
>> used, so I did what I've done before: swap the RAID card into that, swap the
>> drives, even putting them in the same bays as the dead one, and boot.
>>
>> Nope. 100% of the time, when it hits switching roots, it kernel panics.
>>
>> Adding rdshell to the kernel line does nothing, I get no grub rdshell. Booting
>> from a flash drive into rescue, it finds everything, *perfectly*, and mounts
>> it all, including the RAID data drives.
>>
>> I've rebuilt the initrd, and no joy.
>>
>> Anyone have an idea?
>
> Did you try running "grub-install" after your rescue-mode boot and
> chroot into /mnt/sysimage? If that doesn't fix it there is probably
> something different about the device/naming of the root partition.
>
Put another drive in, and had all kinds of grief before, right before I left
yesterday, it FINALLY came up. After disconnecting the RAID (I may have wiped
one of the RAID devices - luckily this is a backup system). I'm guessing the
damn thing, every time I changed drives, reenumerated, ignoring my previous
settings, and deciding on its own what should should be the first drive.
Wonder if the battery needs replacing....
Of course I tried grub-install. At least two tries, after pxebooting to linux
rescue (and it would mount everything *perfectly*), I'd see root was
/dev/sda... but grub-install announced that sda had no BIOS entry....
mark
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
