[Bioclusters] OpenPBS problems

Tim Cutts bioclusters@bioinformatics.org
Tue, 2 Dec 2003 18:07:27 +0000


On 02-Dec-03, Rayson Ho wrote:
> 
> Sending SIGSTOP is different from swapping the whole process to disk!
> 
> All batch systems (SGE, PBS, LSF) can send SIGSTOP to the job, but
> stopping the process still consumes the system resource. Eventually,
> the swap partition will be used up by suspended jobs.

Sure.  OK, I misunderstood the question then, and checkpointing is
indeed what's wanted.

I have yet to see anyone write a properly checkpointable bioinformatics
pipeline.

Tim

-- 
Dr Tim Cutts
Informatics Systems Group
Wellcome Trust Sanger Institute, Hinxton, Cambridge, CB10 1SA, UK