LHC Computer Grid - Rollout > [mailto:[log in to unmask]] On Behalf Of Bly, > MJ (Martin) said: > Apropos the condor jobs: the problem appears to be that they are > 'still' running after the legitimate jobid to which they are > associated is done and gone. Do they change the process group? As far as I remember PBS does a kill on the pgid of the parent process when the job ends. I guess the accounting may also be based on the pgid? Stephen