[mythtv-users] Scheduling new recordings requires mythtv restart

[Joel]

Michael T. Dean wrote:
> On 04/22/2009 06:18 PM, Joel wrote:
>>  I have been running mythfilldatabase out of cron.daily, rather than
>>  having the backend call it.  If it's better to have the backend run
>>  it for some reason, I can switch to that method.
> 
> I recommend switching and letting the backend call it.

Ok.  I switched to letting the backend call it.

> The errors you're getting are due to the socket errors.  A lot of people 
> seem to get them.  I don't know that anyone knows any specific reasons 
> why some people get them a lot and others don't.
> 
>> 2009-04-22 11:48:43.601 MythSocket(8187d90:-1): writeStringList: 
>> Error, called with unconnected socket.
>> 2009-04-22 11:48:43.601 MythSocket(8187d90:-1): readStringList: Error, 
>> called with unconnected socket.
> 
> Those socket errors are preventing mythfilldatabase from telling 
> mythbackend to run a reschedule (thus the error you quoted).  I'd ask 
> other users of your distro if they've seen similar socket errors and 
> whether they've found a way to fix them.

I'm building from svn regularly on slackware 12.1.  I doubt there's that
large of a userbase on slackware, but if anyone has solved this please
tell me how.  My only configure options are --disable-distcc and
--enable-proc-opt.  Processor is a core 2 quad Q6600.

 From the logs it looks like it's making a connection and negotiating
the protocol version, then failing when it tried to read or write
additional data.  Any thoughts on how to continue debugging why the
socket is failing and how to fix it?

 From the backend logs, I see the following, which is interesting
because it looks like mythfilldatabase segfaulted this time:

2009-04-22 23:06:47.174 Running mythfilldatabase
2009-04-22 23:11:48.276 mythfilldatabase still running, skipping checks.
2009-04-22 23:16:38.878 MainServer::HandleAnnounce Monitor
2009-04-22 23:16:45.880 MainServer::HandleAnnounce Monitor
2009-04-22 23:16:51.508 adding: zion as a client (events: 0)
2009-04-22 23:16:51.565 adding: zion as a client (events: 1)
2009-04-22 23:16:51.583 MythSocket(8430550:-1): writeStringList: Error,
socket went unconnected.
2009-04-22 23:16:51.587 MainServer::HandleAnnounce Monitor
2009-04-22 23:16:51.607 adding: zion as a client (events: 0)
2009-04-22 23:16:53.283 mythfilldatabase still running, skipping checks.
2009-04-22 23:17:21.620 MainServer::HandleAnnounce Monitor
2009-04-22 23:17:21.626 adding: zion as a client (events: 0)
2009-04-22 23:17:21.628 Received a remote 'Clear Cache' request
sh: line 1:  4945 Segmentation fault
/usr/local/bin/mythfilldatabase -v most
>>/usr/local/var/log/mythtv/mythfilldatabase.log 2>&1

The mythfilldatabase logs look about the same except they stop at the
clear cache request right before the segfault appears to have occurred.
  Log is here if it is of additional interest:
http://ebel.mybox.org/mythfilldatabase2.log

> 
> It's also possible that mythfilldatabase running in the same environment 
> as your mythbackend will fix the issue--another good reason to let 
> mythbackend run mythfilldatabase.

This doesn't seem to be the case.  My last run was called by mythbackend
and seems to have done no better.

Thanks again,
Joel