[mythtv-users] Mythbackend aborting while many mythbackend --generate-preview processes

[John P Poet]

On Fri, Nov 20, 2009 at 10:16 PM, Bob Cottingham <bobnvic at gmail.com> wrote:
> On Fri, Nov 20, 2009 at 11:03 PM, John P Poet <jppoet at gmail.com> wrote:
>> On Fri, Nov 20, 2009 at 9:26 PM, Bob Cottingham <bobnvic at gmail.com> wrote:
>>> My backend just aborted and my wife isn't happy because it didn't
>>> record Medium.  There were 14 mythbackend --generate-preview 0x0
>>> processes running.
>>
>> I can't help with a fix, but I can with a work-around.  Add a cron
>> entry to kill the "run away" processes.  I have mine set to run every
>> 7 minutes:
>>
>> */7 * * * * pkill -f generate-preview
>
> So you are having this issue as well?  Did you submit a ticket for it?
>  I found a few emails on the mailing list with people having the issue
> but no real discussion or resolution.  I will give this a try as it is
> currently extremely disruptive - my slave backend will only run for a
> couple hours at most and my master backend stop commflagging after a
> couple days and occasionally goes so far as to abort.  I really don't
> understand why it is so much worse for the slave backend.

Actually, I had completely forgotten about the issue until you brought
it up.  I have had that in my crontab for many months.  I have not
tried taking it out of my crontab to see if I am actually still having
the problem.  I guess I should.....  But only when nothing critical is
scheduled to record ;-)

I assumed the problem had to do with the version of ffmpeg myth was
using when it first started happening.  I did not create a ticket,
because I knew that a new ffmpeg synce was scheduled before 0.22 was
to be released.  I am guessing by your email, that the new ffmpeg sync
did not fix the problem.

John
-- 
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?