| 1 |
On Sunday 25 July 2010 09:18:33 Dale wrote: |
| 2 |
> Alan McKinnon wrote: |
| 3 |
> > On Sunday 25 July 2010 06:57:43 KH wrote: |
| 4 |
> >>> You said you ran e2fsck and it was OK. What was the command? |
| 5 |
> >>> |
| 6 |
> >>> |
| 7 |
> >>> |
| 8 |
> >>> Normally with an e2fsck on a journalled fs, the app will replay the |
| 9 |
> >>> journal and make a few minor checks. This takes about 4 seconds, not |
| 10 |
> >>> the 40 minutes it takes to do a ful ext2 check. |
| 11 |
> >>> |
| 12 |
> >>> |
| 13 |
> >>> |
| 14 |
> >>> I think you might need to fsck without the journal. I know there's a |
| 15 |
> >>> way to do this but a cursory glance at the man page didn't reveal it. |
| 16 |
> >>> Maybe an ext user will chip in with the correct method |
| 17 |
> >> |
| 18 |
> >> Hi, |
| 19 |
> >> |
| 20 |
> >> I ran on the two partitions e2fsck /dev/sde3 as well as fsck.ext3 |
| 21 |
> >> /dev/sde3 . Yes, it only took some seconds. |
| 22 |
> > |
| 23 |
> > It's been a long time since I used ext3 so some of this might be wrong. |
| 24 |
> > |
| 25 |
> > An fsck that takes a few seconds is using the journal, which might not |
| 26 |
> > uncover deeper corruption. You should try disabling the journal (I |
| 27 |
> > couldn't find the way to do that though), but this will also work: |
| 28 |
> > |
| 29 |
> > Boot of a LiveCD, mount your root partition somewhere using type "ext2" |
| 30 |
> > and fsck it. This will invalidate the journal but that's OK, it gets |
| 31 |
> > recreated on the next proper boot. Let the fsck finish - it will take a |
| 32 |
> > while on a large fs. |
| 33 |
> > |
| 34 |
> > When done, reboot as normal and see if the machine boots up properly. |
| 35 |
> |
| 36 |
> And I would stand guard to make sure housekeeping doesn't come around. |
| 37 |
> ;-) Cutting power during all this wold not be good. |
| 38 |
|
| 39 |
KH, I think that this may not be related to a fs error as such. |
| 40 |
|
| 41 |
Yes, pulling the plug may have caused fs corruption. However, more likely is |
| 42 |
that pulling the plug did not allow you to do something that you should have |
| 43 |
done after you finished upgrading to grub-0.97-r9. The latest installation of |
| 44 |
grub asks you to reinstall in the MBR and point its root to wherever your |
| 45 |
/boot is. GRUB's fs and its drivers may have changed and therefore the old |
| 46 |
boot loader code is looking for files that no longer exist. |
| 47 |
|
| 48 |
So you'll probably be alright again if you boot with a fresh systemrescue |
| 49 |
LiveCD and run grub and then root (hd....) and setup (hd0) before you quit and |
| 50 |
reboot. |
| 51 |
|
| 52 |
If that doesn't work then you most likely have a fs problem. |
| 53 |
|
| 54 |
HTH. |
| 55 |
-- |
| 56 |
Regards, |
| 57 |
Mick |
Archives