[Lumiera] Question about Scheduler capabilities

[Christian Thaeter]

Am Sat, 9 Jul 2011 12:06:37 -0400
schrieb Hendrik Boom <hendrik at topoi.pooq.com>:

> On Sat, Jul 09, 2011 at 04:31:48AM +0200, Ichthyostega wrote:
> > 
> > Hi Christian,
> > 
> > thanks for the detailed explanation (much appreciated);
> > for now I'll just concentrate on one detail...
> > 
> > Christian Thaeter schrieb:
> > 
> > > I am a bit worried that we must watch carefully that there are no 
> > > failure-avalances,
> > 
> > Why, what would be the problem?
> > My expectation was rather that we don't propagate in forward
> > direction (i.e. we don't notify successor jobs if a predecessor
> > fails). Rather, for a given job now able to be scheduled, we check
> > if all prerequisites are there and only at that point even notice
> > that a prerequisite was marked as failure. This way, there can't be
> > an avalanche of recursive calls. Rather, we just mark this one as
> > failure too and go ahead with the next job able to be scheduled?
> 
> If the prerequisite of some job fails, that it might be useless to
> run the job.  (this depends on the job, of course).  In that case, it
> may also be useless to run any of the other prerequisites of that job.

good point, and doable by the way I want to implement it.
Prerequisites will be computed in parallel (as much cores you have,
the time you know about a failure most others are possibly already
running if not finished) and fill the cache for later reuse (certain
prerequisites are universally useable). There may be less opportunity
to abort them and certainly wont work for all, its prolly still a good
idea to investigate.

	Christian


> 
> -- hendrik
> _______________________________________________
> Lumiera mailing list
> Lumiera at lists.lumiera.org
> http://lists.lumiera.org/cgi-bin/mailman/listinfo/lumiera
> http://lumiera.org/donations.html