Re: Bad interactive behaviour in 2.5.65-66 (sched.c)

[Mike Galbraith <efault@gmx.de>]

At 07:06 AM 3/31/2003 +1000, Con Kolivas wrote:
>On Mon, 31 Mar 2003 00:14, Jens Axboe wrote:
> > On Sat, Mar 29 2003, Robert Love wrote:
> > > On Sat, 2003-03-29 at 21:33, Con Kolivas wrote:
> > > > Are you sure this should be called a bug? Basically X is an interactive
> > > > process. If it now is "interactive for a priority -10 process" then it
> > > > should be hogging the cpu time no? The priority -10 was a workaround
> > > > for lack of interactivity estimation on the old scheduler.
> > >
> > > Well, I do not necessarily think that renicing X is the problem.  Just
> > > an idea.
> >
> > I see the exact same behaviour here (systems appears fine, cpu intensive
> > app running, attempting to start anything _new_ stalls for ages), and I
> > definitely don't play X renice tricks.
> >
> > It basically made 2.5 unusable here, waiting minutes for an ls to even
> > start displaying _anything_ is totally unacceptable.
>
>I guess I should have trusted my own benchmark that was showing this was 
>worse
>for system responsiveness.

I don't think it's really bad for system responsiveness.  I think the 
problem is just that the sample is too small.  The proof is that simply 
doing sleep_time %= HZ cures most of my woes.  WRT contest and it's 
io_load, applying even the tiniest percentage of a timeslice penalty per 
activation and no other limits _dramatically_  affects the benchmark 
numbers.  (try it and you'll see. I posted a [ugly but useful for 
experimentation] patch which allows you to set these things and/or disable 
them from /proc/sys/sched)

I'm trying something right now that I think might work.  I set 
MAX_SLEEP_AVG to 10*60*HZ , start init out at max, and never allow it to 
degrade.  Everyone else is subject to boost and degradation, with the 
maximum boost being MAX_SLEEP_AVG/20 (which is still a good long sleep, and 
the max that one sleep can boost you is one priority).  When you start a 
cpu hogging task, it should drop in priority just fine, and rapid context 
switchers shouldn't gain such an advantage.  We'll see.  Tricky part is 
setting CHILD_PENALTY to the right number such that fork()->fork() kind of 
tasks don't drop down too low and have to crawl back up.  Contest falls 
into this category.

Anyway, I think that inverting the problem might cure most of the symptoms ;-)

         -Mike