How to prevent parpool Error

Hi all,
I would like to ask a question related to parpool error:
Error using parpool (line 145)
Parallel pool failed to start with the following error. For more detailed information, validate the profile 'local' in the Cluster Profile Manager.
--------------- Stack Trace --------------
Full Message:
Error using parpool (line 145)
Parallel pool failed to start with the following error. For more detailed information, validate the profile 'local' in the Cluster Profile Manager.
Identifier: parallel:cluster:PoolRunValidation
------------------------------------------
I used MATLAB by submitting jobs that contains parpool work using SLURM, and when I submit 6~7 jobs, it randomly shows parpool error.
After the error, parpool becomes unable to open until I remove local_cluster_jobs under the directory one above prefdir.
I wonder why this error happens, and if there is any ways to prevent the error before occurance. I also wonder if increasing number of workers in parpool may solve the issue?
Please advise. Thank you.

 Risposta accettata

Raymond Norris
Raymond Norris il 24 Ago 2021

1 voto

There's a slight change it's because multiple jobs are running at the same time, all contending for the same subfolder. If so, this should be resolved in R2021a. In either case, contact Technical Support (support@mathworks.com) and they can either show you a workaround prior to R2021a or help troubleshoot the issue if it's something else.

2 Commenti

I have tried again in MATLAB R2021a, and i don't see the same error any longer. Thank you!
I'm facing a similar problem while using R2020a. Is there any fix to this ?

Accedi per commentare.

Più risposte (0)

Categorie

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by