extents_fl errors on ext3 partition

[Mad Professor]

My 1.5TB raid 5 storage array has ext3 journal partition.

I was using CentOS 5.2 but had to change the O/S to Opensuse 11.2 due to being unable to run the latest 5.5 due to my raid controller drivers breaking udev.

Anyways I remount the array in opensuse using fstab and uuid and everything is fine, so I shutdown and put the case covers back on and moved it back into the rack.

Now when I reboot I get dropped to maintenance mode saying /storage is not clean, run fsck without -a or -p options. Even tho it was clean on CentOS, I guess opensuse is freaking out about the ext3 journal data.


I ran fsck and saw something about Ext3 Filesystem being cleared running ext 2 filesystem check

 Inode *insertrandom number* has EXTENTS_FL flag set on filesystem without extents support. Clear?

I went about 5 minutes answering yes then aborted due to too many yes answers.

After searching google I don't see alot except that it's bad and expect dataloss.

Right now I mounted it as ext2 *because I can no longer mount as ext3* copying data to one of the drives from a broken raid 1 array. I have 1.1TB of data, I do not have any backups due to external drive I use to back it up being broken and being RMA'd to manf.

Data is not reproducible , We're talking about serious data loss.



Any ideas what extents_fl is? should I allow fsck to repair the filesystem?

[Mad Professor]

So This has been solved.

Basically I've backup my array before unleashing fsck -y on it. It removed all the journal inodes on the disc essentially dropping it to ext2, I then use tune2fs -j /dev/sdb1 which then recreated the journal and set data=journal in fstab. which brought it back up to ext3, I don't think I suffered any data loss but I didn't have time to verify the files.