[mythtv-users] A plea for help

[Paul King]

Dan Harper wrote:
> One thing I've had success with also to reproduce problems is to  
> compile a kernel on the box.  This seems to give a good thrashing on  
> memory and disk.  I've had memory problem identified when compiling a  
> kernel that memtest86 didn't pick up.
>
> The give away for RAM is if you compile the kernel twice and crashes  
> occur in different locations of the source tree (ie: some level of  
> randomness).
>
> Dan
>
>
> On 24/09/2007, at 11:46 AM, Brett Kosinski wrote:
>
>   
>> That's all I can think of for now.  Please, anyone who has any idea  
>> how I can troubleshoot this or give me any ideas, let me know.  My  
>> wife gets very upset when her soap-operas don't record,  and with  
>> the new fall-season shows coming up (Heroes!),  I'd hate to miss  
>> any of those either :-)   Thanks!
>>
>> You didn't mention the filesystem, disk layout (LVM, RAID, etc),  
>> kernel version, etc, you were using.  I can say that, in the past,  
>> I've had kernel oopses specifically because of XFS, which occurred  
>> during times of high disk activity (bittorrent was a great way to  
>> trigger the problem).
>>
>> In addition, it sounds like this is a combo FE/BE.  You might try  
>> shutting down X and letting the backend run for a while.  This will  
>> allow you to see any kernel oopses that are occuring, as they will  
>> be dumped the console (just make sure to disable console  
>> blanking).  Of course, this means your Myth system will be  
>> unavailable for use during that time, but you could run a test like  
>> this over a few nights, until a crash triggers (assuming one does).
>>
>> Brett.
>>
>>
>> _______________________________________________
>> 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
>
>   
Hi

I would also point to memory. I have found memtest86 less than reliable. 
I had a frontend on Linuxmce which would crash when loading mythrtv and 
when launching a KDE desktop at slightly different times on each. I ran 
memtest and it came back saying everything was ok. I replaced the memory 
anyway as I was still convinced this was the problem and everything runs 
ok. I also work as a computer service engineer and from experience with 
lots of machines randon rashes 99% of the time seem to come down to 
memory problems.

With the price of memory these days its worth just grabbing a new stick 
to try.

Regards