hi some basic questions 1)which OS and version 2)which SGE version and upgrade from which version 3)what is the jobs script 4)queue configurationsome time it is easier just install new version by using different PORT then add various customer configuration
when everything looking good then you migrate from old to newthis way you will have two version of GE running on the cluster and you have failback option
you use different settings.sh to access different version
regards On 5/30/2012 3:57 AM, Vic wrote:
Fixed now!Well, I spoke too soon, didn't I? I'm still getting sporadic failures; it seems to happen a while after restarting qmaster. I'm back to getting: Unable to initialize environment because of error: range_list containes no elements Exiting. I've got MAX_DYN_EC 10000 At the top of my qconf -mconf settings. *Additionally*, running qsub against the other queue is giving me problems as well; I've got stdout set to go to $HOME/.cluster/$JOB_ID. Looking at qmon for a failed job shows me that this is so; hitting the "qalter" button shows me that stdout is set to: /home/eianpoo/.cluster/$JOB_ID/stdout However, when I look at the "why" report for why the job failed, I get: Error for job 64288: 05/30/2012 08:44:32 [94601:28096]: error: can't open output file "/home/eianpoo/64288/stdout": No such file or directory As you can see, the ".cluster" part of the path is absent. Can anyone help? My entire grid is essentially down :( Thanks. Vic. _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
--
<<attachment: laotsao.vcf>>
_______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
