Hi, Am 13.10.2013 um 10:44 schrieb Mahbube Rustaee:
> sometimes there are jobs that qstat reports "r" state , but really aren't. You mean they finished already and the qmaster didn't realize it? > I need a monitoring script that check those jobs and report to qdel from sge. > Is such script? No, as it shouldn't happen under normal circumstances. The qmaster will wait for the execd on the node to inform him about the end of a job. You have to create a script that checks whether a shephered for a given job number still exists on the machine it is supposed to and handle timeouts due to network problems properly. This is the reason why it's not done automatically. -- Reuti > thx > _______________________________________________ > users mailing list > [email protected] > https://gridengine.org/mailman/listinfo/users _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
