<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Mon, Apr 15, 2013 at 10:22 AM, Rich West <span dir="ltr"><<a href="mailto:Rich.West@wesmo.com" target="_blank">Rich.West@wesmo.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Since upgrading to 0.26 and getting the system stabilized, all of my<br>
commercial flagging jobs are failing with "Failed with exit status<br>
140". I've done some googling which suggests that something is killing<br>
off the commflag process, but I cannot find any evidence of that.<br>
<br>
The system is a brand new install of Fedora 17 with a 0.26 db upgrade<br>
over the years. It has plenty of memory and disk space. Because<br>
everything is being encoded by external boards (hdhomerun, pvr150's),<br>
the CPU is rather bored, too. The DB tables are all good (nothing<br>
needing to be repaired).<br>
<br>
In looking at the logs:<br>
<br>
1. a majority of the mythcommflag*.log files stop at "Finding Logo" with<br>
no further information.<br>
<br>
2. he remaining ones (only 14 so far), go as far as "CoreContext<br>
ClassicCommDetector.cpp:536 (go) - 90% Completed @ 52.179 fps." and then<br>
generate:<br>
Decoder avformatdecoder.cpp:4341 (ProcessAudioPacket) - AFD: Unknown<br>
audio decoding error<br>
Decoder avformatdecoder.cpp:4556 (GetFrame) - decoding error<br>
eno: Unknown error 541478725 (541478725)<br>
CoreContext main.cpp:941 (FlagCommercials) - Finished, 2 break(s) found.<br>
<br>
And the web interface lists all of the jobs with the 140 error..<br>
<br>
Any ideas as to where to start digging for this one?<br>
<br>
Thanks!<br>
-Rich<br>
_______________________________________________<br>
mythtv-users mailing list<br>
<a href="mailto:mythtv-users@mythtv.org">mythtv-users@mythtv.org</a><br>
<a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a><br>
</blockquote></div><br></div><div class="gmail_extra">I think there are some ffmpeg bugs that are causing the majority of problems with commflagging in 0.25+. Additionally, there is a race condition which causes the commflagging job to abort when it hits end-of-file while operating on a recording in progress. The workaround for this is to change the setting which prevents commflagging jobs from running until the recording is finished. For me, commercial flagging was near-perfect in 0.24, but is quite broken since then. I'm lucky if it finds one commercial break per show now.<br>
<br>Karl<br></div></div>