[prev in list] [next in list] [prev in thread] [next in thread] 

List:       mythtv-dev
Subject:    [mythtv] Master back-end bug report (And front-end as well...)
From:       Rudy Zijlstra <mythtv () edsons ! demon ! nl>
Date:       2004-04-29 22:17:50
Message-ID: 40917F0E.5080805 () edsons ! demon ! nl
[Download RAW message or body]

Hi all,

I've likely found a rather repeatable bug.

Config:
Master back-end with encoder 1 (PVR-250) and sql server and (nfs 
exported) file storage for mythtv
EPIA M6000 slave backend with encoder 2 (PVR-350) and front-end 
connected to TV using TV-out of PVR.

Situation: slave backend already running
start master back-end
slave will connect to master well.
reboot slave backend
after reboot slave re-connects
start using myth to browse through recorded programs
and observe master backend crashing with: mutex destroy failure, 
resource busy or in use.

Logs from the master backend:
root@repeater:~# Starting up as the master server.
2004-04-29 16:05:48 mythbackend version: 0.15.20040331-1 www.mythtv.org
2004-04-29 16:05:48 Enabled verbose msgs : important general
2004-04-29 16:05:48 adding: tv-server as a slave backend server
2004-04-29 16:05:50 Found changes in the todo list.
2004-04-29 16:05:50 Scheduled 18 items in 0.319817 seconds.
2004-04-29 16:05:50 Seem to be woken up by USER
2004-04-29 16:09:56 MainServer::HandleAnnounce Playback
2004-04-29 16:09:56 adding: tv-server as a client (events: 0)
2004-04-29 16:09:57 MainServer::HandleAnnounce Playback
2004-04-29 16:09:57 adding: tv-server as a client (events: 1)
2004-04-29 16:13:08 adding: tv-server as a slave backend server
2004-04-29 16:13:08 Found changes in the todo list.
2004-04-29 16:13:09 Scheduled 18 items in 0.322945 seconds.
2004-04-29 16:20:32 MainServer::HandleAnnounce Playback
2004-04-29 16:20:32 adding: tv-server as a client (events: 0)
2004-04-29 16:20:33 MainServer::HandleAnnounce Playback
2004-04-29 16:20:33 adding: tv-server as a client (events: 1)
2004-04-29 16:20:44 Slave backend: tv-server has left the building
Mutex destroy failure: Device or resource busy


As you can see, its CVS of march 31.

Related bug: the frontend waits forever on the crashed backend and does 
not return. At least not within 5 minutes, which is how long my patience 
was.

Questions:
- Has any code gone into CVS to solve something like this?
- Can somebody point me to a description of the protocol used? I'd like 
to take a look at it, C++ is not something i easily read... And time i 
tend to be short on.
  With a current description I might come up with possible improvements.

Regards,

Rudy
_______________________________________________
mythtv-dev mailing list
mythtv-dev@mythtv.org
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
[prev in list] [next in list] [prev in thread] [next in thread]