Mounted partition is suddenly read only

[Steve Flynn]

On 18 December 2012 17:26, JD <jd1008 at gmail.com> wrote:


> Yes Steve, it is one of those drives.
> The problem is, only ONE mounted partition out of 4 mounted partitions
> of the drive got remounted read only. The other ones remained read/write.
> So, it is not the case that the drive went to sleep.
>
> Aye - that's what made me ask the question - only one partition going on
strike rules out the entire drive dozing off.



> I really should have looked at syslog. My bad!!
> Syslog says:
>
> syslog.1:Dec 17 19:12:34 cme kernel: [160907.464516] Aborting journal on
> device sdb3.
> syslog.1:Dec 17 19:12:34 cme kernel: [160907.464611] EXT3-fs (sdb3): error
> in ext3_reserve_inode_write: Journal has aborted
> syslog.1:Dec 17 19:12:34 cme kernel: [160907.464617] EXT3-fs (sdb3):
> error: remounting filesystem read-only
> syslog.1:Dec 17 21:43:58 cme kernel: [169991.763211] EXT3-fs (sdb3):
> error: ext3_remount: Abort forced by user
> syslog.1:Dec 17 21:59:22 cme kernel: [170915.743954] EXT3-fs (sdb3):
> error: ext3_remount: Abort forced by user
> syslog.1:Dec 17 22:01:25 cme kernel: [171039.429121] EXT3-fs (sdb3):
> error: ext3_remount: Abort forced by user
> syslog.1:Dec 17 22:29:57 cme kernel: [172750.931413] EXT3-fs (sdb3):
> error: ext3_remount: Abort forced by user
> syslog.1:Dec 17 22:44:59 cme kernel: [173653.286490] EXT3-fs (sdb3):
> error: ext3_put_super: Couldn't clean up the journal
> syslog.1:Dec 17 22:49:54 cme kernel: [173948.443870] EXT3-fs (sdb3): using
> internal journal
> syslog.1:Dec 17 22:49:54 cme kernel: [173948.443877] EXT3-fs (sdb3):
> mounted filesystem with ordered data mode
>
> And so I wonder what caused
>
EXT3-fs (sdb3): error in ext3_reserve_inode_write: Journal has aborted ?
>
> Guess I will be googling this for a while :)
>
>
At the risk of stating the obvious (and this is just an educated guess on
my part - I've no googled any of this as I'm pretty sure you'll be well
ahead of the game here) it looks like the filesystem was trying to grab an
inode to extend the journal and it failed. Might be something as simple as
a previous dirty unmount of the drive (I think you said these were
external) had left this part of the disk a bit creaky.

I'd be running that filesystem through badblocks to try to map out the
dying part of the drive, if that's what's going wrong repeatedly.

Anything in the SMART stats of the drive indicating there's an issue with
the physical medium ("smartctl" should be able to dump the stats for you -
don't think it's installed by default)

-- 
Steve

When one person suffers from a delusion it is insanity. When many people
suffer from a delusion it is called religion.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-users/attachments/20121218/86962527/attachment.html>