[SOLVED] Aborted journal and volume remounted read-only??

rjo98 · 06-04-2010, 05:46 PM

I have errors like this in my messages file, and nobody can write to the volume its talking about. i'm assuming i need to reboot and do an e2fsck but what options would you recommend? -fyc i'm thinking but not sure. thanks for the help!!

journal_bmap: journal block not found at offset 5132 on hdb5
Aborting journal on device hdb5.
__journal_remove_journal_head: freeing b_committed_data
journal commit I/O error
ext3_abort called.
EXT3-fs error (device hdb5): ext3_journal_start_sb: Detected aborted journal
Remounting filesystem read-only

alunduil · 06-04-2010, 05:52 PM

Actually, what you need to do is remove the journal from the filesystem and then do the e2fsck -f. Then readd the journal and reboot. Of course to do all this you need the filesystem unmounted.

Code:

tune2sf -O ^has_journal /dev/<dev>
e2fsck -f /dev/<dev>
tune2sf -j /dev/<dev>

Regards,

Alunduil

rweaver · 06-04-2010, 05:54 PM

What you're going to need to do is reboot to single user mode, make sure drive is unmounted and then fsck it. I would suggest -fn first run to see what it says... after that if necessary -fc and if there are tons of errors -fcy and you know you want to answer yes on everything... first thing first though, I would be get back up off that drive NOW.

rjo98 · 06-04-2010, 05:55 PM

Quote:

Originally Posted by alunduil

Actually, what you need to do is remove the journal from the filesystem and then do the e2fsck -f. Then readd the journal and reboot. Of course to do all this you need the filesystem unmounted.

Code:

tune2sf -O ^has_journal /dev/<dev>
e2fsck -f /dev/<dev>
tune2sf -j /dev/<dev>

Regards,

Alunduil

Does removing the journal make me lose all my data or no? If it will just remove it then it can be re-added without data loss I'd be ok with that.

rjo98 · 06-04-2010, 05:56 PM

Quote:

Originally Posted by rweaver

What you're going to need to do is reboot to single user mode, make sure drive is unmounted and then fsck it. I would suggest -fn first run to see what it says... after that if necessary -fc and if there are tons of errors -fcy and you know you want to answer yes on everything... first thing first though, I would be get back up off that drive NOW.

luckily (i guess) this is my backup server, so its backups of all my other servers. I have a -fcy running now, figure might as well check it all since i have some time. but what about the bad journal thing?

alunduil · 06-04-2010, 07:35 PM

Yes, the journal can be added and removed from ext3 at will.

Regards,

Alunduil

rjo98 · 06-04-2010, 07:51 PM

So with those messages i was getting, is it absolutely necessary, a good idea, or just possible that i will need to drop and recreate the journal? Also, how long does that process take roughly?

alunduil · 06-04-2010, 07:56 PM

Everything I've been told about dealing with an aborted journal says that you need to drop the journal, fsck, and re-add the journal. Simply get into an environment where you can unmount and work on the device in question and run the three commands that I suggested earlier.

Regards,

Alunduil

rjo98 · 06-04-2010, 07:59 PM

ok, thanks, i'll give it a shot after my e2fsck -fcy finishes.

alunduil · 06-04-2010, 08:39 PM

Here's to you not losing any files.

Regards,

Alunduil

rjo98 · 06-04-2010, 08:40 PM

Thanks, I got my fingers crossed!

rjo98 · 06-05-2010, 09:25 AM

So once my original e2fsck finishes, done without dropping the journal, do I still need to do another one once I drop the journal or can I just drop and create it skipping the e2fsck in the middle of those 2 commands?

alunduil · 06-05-2010, 11:13 AM

I'm not sure but I wouldn't take the risk of not doing it. Let us know how things go when it's all done.

Regards,

Alunduil

rjo98 · 06-05-2010, 11:28 AM

I just removed the journal and am now e2fsck'ing. Will let you know how it turns out. Thanks for answering!

rjo98 · 06-05-2010, 12:11 PM

Journal rebuilt, said volume was clean on startup, and I just successfully did a du -sh * on the entire volume, which I couldn't do before! I think I'm ok!
I do have a question though, now my lost+found is 17GB though. I think I saw it say it was moving stuff to there during the e2fsck. All the file names in there look like a number sign than an inode number maybe? What are all those files? Does this mean I lost 17GB of data?