[Bioperl-pipeline] pipeline connections

[shawnh@worf.fugu-sg.org]

> If it fails does it get rerun later? Isn't it most likely that if one
> fails all will fail because that node is screwy? How do you let LSF know
> things are not going well and that it should migrate the runner.pl job to
> another node?

If it fails, it will simply get updated as failed. runner.pl doesn't do the fetching
of failed jobs. it just runs the jobs the manager tells it to. Then next time round, PipelineManager
will fetch it again. It has a 1/60 of getting sent to the same node.
Right now, PipelineManager has no idea whether nodes are working. It just does bsub.
LSF should be managing where jobs get sent off.

shawn



> 
> Elia
> 
> > 
> > so we either have to call the Batchsubmission object n times or once.
> > 
> > 
> > shawn
> > 
> > 
> > On Mon, 23 Sep 2002, Elia Stupka wrote:
> > 
> > > > The idea of batching a bunch of jobs to nodes is 'in the future' we
> > > > can save time...just issuing one bsub command and the node can run n
> > > > jobs at once.
> > > 
> > > But you agree it's not worth doing for genewises, right?
> > > 
> > > Elia
> > > 
> > > ********************************
> > > * http://www.fugu-sg.org/~elia *
> > > * tel:    +65 6874 1467        *
> > > * mobile: +65 9030 7613        *
> > > * fax:    +65 6779 1117        *
> > > ********************************
> > > 
> > > 
> > > _______________________________________________
> > > bioperl-pipeline mailing list
> > > bioperl-pipeline@bioperl.org
> > > http://bioperl.org/mailman/listinfo/bioperl-pipeline
> > > 
> > 
> > 
> 
> Elia
> 
> ********************************
> * http://www.fugu-sg.org/~elia *
> * tel:    +65 6874 1467        *
> * mobile: +65 9030 7613        *
> * fax:    +65 6779 1117        *
> ********************************
> 
> 

-- 
********************************
* Shawn Hoon
* http://www.fugu-sg.org/~shawnh
********************************