Groups.grid.org

Hi all,

I have developed an application that makes use of a grid environment to expedite execution when possible. I am currently porting to SGE 6.2 from Apple's Xgrid and have two questions which I have not yet been able to answer:

1) Is there a reliable way to monitor job execution so that I know when a job is finished? Currently, after submitting jobs to the grid, my application polls the sent jobs one by one using qstat -j . While the jobs are pending or running, I get some output but when a job has finished, all I get is an error message that the job id does not exist. What I'm looking for is a command that will tell me, based on job id, what's the status of the job, whether pending, running or finished or anything else. In Xgrid, this is accomplished with xgrid -job attributes -id .

2) Is there a way, after submitting a binary job in the grid, to get it's return value (aka exit code)? My application currently relies on this to determine if the execution of a job was successful and branches accordingly. In case it's not possible, what would be best strategy to determine job success or failure? I've tried monitoring the content of the jobname.o and jobname.e files that are produced but that leads to inconsistent results: some jobs finish successfully while still having stderr output.

Thanks and have a good weekend!

Burt