[mythtv-users] On the verge of giving up

[osma ahvenlampi]

On 9/27/07, Greg Estabrooks <greg at phaze.org> wrote:
> > That's what they say. My backend/frontend combo does it all the time,
> > though. Perhaps it is the menu action, perhaps it's OTA EPG updates
> > from the DVB channels -- in any event it's a major drag. I would have
>
>  Actually that would cause BFSQ to happen more often. Since it's
> constantly updating the program info and those updates need to be considered
> for scheduling purposes. I don't know off hand how often
> it triggers scheduling updates though. If it's more than every few minutes
> then I'd say something is wrong . It certainly should not occur every few
> seconds.

MySQL slow query log tells me the median interval is around 2-3
minutes, but sometimes it is triggered 20 seconds after the last
execution. Each time it takes 10-20 seconds to execute.. So that's
about 10% of total runtime. Every query examines two million rows and
returns nearly 8000. A bit excessive, I would say.

> > "ancient" in this case would be the svn-from-three-months-ago, ie
> > about 10x younger than the latest actual release. Sorry, cheap shot,
> > but I couldn't resist.
>
>  How is that a cheap shot? Maybe my sarcasm meter is still broken :)

Well, I guess the last release wasn't really 30 months ago even if it
does feel like that. I'm just used to projects really working way
better if the releases are time boxed to very short intervals - in my
professional life I try to aim for 1 month, two at the outset, but I'm
also really impressed by the progress the GNOME folks make by making a
polished release every six months.

>  If something is constantly triggering the BFSQ every few seconds
>  you would certainly have some issues. You should gather some logs from
> the backend showing EIT updates and howoften the BFSQ is launching and
> submit it to trac.

Last couple of times I submitted anything to trac without a patch
included, I was shot down pretty fast.. and the couple of patches I've
submitted were ignored despite early positive feedback. I realize the
devs are probably just as busy with their other lives as I'm with
mine, but that a good motivation to try to ignore these problems and
think about something else :(

-- 
Osma Ahvenlampi   <oa at iki.fi>       http://www.fishpool.org