Unbootable after filesystem corruption

Bug #675416 reported by Jeroen T. Vermeulen
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Ubuntu
Confirmed
Undecided
Unassigned

Bug Description

I recently had an incident with my ext4 filesystem corruping itself and going read-only. (It does that sometimes, since lucid or so, but that's a separate bug). I had two of those incidents over the past few days. The first lost me some data but otherwise seemed to leave the system in a workable state; the second left me with an unbootable system.

None of the installed kernels (one from Lucid, one from Maverick, and a 2.6.36 mainline one) would boot. It looked as if I was thrown into busybox, though there were lots of output and no prompt. The output looked like a list of available devices.

After a manual fsck from a live CD the system boots fine again. Interestingly the fsck didn't even fix much; there were perhaps a few dozen lines of output.

Revision history for this message
Fabio Marconi (fabiomarconi) wrote :

Hello Jeroen
seems something hardware related, but can you please attach here
/var/log/syslog
before and after the fsck ?

Changed in ubuntu:
status: New → Incomplete
Revision history for this message
Jeroen T. Vermeulen (jtv) wrote :
Revision history for this message
Jeroen T. Vermeulen (jtv) wrote :
Revision history for this message
Jeroen T. Vermeulen (jtv) wrote :

Hi Fabio,

Thanks for looking into this. Definitely some weird stuff in the log before things went bad.

Revision history for this message
Fabio Marconi (fabiomarconi) wrote :

Hello Jeroen
In syslog 'befor' we have just much errors related to bluetooth:
Nov 11 20:01:45 salamis kernel: [20167.757764] hci_scodata_packet: hci0 SCO packet for unknown connection handle 0
Nov 11 20:01:45 salamis kernel: [20167.757772] hci_scodata_packet: hci0 SCO packet for unknown connection handle 0
Nov 11 20:01:45 salamis kernel: [20167.757779] hci_scodata_packet: hci0 SCO packet for unknown connection handle 0
Nov 11 20:01:45 salamis kernel: [20167.757787] hci_scodata_packet: hci0 SCO packet for unknown connection handle 0
But all this problems, in my opinion, means that there's something much serious, then make a backup of yours important data.
Please start the "Memory check" from the boot menu. If that runs successfully, please start the rescue system in the boot menu and select the "File system check" option in the next menu. If the file system corruption is serious, the safest option would be to reinstall your system.
If not solved probably the HD is in a bad state.
Sorry for that
Fabio

Revision history for this message
Jeroen T. Vermeulen (jtv) wrote :

Thanks. I've been seeing regular filesystem corruption for what must be close to a year now so I ran the full memory test at one point. It showed no problems whatsoever. Also, another user (also with nVidia and ext4 IIRC) ran into very similar symptoms which made me suspect the driver at the time. I switched to nouveau and, coincidence perhaps, didn't seem to have the problem for a while.

Is the "File system check" option different from the manual fsck I did as reported above? IIRC I usually run it as e2fsck -v -f <device> (but I may have used plain fsck as well).

Revision history for this message
Jeroen T. Vermeulen (jtv) wrote :

I'm having the problem again, though in a more serious form that leaves Maverick completely unbootable, and a Maverick live CD won't let me fsck the partition. But I took the laptop back to the shop for a full hardware check-up, and it showed no problems. This includes a hard-disk check and yet another memory check.

Changed in ubuntu:
status: Incomplete → Confirmed
Revision history for this message
Jeroen T. Vermeulen (jtv) wrote :

By the way, bug 688541 shows that it's not at all unthinkable that this could be an OS problem.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.