Hi,

A couple of linodes face a weird ext4 issue randomly, as seen in the log below:

blkfront: barrier: empty write xvdc op failed
blkfront: xvdc: barrier or flush: disabled
end_request: I/O error, dev xvdc, sector 46399544
end_request: I/O error, dev xvdc, sector 46399544
Aborting journal on device xvdc-8.


Jun 09 22:11:39 [kernel] EXT4-fs error (device xvdc): ext4_journal_start_sb:327: Detected aborted journal
Jun 09 22:11:39 [kernel] EXT4-fs (xvdc): Remounting filesystem read-only
Jun 10 01:31:57 [kernel] EXT4-fs error (device xvdc): ext4_put_super:818: Couldn't clean up the journal

The servers are running official linode kernel 3.0.18-x86_64-linode38

I tried switching to the 3.2.1 kernel and a custom kernel as well, but no use. It develops the fault a few moments after booting.

And according to this thread: viewtopic.php?t=8259 and the logs posted above, barriers are already disabled.

How to fix this? Also, it really doesn't make sense for our I/O needs to use ext3, it would be a *big* performance penalty.

What's in your /etc/fstab?

_________________
Paid support
How to ask for help
1. Give details of your problem
2. Post any errors
3. Post relevant logs.
4. Don't hide details i.e. your domain, it just makes things harder
5. Be polite or you'll be eaten by a grue

It turned out to be a problem with that specific machine. After hours of debugging, when the bug wasn't found, we deleted the machine. As such one was extra in our LB, so that solved. None of the others have/had the problem.