Hello,
I am unable to mount my ext4 root drive.
I get this error:
EXT4-fs: ext4_check_descriptors: Block bitmap for group 2688
not in group (block 4205711663)!<3>EXT4-fs: group descriptors
corrupted!
I had a similar error some time ago, but I thought it was power
related. I fixed the problem by running a fcsk.
I have since purchased a ups so I can at least rule that out.
Also, after I fixed the error I never saw any ata errors or
kernel oops in dmesg.
I am running Fedora 10 with kernel
2.6.27.12-170.2.5.fc10.x86_64.
Root drive is on a lvm.
The physical disk is a 1.5 T.B Seagate (non-affected bad
firmware drive).
Any tips to debug this?
Jimmy Rentz
On Tue, Feb 03, 2009 at 07:15:31PM -0500, Jimmy Rentz wrote:
> I am unable to mount my ext4 root drive.
>
> I get this error:
> EXT4-fs: ext4_check_descriptors: Block bitmap for group 2688
> not in group (block 4205711663)!<3>EXT4-fs: group descriptors
> corrupted!
>
> I had a similar error some time ago, but I thought it was power
> related. I fixed the problem by running a fcsk.
You'll need to fix it using fsck again. I assume the back up
superblocks will have the correct value.
It would be interesting to see if there are any other errors picked up
by e2fsck.
> Any tips to debug this?
Can you reliably replicate this problem? I've not seen any similar
reports of problems like this.
- Ted
On Wed, 4 Feb 2009 01:24:00 -0500
Theodore Tso <[email protected]> wrote:
> On Tue, Feb 03, 2009 at 07:15:31PM -0500, Jimmy Rentz wrote:
> > I am unable to mount my ext4 root drive.
> >
> > I get this error:
> > EXT4-fs: ext4_check_descriptors: Block bitmap for group 2688
> > not in group (block 4205711663)!<3>EXT4-fs: group
> > descriptors corrupted!
> >
> > I had a similar error some time ago, but I thought it was
> > power related. I fixed the problem by running a fcsk.
>
> You'll need to fix it using fsck again. I assume the back up
> superblocks will have the correct value.
>
> It would be interesting to see if there are any other errors picked up
> by e2fsck.
I reran fsck and that fixed the problem.
Here is a sample of the errors:
Free blocks count wrong for group #1 (14705, counted=19112).
Fix? no
Free blocks count wrong for group #2 (12753, counted=11404).
Fix? no
Free blocks count wrong for group #3 (20531, counted=17213).
Fix? no
...
Free blocks count wrong for group #11160 (7673, counted=3577).
Then some errors about inode and directory counts wrong.
I also forgot that I have a log of the fsck errors when I
had this problem before.
>
> > Any tips to debug this?
>
> Can you reliably replicate this problem? I've not seen any similar
> reports of problems like this.
>
> - Ted
I am not sure how to reproduce this problem. It seems random
enough. The system is a mythtv box so it happens to
poweron/poweroff on a regular basis to record.
Jimmy Rentz wrote:
> I am not sure how to reproduce this problem. It seems random
> enough. The system is a mythtv box so it happens to
> poweron/poweroff on a regular basis to record.
Just to be clear, it does this in a controlled manner, right.
-Eric
On Wed, 04 Feb 2009 10:18:52 -0600
Eric Sandeen <[email protected]> wrote:
> Jimmy Rentz wrote:
>
> > I am not sure how to reproduce this problem. It seems random
> > enough. The system is a mythtv box so it happens to
> > poweron/poweroff on a regular basis to record.
>
> Just to be clear, it does this in a controlled manner, right.
>
> -Eric
You mean whether the poweron/poweroff part is done cleanly?
The system sets a wakeup time to poweron (to
record) then calls poweroff (via a script).
Jimmy Rentz wrote:
> On Wed, 04 Feb 2009 10:18:52 -0600
> Eric Sandeen <[email protected]> wrote:
>
>> Jimmy Rentz wrote:
>>
>>> I am not sure how to reproduce this problem. It seems random
>>> enough. The system is a mythtv box so it happens to
>>> poweron/poweroff on a regular basis to record.
>> Just to be clear, it does this in a controlled manner, right.
>>
>> -Eric
> You mean whether the poweron/poweroff part is done cleanly?
> The system sets a wakeup time to poweron (to
> record) then calls poweroff (via a script).
Right. I just wanted to make clear to anyone who hasn't used mythtv
that it can cleanly shut down & restart; I didn't want "poweroff" to be
mistaken for "power fail" :)
-Eric