[GE users] more info for "rsh zombies when using mpich2" -- johnny layne

Reuti reuti at staff.uni-marburg.de
Mon Jul 30 21:28:09 BST 2007


Am 30.07.2007 um 17:25 schrieb Johnny Layne:

> hi all again,
>    OK I _am_ using individualized .smpd files for each job (sorry,  
> I really need more sleep).  I wonder what happens when you launch  
> numerous of these jobs and parts of them run on the same node, for  
> instance a 4 processor machine divided amongst 2 jobs, and then  
> stopmich2.sh runs.  Sorry for the many posts, I will get busy  
> investigating & waiting for info.  Thanks,

I assume you mean the daemon-based smpd method. So you are using the  
"-smpdfile <filename>" option?

Although MPICH2 stores the used machines in the .smpd file, it is not  
used for the shutdown (so you could still stay with one file).  
Instead stopmpich2.sh will shutdown the daemons in the reverse order  
they were started on the nodes. As each job uses a different port  
number, only the daemons belonging to the job will be addressed.

Just be sure, that also your job uses the specific port number in the  

-- Reuti

To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net

More information about the gridengine-users mailing list