Since the arrival og dvb-t in Denmark, I've been playing with mythtv - with mythbackend on my qnap ts-119 (1200mhz, 512mb ram - installed debian on it), inspired by Andreas Back Aaen running mythbackend on a 400mhz Buffalo linkstation.
Unfortunately I decided to go with v0.22 (the latest), and it seems it has several issues with the dvb-t signals in this region :(
Subtitles delayed on tv2, mpeg4 streams not shown correctly, or not seekable etc. (see http://forum.recordere.dk/forum_posts.asp?TID=69424&PN=7&title=mythtv-02... for details on some of the problems and possible solutions).
Some have reported that v0.21+fixes with a few patches works like a charm.. So yes - I'm considering going back a version.. :)
One issue I've had is that mythbackend simply stops doing its job (recording).. but the backend log still logs as if it's actually doing something(except recording).. I haven't had the time to setup a new backend box (my current one is too low powered to run with gdb attached) so I can't genereate a backtrace unfortunately :(
I've reported the bug, but that's where it stands for now: http://svn.mythtv.org/trac/ticket/7647
If any of you suffer the same problem - feel free to help with information to the bug - and here's a script I wrote that restarts the backend when it freezes - I run it via cron every 5 minutes.
http://vsen.dk/files/check_mythtv.sh (remember to change paths to storage folder, and mythweb url).
The script checks status using mythweb (I've found that it doesn't answer when this problem arises - or it answers with errors) and then ensures that the backend isn't actually recording (writing anything to the storage folder).
It works rather well, as the issue is detected everytime within 5 minutes - so it saves my bacon :)
Sun, 01/31/2010 - 05:55
Fri, 02/12/2010 - 15:05
I set my backend to NOT run comflag while recording.
My backend is a 1200mhz arm (a qnap ts-119) running Debian with only 512mb ram.
I'm suspecting mythtv of having some sort of (temp) memory waste, while recording - which I think is the source of my problems.
I'm trying to find the time to create a new setup, on a dual-atom box with plenty of memory instead and see if the problems still continue.