[mythtv] EITScanner caused playback issues

Tom Lichti tom at redpepperracing.com
Wed Jul 19 18:59:20 UTC 2006


Daniel Kristjansson wrote:
> On Wed, 2006-07-19 at 14:42 -0400, Tony Lill wrote:
>   
>> Janne Grunau <janne-mythtv at grunau.be> writes:
>>
>>     
>>> On Tuesday 18 July 2006 23:14, Jan Kleinig wrote:
>>>       
>>>> I have some problems with the EIT-Scanner at the moment, depending
>>>> stumbling tv- or recording-playback when the EIT-scanner inserts his
>>>> data into the database. Would it help me building mythtv-eit-branch?
>>>>         
>>> Currently not. there are no changes in the eit branch. 
>>>
>>> Can you give us specifications of your backend/frontend computer.
>>>
>>> We should maybe decrease the Chunksize and increase the minimal 
>>> reschedule time for the realease.
>>>       
>> I submitted a patch for ticket #1660 that creates an asynchronus
>> database insertion thread. I used it to solve a similar problem with
>> database inserts hosing recording from ivtv. If you're just inserting
>> stuff for use later, it should solve your problem too.
>>     
>
> There is an asynchronous DB insertion thread for EIT data,
> and buffering which takes care of the DB inserts in the DVB
> recorder (buffering is needed for ivtv too, since DB inserts
> aren't the only thing that can hose an ivtv recording).
>
> The problem is probably with the rescheduling of recordings
> which can eat up a lot of resources; there is no way to avoid
> the reschedule hit, it can only be made less frequent and risk
> not recording a program you want to watch.
>
> How often does the system stumble when EIT is running?
>
> -- Daniel
>   
Slightly OT, but if the DB is on a machine that is different from the 
one doing the recording (and is not a MythBackend) would that pretty 
much eliminate and DB related problems with recording?

Thanks
Tom



More information about the mythtv-dev mailing list