Hi,

Yesterday, I rebooted my CentOS 7 laptop after couple of days and it got updated automatically. I tried to take a backup image of the root (/) partition with clonezilla, but it failed with the error message "extfsclone.c bitmap error at group 1".

I did lots of research and tried the following options.

1. Tried using the option fsck-chk-part as part of clonezilla backup, but got the same error.
2. Tried fsck.ext4 /dev/sda6 in partedmagic and got an error message saying "e2fsck is old. Please update to new version".
3. Tried booting with latest elementary OS live cd and ran the same command, but it did not report any errors.
4. Opened up gparted and it shows an exclamation mark (!) at /dev/sda6 and when I double click it, it shows "/dev/sda6 has a bad superblock". Gparted shows the fstype as ext4.
5. I booted with hiren boot cd and tried partition recovery and it shows /dev/sda6 as "unformatted".

I saw couple of guides related to fixing the superblock and I will try them today.

But, my question is can it be fixed by simply formatting the /dev/sda6 partition?

Thanks,
Balaji.

If there is nothing wrong with your hardware then you can fix the bad superblock problem by formatting the partition. However if you format the partition then you will lose all of your data on the partition.

Before you do that I suggest that you try running fsck against an alternate superbock. Here is an explanation of how to do that:

http://www.cyberciti.biz/faq/recover...ted-partition/

-----------------------
Steve Stites

Thanks for your reply. I tried this with the backup at 32768. It appeared to have fixed something, but when I run clonezilla, I still got the bitmap error. I ran the same e2fsck and tried the backup super block from 32768. the command runs, I get the final output as "file system was modified" though clonezilla is still reporting the same error.

I mounted /home on /dev/sda7 where most of my data resides. So, if formatting /dev/sda6 will resolve the issue, I will try and reinstall linux.

I tried the imaging with todobackup, and it is reporting that /dev/sda6 is unformatted. I did a surface test and it came out clean.

are there any other HDD diagnostic tool that I can use to confirm that it is not a hardware issue?

Also, should I try the backup super block from different location, other than 32768?

I suggest that you try gparted again and see whether it still thinks that the first superblock is bad.

------------------------
Steve Stites

No, Gparted is not showing any errors. Even running sudo fsck.ext4 -f -v -y /dev/sda6 is not throwing any errors. But, clonezilla is continuing to report the bitmap error.

Thanks
Balaji.

If everything works with the file system except clonezilla then there is likely a bug in clonezilla. If that is the case then I suggest you report the clonezilla bug to the clonezilla developers and try using a different backup program. I recommend that you use rsync or one of the GUI programs which interface with rsync to do your backups.

------------------------
Steve Stites

It is not only clonezilla, I tried two other imaging softwares and they are reporting that the root partition is unformatted.

Thanks,
Balaji.

Previous kernel(s) boot without any problems?

Yes, the CentOS update added two new boot entries, one with debugging. I tried the new entry without debugging and did not boot. The previous kernel entry boots fine with out any issues.

Thanks,
Balaji.

Ummm.. The documentation for clonezilla has one warning: "The partition to be imaged or cloned has to be unmounted." (reference: http://clonezilla.org/)

It is rather unlikely that the root partition is unmounted (it might accept single user mode with root mounted read-only though).

Any time a filesystem is actually mounted, and active, it will have bitmaps modified in memory, and possibly only partially updated on disk. This always leads to imaging problems as the resulting copy is corrupted due to the active modification of the partition.