[GE users] semaphore leftovers

Reuti reuti at staff.uni-marburg.de
Fri Feb 11 18:30:36 GMT 2005


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

Ron,

Quoting Ron Chen <ron_chen_123 at yahoo.com>:

<snip>
> And also, if we place a loop in the stop_proc_args, we
> need to use rsh. In some cluster env, using rsh is not
> allowed. One way around this is to allow prolog/epilog
> for parallel slave tasks. The code change should be
> small.
> 
> (In case you are interested, SGE 6.0u3:
> source/daemons/execd/exec_job.c:990 just removing the
> check for should work)

thanks. This would be, independent from the actual theme, a feature I would 
like to have. I will look into it next week. At a brief look, this would be 
executed on the slave nodes for each qrsh call, and on the head node in 
addition before and after the job? So the epilog/prolog scripts must be 
carefully prepared. Maybe this is the reason, that they are not executed for 
PEs by default.

And I see a little clear now with TM. For special applications the well 
established qrsh interface can stay there, and the TM can be used if it's 
supported by the parallel method chosen (directly as LAM or with a special 
mpiexec). I understand stdio thing, which is not forwared by SGE on it's own. 
But when there are interfaces for ch_p4,... , could there also be a rsh-like 
interface to the TM and replacing qrsh -inherit? Something like a 
script-interface to the TM, accessible direct from the shell; and the 
rsh-wrapper will call the TM instead the up to now used qrsh -inherit method?

Cheers - 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