[mythtv-commits] Ticket #9123: frontend and backend get stuck after live tv channel change using HD-PVR
MythTV
mythtv at cvs.mythtv.org
Thu Oct 21 07:33:37 UTC 2010
#9123: frontend and backend get stuck after live tv channel change using HD-PVR
----------------------------------------------------------+-----------------
Reporter: Johnny Stenback <mythtv-users@…> | Owner: cpinkham
Type: defect | Status: infoneeded_new
Priority: minor | Milestone: unknown
Component: MythTV - General | Version: Trunk Head
Severity: low | Resolution:
Keywords: | Ticket locked: 0
----------------------------------------------------------+-----------------
Comment (by Johnny Stenback <mythtv-users@…>):
Ok, trying to answer most questions here.
The hardware of the mbe is a 2.93GHz Intel Core2 duo with 4 gigs of ram an
plenty of disks.
The hardware of the frontends (both identical) is a Zotac IONITX based
system with a single core Intel Atom 230 and, Nvidia ION graphics.
The network is gigabit ethernet all the way.
I did a bunch more digging today, and while the behavior has changed to
the point where I don't consistently see the lockup any more, or have not
been able to reproduce it today at all, I still see issues which look like
they're rooted in the same underlying problem.
What I'm seeing now (both on the standalone frontend and on the combined
sbe and frontend) is that live tv starts out fine, but after the first
channel change the osd never goes away, it remains open after the video
starts playing and it's saying that the signal lelvel is 0%, and some
other value flashes by at a pretty high frequency. This matches what I see
in the logs, when this happens I see a flood of "MythEvent: SIGNAL 1"
messages in the logs. And what's interesting is that these messages keep
going in the logs even after I stop watching live tv.
My theory so far is that we somehow fail to exit the loop in
SignalMonitor::MonitorLoop(), if that's the case I would expect the
behavior I'm seeing, more or less. Per the logs, SignalMonitor::Stop() is
called, but if for some reason the member "running" is not set at that
point we would fail to stop the monitor loop. I'm working on a way to
prove this, but if anyone has other ideas I'm all ears.
I'll attach the logs from both the sbe and the mbe, the mbe happened to be
running with -v general,important,network,record,channel, and the sbe with
-v important,general,extra,network,channel,record.
Oh, and yes, always stream from master backend is on here, I had it off
earlier, but that seemed to have made it impossible to watch any live tv
on my standalone frontend (I kept getting "Error opening jump program file
buffer", and then dropped back to the main menu), but that's another issue
for another ticket that I'll investigate more later if it persists.
--
Ticket URL: <http://svn.mythtv.org/trac/ticket/9123#comment:8>
MythTV <http://www.mythtv.org/>
MythTV Media Center
More information about the mythtv-commits
mailing list