[gridengine users] Evaluating different scheduler strategies by simulating a past workflow

Reuti reuti at staff.uni-marburg.de
Thu Sep 29 18:45:55 UTC 2011


Am 29.09.2011 um 15:02 schrieb Fabio Martinelli:

> Dear Sun Grid Engine colleagues
> is there a way to apply again a past workflow by using information  
> like that stored inside the ARCO DB, the reporting file or the  
> accounting file ?
> obviously with a shorter time scale and we don't mind about about  
> memory consumption and I/O, just slots assignment.

for this you will need to set up a share-tree policy. If you don't  
take the CPU usage into account, one can set: ACCT_RESERVED_USAGE and  
SHARETREE_RESERVED_USAGE in SGE's configuration (man sge_conf). As it  
then reflects the granted CPU time (and not the actual used one), you  
can devide it (the CPU time) by wallclock time and get the slot count.

> basically we need to understand if the actual scheduler policy is  
> "fair", at least according to our personal concept of "fairness";

Well, this is only your personal view. What do you judge as "fair"?  
Same slot count in the cluster for running jobs? Same slot count in  
the last 7 days, independet how long they ran?

-- Reuti

> so far it's unfair and as a reaction we should tune some scheduler  
> parameters and observe the Sun Grid Engine behavior but as jobs take  
> hours or days to complete this tuning process is simply to long to  
> manage.
> just to cite a concrete solution, I never tried this Moab Simulator
> http://www.adaptivecomputing.com/resources/docs/mwm/6-1/Content/topics/analyzing/simulations.html
> but the 'simulation' concept and the related tools seem to be there,  
> so I wonder what about {Sun} Grid Engine.
> I'm using 6.2-5 on Linux 64bit
> many thanks,
> best regards
> Fabio Martinelli
> _______________________________________________
> users mailing list
> users at gridengine.org
> https://gridengine.org/mailman/listinfo/users

More information about the users mailing list