radical-cybertools / radical.pilot

RADICAL-Pilot
http://radical-cybertools.github.io/radical-pilot/index.html
Other
54 stars 23 forks source link

Multiple Pilots Being Generated when Running FACTS on Amarel #2965

Closed AlexReedy closed 6 months ago

AlexReedy commented 1 year ago

Hey all, I've been noting an issue recently where when I am running large configurations of FACTS runs (e.g. nsamples=20000 and nlocations=66190) there will be anywhere from 2-4 pilots created in the sandbox. It looks like some resubmission of the pilot but I am unable to diagnose the cause.

mtitov commented 1 year ago

@AlexReedy FACTS uses EnTK, and EnTK has a default number of attempts equal to 3, you can change it with a corresponding parameter here https://github.com/radical-collaboration/facts/blob/main/runFACTS.py#L76 - add reattempts=1 (parameter reattempts actually sets the number of attempts)

mtitov commented 1 year ago

(*) New pilots start only if the previous ones failed to complete the execution of a workflow

mtitov commented 1 year ago

@AlexReedy Hi Alex, did you decrease the number of attempts as I wrote earlier, or it is not an issue anymore and one pilot completes everything?

andre-merzky commented 6 months ago

Closing this for lack of activity.