[mythtv-users] Problem with commflag queuing on slave backends. 0.21

[Ryan Pisani]

> List --
>
> I seem to have an interesting issue with NFS based recordings and job
> queuing happening in 0.21. I have searched gossamer but I have not found
> any relevant information for the issue I'm seeing.
>
> Here goes:
>
> I have a pretty simple setup, with an MBE that runs 24/7 that holds the
> master NFS volume for all recordings to go to. I then have SBE/frontend
> that comes on at wake-up periods of the day to check for recordings and do
> it's jobs, etc.
>
> Since upgrading to 0.21 I'm seeing a repetitive issue where a slave
> backend comes online and it ends up queuing a mythcommflag job (always as
> user job #4) for a recording that was recorded on the MBE and had already
> completed commflagging.
>
> The problem presented is that this job just sits in the queue indefinitely
> and my shutdown script never gets a green-light to shutdown as
> mythshutdown --check returns a non-zero status.
>
> The more interesting part is that I have the flag set in the settings
> database to "commflag on original recording host only". Yet the MBE still
> queues up this job for the SBE.
>
> So in the job outlined below. I manually powered on the SBE at ~ 1:30AM
> and the commflag job in question was completed for nearly 2 hours.
>
> Relevant Lines from MBE log:
>
>
> 2008-03-21 22:00:53.871 JobQueue: Commercial Flagging Starting for World's
> Most Amazing Videos "Nature's Fury" recorded from channel 1121 at Fri Mar
> 21 21:00:00 2008
> 2008-03-21 22:00:53.954 commflag: Commercial Flagging Starting: World's
> Most Amazing Videos "Nature's Fury" recorded from channel 1121 at Fri Mar
> 21 21:00:00 2008
> 2008-03-21 22:01:31.209 scheduler: Finished recording: World's Most
> Amazing Videos "Nature's Fury": channel 1121
> 2008-03-21 22:01:31.751 Finished recording World's Most Amazing Videos
> "Nature's Fury": channel 1121
>
>
> Relevant entries in the jobqueue table: Notice the "queued" problem job
> was scheduled before the first one finished... not sure what explains
> that.
>
> mysql> select * from jobqueue where chanid="1121";
> +------+--------+---------------------+---------------------+------+------+-------+--------+---------------------+----------+------+-----------------------+---------------------+
> | id   | chanid | starttime           | inserttime          | type | cmds
> | flags | status | statustime          | hostname | args | comment
>       | schedruntime        |
> +------+--------+---------------------+---------------------+------+------+-------+--------+---------------------+----------+------+-----------------------+---------------------+
> | 4525 |   1121 | 2008-03-21 21:00:00 | 2008-03-21 20:59:32 |    2 |    0
> |     2 |    272 | 2008-03-21 22:35:28 | frouse2  |      | 6 commercial
> break(s) | 2008-03-21 20:59:32 |
> | 4528 |   1121 | 2008-03-21 21:00:00 | 2008-03-21 22:01:31 | 2048 |    0
> |     0 |      1 | 2008-03-21 22:01:31 | frouse2  |      |
>       | 2008-03-21 22:01:31 |
>
>
>
> Any suggestions?  Please note this started being a problem after the move
> to 0.21 and was not exhibited in 0.20.2.
>
> thanks
> Ryan
>

Ignore this. It turns out that "user job 4" was checked in the "GENERAL"
setup page under TV Settings. This apparently overrides the "Only commflag
on recording host" setting in MythTV-Setup screens.