strange one that i've seen twice now.

mythcommflag starts and runs forever. in mythweb it says "building head start buffer". TOP shows it running at near 190% CPU (2x) been running for hours.

I also can't kill it. pkill mythcommflag does nothing.

This is the second time in two days this has happend, can't find anything in BE logs. NOTE I'm using the re-nice babysitter script and have been for(ever).

Last time a reboot 'fixed' it for a day.

Thoughts?

Corrupted recording....

I've had this happen a few times before. The reason rebooting "clears" it is that the file is left marked "in use by comm flagging" so the next time it runs and scans for new files it bypasses the borked ones. If you clear the status in the DB or use the --force flag or the like, the current instance of comm flagging will loop forever on the exact same recordings.

The only real solution I've found is to delete the bad recordings. You'll generally find that these will either hang or kill transcoding too.

OK thanks. So a corrupt recording happens 2 days in a row, 2 different tuners, 2 different sources. I'm hoping this isn't the begining of the HDD problem...

Errors do tend to cluster. Often because there is common causal factor in effect. Last time I saw this was terribly windy, rainy, miserable weather, before that when they were doing cable work in my neighborhood.... YMMV.

OK thanks. We have been getting weather the last week. I'll keep watch...