[mythtv-users] commercial flagging and transcode jobs dying

[jafa]

jafa wrote:
>> Hi guys,
>>
>> Yesterday I noticed my machine was getting behind on the job queue... it 
>> was working on a transcoding job but very slowly. A quick check showed 
>> that the mythcommflag was using 80% of the CPU even though there were no 
>> active commercial jobs (according to mythtv system status). BTW - I have 
>> the cpu option for the job queue set to high (nice = 0) and limited to 
>> one job at a time.
>>
>> I set up another mythtv backend to provide some more CPU power so it 
>> could catch up.... This morning I noticed that a transcoding job on the 
>> second backend was stuck at 50% progress. The mythtranscode process was 
>> not running. No obvious error messages.
>>
>> I did an svn update, recompiled, and now it is transcoding the same job 
>> again. It will be interesting to see if it has the same problem in the 
>> same place.
>>   
>>     
> Both good and bad... the problem is repeatable :-)
>
> Mythtranscode died in the same place... 1 hour into the 2 hour program.
>
> Mythbackend shows:
>
> 2006-07-03 09:56:50.592 XDS: Current Program Length 2:00 Time in Show 
> 0:59.00
> 2006-07-03 09:57:33.627 XDS: Current Program Length 2:00 Time in Show
> 1:00.00
> 2006-07-03 09:58:06.048 TFW: safe_write(): funky usleep
>
> The recording plays ok in mythtv.
>
> Nick
>   
Ok - found the problem with transcoding - There was a 2GB file size 
limit for write imposed by samba as I had not enabled large file support 
(lfs).

Hopefully the svn update will have fixed the commflag problem.

Nick