<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style></head>
<body class='hmmessage'><div dir='ltr'>
<br><br><div><div id="ecxSkyDrivePlaceholder"></div>> >> It actually seems worse under 0.25. Now commflag jobs are failing or<br>> >> deciding half way through a show that they're finished. The attached log<br>> >> is<br>> >> basically for 2 commflag jobs that I have been trying to run tonight. One<br>> >> failed so I re-queued it. Then both came back as "finished" but the shows<br>> >> were still recording.<br>> >> I then re-queued both again under mythbackend oto see what would happen.<br>> >> Both went a little longer, but then decided they were done at about the<br>> >> 50% - 70% point.<br>> > It appears that jobs running from the SBE that run in real time are deciding<br>> > they're done when they catch up to the actual recording in progress. This<br>> > has happened consistently now with every recording from tonight over several<br>> > re-queues, under mythjobqueu or mythbackend. Here's the releveant logs from<br>> > two of them...<br>> ><br>> > --SNIP--<br>> Do the commflag jobs report "Failed with exit status 140"?<br>> I have gotten this randomly on commflag jobs since upgraded to 0.25 (on <br>> Ubuntu 10.04).<br>> Hasnt concerned me that much, as commflagging seems to be pretty random <br>> in my part of the world anyway.<br>> (ie, some recordings its almost perfect, other recordings it will detect <br>> 0 commercials, even when the commflag job runs successfully)<br><br></div><div>One of them reported <span style="font-size:10pt">"Failed with exit status 140" the first time it ran. The rest (about 3 or 4 tries of the two included in those logs) simple reported "Finished 1 Commercial Break Found (then 2, then 3, etc)" when the caught up to the recording in progress. This only seemed to happen for jobs being run from mythjobqueue or mythbackend on the SBE. Once the recordings finished, I re-queued again and then ran through the whole program with no problem from the SBE, so it appears to be an issue with commflags running from a SBE in real time that catch up to the recording in progress. Any jobs that ran from the MBE performed as expected and simply slowed down and kept step with the recording in progress when they caught up to it.</span></div><div><span style="font-size:10pt"><br></span></div><div><span style="font-size:10pt">At this point, I have moved all job tasks the MBE and shut down the SBE completely. However, I am more than happy to fire it back up to do jobs to help figure it all out.</span></div> </div></body>
</html>