I think SGE handles it by setting the hard limit at job startup time... Run the following script inside SGE, what is the output?? #!/bin/sh ulimit -a Rayson --- Barry J Mcinnes <Barry.J.Mcinnes at noaa.gov> wrote: > Hi Chris, > did you get anywhere with this request? > We have tried a test PBS cluster and cannot get it to limit the > execution size either, although we are still trying to get them to > make this happen. > > > Hi Barry, > > This is probably a Grid Engine issue; enforcing memory usage limits > for > the purpose of sending suspend/terminate signals etc. has long been > part > of the SGE feature set. > > You may want to post this message to the Grid Engine users mailing > list: > users at gridengine.sunsource.net to see if anyone else is doing > memory > limit enforcement on Mac OS X with SGE 6 > > I did a brief search through the open bug reports but did not see any > open issues that match your problem. > > That said though, I've only seen jobs trip SGE queues into error > state > 'E' when the job itself failed in a spectacular manner. It's odd that > it > errors out several queues and the runs on a different box. > > If you are willing to share your test code I'd be interested in > trying > to replicate on one of the G5 clusters I have access to. > > Regards, > Chris > -- > ------ > Barry Mc Innes > Email: Barry.J.McInnes at noaa.gov > Phone: 303-4976231 FAX: 303-4977013 > Smail: NOAA/CDC > 325 Broadway R/CDC1 > Boulder CO 80305 > ------> _______________________________________________ > Bioclusters maillist - Bioclusters at bioinformatics.org > https://bioinformatics.org/mailman/listinfo/bioclusters > __________________________________ Celebrate Yahoo!'s 10th Birthday! Yahoo! Netrospective: 100 Moments of the Web http://birthday.yahoo.com/netrospective/