[mythtv] [mythtv-commits] Ticket #6192: Long delays processing events

[Eric Bosch]

Strangely,  just recently had a segfault occur while LiveTV was running, 
but my bad, I didn't have the debugger up, so no backtrace.  I have it 
running now in GDB and will let you all know if it happens again and 
will send the backtrace.


Shane wrote:
> On Mon, Feb 9, 2009 at 5:38 PM, Chris Pinkham <cpinkham at bc2va.org> wrote:
>> * On Mon Feb 09, 2009 at 05:30:52PM -0500, Shane wrote:
>>> Yeah, I thought there was probably other reasons for the backend getting into
>>> the loop sending events to itself too.
>>>
>>> The one I looked into from a while back was caused (I think) by the mythbackend
>>> preview generation process on the backend inadvertently registering for events
>>> and if timing was right it would cause the loop.
>> My change won't prevent this if it was occurring.  My change only
>> prevents the master backend from connecting back to itself.  Only the
>> master should be relaying events, so if another process runs on the
>> master server and connects to the master mythbackend process it
>> shouldn't cause event loops.
> 
> Yeah, it was only an old theory as to the cause of the event relaying loop.
> 
> The IsMasterBackend() check you added at the top of ConnectToMasterServer() is
> similar to the debugging patch I posted. It will print a message in the same
> circumstances and hi-lite any other lurking instances.
> 
> I can't reproduce the problem and it only strikes here rarely, it
> seemed to occur more
> frequently for Eric. I'll keep a watch out for those messages in the
> logs, hopefully
> Eric will report back if he sees them too or the original problem returns.
> 
> Perhaps, a bit premature closing the ticket. It can be re-open if needed.
> 
> Shane
> _______________________________________________
> mythtv-dev mailing list
> mythtv-dev at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
>