cexec is a handy set of python scripts for executing a command on multiple nodes. Check out its man page.
Sometimes, if lam does not shutdown properly it leaves orpan processes which prevent you from booting. You can check for orphan processes by typing "cexec -p ps -axuw|grep $USER". This will list all process you have running on each node, prefaced with the cluster, node name. Lam stores some runtime files in /tmp/lam-$USER@$HOSTNAME, i.e. /tmp/lam-fairchild@nile. Sometimes the existence of this directory will stop lam from booting. You can check for these files by using the cexec tool again. "cexec -p ls /tmp|grep $USER". First, try the wipe command. If that does not eliminate the processes try, "cexec tkill", and check for any processes. If you still have orphaned lam processes, ssh to the machine/s and kill them. You can remove the /tmp/lam-$USER* direcortories using the cexec tool. "cexec rm -r /tmp/lam-$USER*".
You can find documentation on using Sun GridEngine here
If you are still having problems, please submit a trouble ticket to the PBPL Helpdesk with a clear, specific description of the problem.
