[mythtv-users] mythbackend crashes every 6-7 hours

[Greg Oliver]

On Sun, Oct 31, 2010 at 7:04 AM, Greg Oliver <oliver.greg at gmail.com> wrote:
> On Sun, Oct 31, 2010 at 6:57 AM, Rob Verduijn <rob.verduijn at gmail.com> wrote:
>> Hello,
>>
>> Hunting for very obscure ram memmory flaws sounds like a ghost chase to me.
>> So I'll classify that one under gut feelings and assumptions and
>> therefore won't waste time on it.
>> (Even if I had any spare ram, which I don't)
>>
>> But the valgrind suggestion sounds like a more reasoning aproach to me.
>> Do you have a link to a howto, because I'm not really familiar with valgrind.
>> I will get on it as soon as I figure out what valgrind is and how to use it.

Like was said though, it is highly unlikely a software issue since
nothing has changed.  It does not have to be bad memory as you know
for a hardware issue to cause this.  After valgrind, I would suggest
what others have - a stress test of sorts..  It really sounds like
something is failing in the machine.

> Here's a quick start.  You need to have the mythtv debug symbols
> packages installed prior though.
>
> http://valgrind.org/docs/manual/quick-start.html#quick-start.prepare
>
> There were quite a few results returned with google "mythtv + valgrind" as well.
>
> -Greg
>
>> Rob Verduijn
>>
>> 2010/10/31 Paul Harrison <mythtv at sky.com>:
>>> On 31/10/10 08:35, Rob Verduijn wrote:
>>>> Hi there,
>>>>
>>>> I've spend some time awake when the folks we're sleeping.
>>>> Ran the memory stress test and guess what.......ZERO!!! ERRORS
>>>> Over one hour wasted.
>>>>
>>>> The physical ram is OK.
>>>>
>>>> I will perform no more hardware related tests that are based on
>>>> assumptions or gut feelings.
>>>>
>>>> Sorry about that I get cranky when I do not get enough sleep.
>>>>
>>>> I'm not running trunk builds.
>>>> I'm running packages from the packman repository.
>>>> http://ftp.gwdg.de/pub/linux/misc/packman/suse/11.3
>>>>
>>>> Rob Verduijn
>>>>
>>>>
>>>>
>>>
>>> Running a memory stress test for only one hour is unlikely to find the
>>> fault, I've run stress tests for 24 hours with no errors but stress the
>>> memory using a real world application like a large compile and it
>>> randomly fails every time. Often the only way to be sure is to swap out
>>> the memory sticks one by one until you find the faulty one.
>>>
>>> If you are sure your hardware is not faulty and the problem is due to a
>>> leak then the best way to get it fixed is to run the backend under
>>> valgrid under the same conditions that causes the problem and post the logs.
>>>
>>> Paul H.
>>> _______________________________________________
>>> mythtv-users mailing list
>>> mythtv-users at mythtv.org
>>> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>>>
>> _______________________________________________
>> mythtv-users mailing list
>> mythtv-users at mythtv.org
>> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>>
>