[mythtv-users] OT: Kernel errors

[Anthony Giggins]

On 18 April 2012 10:43, Michael Watson <michael at thewatsonfamily.id.au>wrote:

> On 18/04/2012 10:35 AM, Frank Phillips wrote:
>
>> On Tue, Apr 17, 2012 at 7:07 PM, Jerry Rubinow <jerrymr at gmail.com<mailto:
>> jerrymr at gmail.com>> wrote:
>>
>>    On Tue, Apr 17, 2012 at 9:02 AM, Mark Lord <mythtv at rtr.ca
>>    <mailto:mythtv at rtr.ca>> wrote:
>>
>>        On 12-04-16 11:36 PM, Jerry Rubinow wrote:
>>        > My mythbackend computer locked up today, the whole computer,
>>        not just mythbackend.  Following is the
>>        > syslog at the time this happened.  It's been rock steady for
>>        months, and I haven't made any changes
>>        > lately.  I rebooted and now it seems to be functioning normally.
>>        >
>>        > Any suggestions for what I should do?  Is this a sign of
>>        disk errors?
>>        ..
>>
>>        No, not with the scanty information provided.
>>        If there are disk errors, then there will be kernel logs along
>>        with them.
>>
>>        Also, "smartctl -a /dev/sdX" will give very good information
>>        about the error state of the drives.
>>
>>
>>    Thanks Mark.  kern.log had the same info as syslog, and smartctl
>>    wasn't revealing.  Sorry for the scanty info, but I'm not sure
>>    what direction to look.
>>
>>    Once more piece of data is that I saw a very high load before it
>>    completely locked up, but the top items in top were not using much
>>    cpu.
>>
>>
>>  That high load is caused by IO wait, which you can see in top as %wa.
>> The longer the disk takes to complete a task, the more processes backup in
>> the queue, causing a high load to be reported. Look closely at your disk,
>> as it most likely has issues.
>>
>>  I find "smart" almost useless with physical problems on hard drives.  I
> find hddtemp in combination with a logging function like cacti or mrtg to
> show up physical hard drives better.  (The temperature tends to go high
> when the disk is having physical issues)
>
> Try a 'cat /var/log/syslog | grep sdX'.  Might reveal something.  You
> might want to go through each of your hard drives (sda, sdb, etc) if you
> have more than one.
>
>
>
> Hi Michael,

OT but can you provide your cacti scripts your using for hddtemp?

Cheers,

Anthony
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mythtv.org/pipermail/mythtv-users/attachments/20120418/1326963c/attachment.html>