Block bitmap differences
 

Hi,

a few weeks ago, an unhealthy combination of firmware in an Adaptec Raid
controller and Seagate disks damaged my Raid6 filesystem. A bunch of
files were damaged or lost at that time after the firmaware was updated
and I had run e2fsck. Luckily, I was able to restore everything from a
backup. A subsequent check with e2fsck reported no errors.


Yesterday, I ran e2fsck -n again, to see if the system is still OK. It
isn't and I have no idea how to interpret the messages (see attachment).


What is the meaning and severity of

- Block bitmap differences?
- Free blocks count wrong for group?


Thanks and regards.

Roland


receiving file list ... done
data/886125/

sent 26 bytes received 166603904 bytes 63820.70 bytes/sec
total size is 1156104247261 speedup is 6939.24
e2fsck 1.40.8 (13-Mar-2008)
Warning! /dev/sdb1 is mounted.
Warning: skipping journal recovery because doing a read-only filesystem check.
/dev/sdb1 contains a file system with errors, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
Block bitmap differences: -(729567354--729567775) -(729630796--729630799) -(729632768--729633174) -(1875509507--1875509511) +(1875775082--1875775086)
Fix? no

Free blocks count wrong for group #22260 (969, counted=0).
Fix? no

Free blocks count wrong for group #22264 (544, counted=0).
Fix? no

Free blocks count wrong for group #22266 (2794, counted=2383).
Fix? no

Free blocks count wrong for group #57236 (5, counted=0).
Fix? no

Free blocks count wrong for group #57244 (14, counted=19).
Fix? no

Free blocks count wrong (1316741037, counted=1316739108).
Fix? no

Inode bitmap differences: +468729888 -468730043
Fix? no


/dev/sdb1: ********** WARNING: Filesystem still has errors **********

/dev/sdb1: 32026820/536870912 files (0.5% non-contiguous), 830719896/2147460933 blocks
_______________________________________________
Ext3-users mailing list
Ext3-users@redhat.com
https://www.redhat.com/mailman/listinfo/ext3-users

Block bitmap differences
 

On Thu, Oct 23, 2008 at 10:00:44AM +0200, Roland Bock wrote:
> Hi,
>
> a few weeks ago, an unhealthy combination of firmware in an Adaptec Raid
> controller and Seagate disks damaged my Raid6 filesystem. A bunch of
> files were damaged or lost at that time after the firmaware was updated
> and I had run e2fsck. Luckily, I was able to restore everything from a
> backup. A subsequent check with e2fsck reported no errors.
>
> Yesterday, I ran e2fsck -n again, to see if the system is still OK. It
> isn't and I have no idea how to interpret the messages (see attachment).

You ran the e2fsck while the filesystem is mounted. So the output
reported is not trustworthy, and block allocation bitmap differences
and free block/inode accounting information being wrong is normal when
running e2fsck -n on a mounted filesystem.

This message, however, is cause for concern:

> /dev/sdb1 contains a file system with errors, check forced.

This means the filesystem noticed some discrepancy (for example, when
freeing a block, it noticed that the block bitmap already showed the
block as being not in use, which should never happen and indicates
filesystem corruption).

I would recommend that you schedule downtime so you can run e2fsck on
the filesystem while it is unmounted. Given the errors that you saw
when running e2fsck while it was mounted, it's unlikely that you will
see anything serious, but it is still something that you should do.

Regards,

- Ted

_______________________________________________
Ext3-users mailing list
Ext3-users@redhat.com
https://www.redhat.com/mailman/listinfo/ext3-users

Block bitmap differences
 

Ted,

thank you for your answers.

Is it normal to encounter file systems with minor errors? We run 8
systems with Ubuntu 8.04 64bit and e2fsck reports "<device> contains
file system with errors" for at least one partition on every machine.


Since there are 4 different types of hardware configurations, I tend to
say that hardware is rather not to be blamed...


If it is not normal, what could be the reasons?
Are there any options to turn on logging which could give more insight
(what would be the performance impact)?



Thanks and regards,

Roland


Theodore Tso wrote:

On Thu, Oct 23, 2008 at 10:00:44AM +0200, Roland Bock wrote:

Hi,

a few weeks ago, an unhealthy combination of firmware in an Adaptec Raid
controller and Seagate disks damaged my Raid6 filesystem. A bunch of
files were damaged or lost at that time after the firmaware was updated
and I had run e2fsck. Luckily, I was able to restore everything from a
backup. A subsequent check with e2fsck reported no errors.


Yesterday, I ran e2fsck -n again, to see if the system is still OK. It
isn't and I have no idea how to interpret the messages (see attachment).


You ran the e2fsck while the filesystem is mounted. So the output
reported is not trustworthy, and block allocation bitmap differences
and free block/inode accounting information being wrong is normal when
running e2fsck -n on a mounted filesystem.


This message, however, is cause for concern:


/dev/sdb1 contains a file system with errors, check forced.


This means the filesystem noticed some discrepancy (for example, when
freeing a block, it noticed that the block bitmap already showed the
block as being not in use, which should never happen and indicates
filesystem corruption).

I would recommend that you schedule downtime so you can run e2fsck on
the filesystem while it is unmounted. Given the errors that you saw
when running e2fsck while it was mounted, it's unlikely that you will
see anything serious, but it is still something that you should do.

Regards,

- Ted


_______________________________________________
Ext3-users mailing list
Ext3-users@redhat.com
https://www.redhat.com/mailman/listinfo/ext3-users

Block bitmap differences
 

Roland Bock wrote:
> Ted,
>
> thank you for your answers.
>
> Is it normal to encounter file systems with minor errors? We run 8
> systems with Ubuntu 8.04 64bit and e2fsck reports "<device> contains
> file system with errors" for at least one partition on every machine.
>
> Since there are 4 different types of hardware configurations, I tend to
> say that hardware is rather not to be blamed...
>
> If it is not normal, what could be the reasons?

Look in your system logs; if the fs is flagged with errors, it should
have issued a message when the error occurred.

-Eric

> Are there any options to turn on logging which could give more insight
> (what would be the performance impact)?
>
>
> Thanks and regards,
>
> Roland


_______________________________________________
Ext3-users mailing list
Ext3-users@redhat.com
https://www.redhat.com/mailman/listinfo/ext3-users

Block bitmap differences
 

Eric,

what should I be looking for? In /var/log I grep'ed for ext and fs (case
insensitively) in all syslog, messages and kern.log files. I found
nothing which indicated an error to me. Just occasional mount/umount
messages and the like.


Well, to be exact: I did find some error messages from the time when we
had hardware issues on one machine. But nothing since these were
resolved two weeks ago. e2fsck was happy then.



Thanks and regards,

Roland

Eric Sandeen wrote:

Roland Bock wrote:

Ted,

thank you for your answers.

Is it normal to encounter file systems with minor errors? We run 8
systems with Ubuntu 8.04 64bit and e2fsck reports "<device> contains
file system with errors" for at least one partition on every machine.


Since there are 4 different types of hardware configurations, I tend to
say that hardware is rather not to be blamed...


If it is not normal, what could be the reasons?


Look in your system logs; if the fs is flagged with errors, it should
have issued a message when the error occurred.

-Eric

Are there any options to turn on logging which could give more insight
(what would be the performance impact)?



Thanks and regards,

Roland





_______________________________________________
Ext3-users mailing list
Ext3-users@redhat.com
https://www.redhat.com/mailman/listinfo/ext3-users