2.6.24-16 won't boot after upgrade to Hardy

[Rob Quill]

Binary package hint: linux-image-2.6.24-16-generic

Yesterday I upgraded my PC running Gutsy to Hardy. Previously the kernel I'd been using was a custom-compiled version of 2.6.22, with a patch applied to solve a problem with SMART. After upgrading to Hardy and rebooting it boots up to a BusyBox. Upon further investigation the messages before it gets to the busy box were to the effect of:

mount: cannot mount /dev/disk/by-uuid/..... on /root: Device or resource busy.

The uuid is the same as the 2.6.22 UUID, so I don't think that is the problem. Doing:

dmsetup remove_all
mount /dev/sda2 /root

Works fine, as does mounting by UUID after dmsetup remove_all has been run. I was given this idea from a bug in (I believe) 2.6.22 where you needed to remove evms. So i did sudo apptitute purge evms, even though it wasn't installed and then did update-initramfs -u -k 2.6.24-16-generic and rebooted but the problem remains.

I should not that the 2.6.22 kernel boots fine and hardy works fine using it as far as I can tell.

Thanks for your help.

Rob

[Rob Quill (rob-quill)]

Just wondering if anyone had any idea how to fix this? I meant to say before that I am using an AMD64 processor.

Any help is much appreciated.

Rob

[Frodon (frodon)]

You should try adding "all_generic_ide" at the end of your hardy kernel entry in menu.lst, this has solved the issue for some users.

Not sure it will work in your case but at least you can try.

[Rob Quill (rob-quill)]

Thanks for your help Frodon. Unfortunately using "all_generic_ide" did not fix the issue, but changing the disk from /dev/disk/by-uuid/.... to /dev/mapper/nvidia_... in menu.lst resolved the problem :D

Thanks again,

Rob

[Leann Ogasawara (leannogasawara)]

Hi Rob,

Did you want to investigate this bug further or are you satisfied with the current solution you have. Let us know. Thanks.

[Rob Quill (rob-quill)]

Hi,

The problem hasn't reoccurred since I changed the entry in menu.lst, so I'm happy.

Rob

[Leann Ogasawara (leannogasawara)]

Thanks Rob. I'm just going to close this report for now. The "Won't Fix" status is a bit harsh I think in this scenario but we have limited states for bug reports. It would be better if we had a "Resolved" state which reflected you have a working solution bug we didn't modify any code on our end. Thanks.