[PLUG] Desktop failure returned

Denis Heidtmann denis.heidtmann at gmail.com
Thu Jul 16 20:56:39 UTC 2015


On Thu, Jul 16, 2015 at 10:45 AM, Denis Heidtmann <denis.heidtmann at gmail.com
> wrote:

>
>
> On Wed, Jul 15, 2015 at 10:49 PM, Dale Snell <ddsnell at frontier.com> wrote:
>
>> On Wed, 15 Jul 2015 22:05:26 -0700, in message
>> CAArUT0jKWejUygaOdt0CQVgXa40MVNc1O--MbkvGoPh4h=-6xw at mail.gmail.com,
>> Denis Heidtmann wrote:
>>
>> > After upgrading to Grub2 on the 13th the hope was that would fix
>> > things. Not so.
>> >
>> > This time the failure happened after it was running for 5-10 minutes,
>> > not on boot.  First symptoms were Chrome failed to start 3 times.
>> > Then Nautilus did not display properly; it closed when I attempted to
>> > view the root directory.  Then the desktop icons were big and
>> > spurious text appeared.  ^ alt bksp yielded:
>> > *Stopping save kernel messages
>> > speech dispatcher disabled;edit /etc/default/speech-dispatcher
>> > WARNING: All config files
>> > need.conf:/etc/modprobe.d/nvidea-current_hybrid.conf.hidden, it will
>> > be ignored in a future release.
>> > *Starting Virtual Box Kernel Modules
>> > *Starting Virtual Box Kernel Module...
>> > *Starting MD monitoring service mdadm--monitor
>> > saned disabled; edit /etc/default/saned  *checking battery state...
>> > [2033.461491] end_request: I/O error, dev sda, sector 606863455
>> > [2033.461491] EXT3-fs error(device: sda1): ext3-_get_inode_loc:unable
>> > to read inode block- inode=18964526,block=75857924
>> > ....8 more similar messages...
>> > ^ ALT DEL produced 3 more messages
>> > Each repeat of ^ ALT DEL yielded an identical message.
>> >
>> > Power off, then restart.  Now it is working fine, AFAICT.
>> >
>> > This is screaming disk problem, yet tests of the disk say it is
>> > fine.  What else could be failing erratically?
>>
>> The disk interface.  I assume that you're using a SATA interface on
>> the motherboard.  Have you tried a different SATA port?  If that
>> fails, you could buy a SATA interface card and plug the drive into it.
>>
>> You could still have an intermittent failure in your drive itself.
>> Frankly, I'd buy another drive and see if it fixes your problem.  If
>> not, well, it doesn't hurt to have a spare drive.
>>
>> Is the failure always located in the same sector/block/inode?  That
>> would point to the drive proper.  If the sectors are different every
>> time, then it's probably something else.  It could even be the power
>> supply for the computer.  Have you tried swapping it out?  (I don't
>> remember what you've done, other than re-load Grub.)
>>
>> Hope this helps.
>>
>> --Dale
>
>
> I have upgraded grub, run e2fsck and smart checks.  I have checked the
> fans.  Now to try  the cables.  Since my last communication I had another
> failure.  On restart I chose the usual Ubuntu.  This time it crashed w/
> "can't open file...Can't open root device "UUID= ...  Kernel panic.
>
> Power off/on; chose recovery mode.  I chose fsck .  This time it looked
> like fsck crashed:
> udevd[766] '/sbin/blkid -o udev -p /dev/sda1' [1434] terminated by signal
> 11(segmentation fault)
>
> But it is running fine now.  Gremlins.
>
> -Denis
>

Checking if failures refer to the same inodes.  Failures show up
differently, but in two instances the same failure messages appeared: "*
Stopping save kernel messages.", followed, some lines later, by:
"end_request: I/O error, dev sda, sector 606863455
EXT3-fs error (device sda1): ext3_get_idnode_loc: unable to read inode
block -- inode=18964521, block=75857924
end_request: ...
<more lines>..."

The numbers in the "end_request" lines for the two instances are
identical.  Subsequent lines were different.

How definitely does this pin the blame on the hard drive?

-Denis



More information about the PLUG mailing list