[27341] in Athena Bugs
Re: Athena 10 crash: dm-2 I/O errors during sustained high-volume
daemon@ATHENA.MIT.EDU (Evan Broder)
Sat Jul 25 13:55:34 2009
Message-ID: <4A6B4703.409@mit.edu>
Date: Sat, 25 Jul 2009 10:55:15 -0700
From: Evan Broder <broder@mit.edu>
MIME-Version: 1.0
To: Matthew Belmonte <belmonte@mit.edu>
In-Reply-To: <200907250701.n6P71Fvv023139@contents-vnder-pressvre.mit.edu>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-Spam-Flag: NO
X-Spam-Score: 0.00
Cc: bugs@mit.edu
Errors-To: bugs-bounces@mit.edu
Hi Matthew -
My inclination is to blame that sort of error on imminent hard drive
failure. As our main cluster testing machine, opus's drive has been
pretty heavily stressed - it's been reinstalled weekly, or sometimes
daily - so I'm not too surprised that it might be going.
Can you or somebody else in SIPB try checking SMART information on the
drive and other stuff like that?
- Evan
Matthew Belmonte wrote:
> This morning I was transferring a large volume of data from a USB disc to the
> local disc /dev/sda2 on opus.mit.edu.
>
> At 2.02am or thereabouts, after leaving the machine unattended for a few
> minutes, I came back and moved the mouse, whereupon gnome immediately crashed
> and caused me to be logged out. On several attempts to log in again, with my
> default session and then without customisations, I was immediately logged back
> out. After rebooting, I was able to log in normally.
>
> /var/log/messages (below) is peppered with messages about I/O errors on dm-2
> beginning around the time of the crash. dmsetup and lvs give the following
> details:
>
> root@opus:/afs/athena.mit.edu/user/b/e/belmonte# dmsetup ls
> athena-swap_1 (252, 4)
> athena-login (252, 0)
> athena-root (252, 1)
> athena-root-real (252, 2)
> athena-login-cow (252, 3)
> root@opus:/afs/athena.mit.edu/user/b/e/belmonte# lvs
> LV VG Attr LSize Origin Snap% Move Log Copy% Convert
> login athena swi-ao 10.00G root 0.22
> root athena owi-ao 111.64G
> swap_1 athena -wi-ao 9.37G
>
>
> Jul 24 07:43:19 opus syslogd 1.5.0#5ubuntu3: restart.
>
[snip]
