On Thu, Dec 10, 2015 at 03:04:32PM +0100, Jerome Poitout wrote: > Running jobs. If team1 is consumming 80% of CPU Time and team2 20%, > then team2 jobs should have higher priority. > > J??r??me > > Le 10/12/2015 12:18, Reuti a ??crit : > > Do you want to honor the running jobs only (functional policy) or also the > > past usage over a moving timeframe of the last 30 days or alike (share tree > > policy)? > > > > -- Reuti Then, as Reuti said, you probably want to set up the functional policy as the primary mechanism for controlling priority. Assign a high value to weight_ticket when compared to weight_urgency. Assign a lot of tickets to weight_tickets_functional and none or few to weight_tickets_share.
The teams can probably be represented by either projects or departments. So give either weight_project or weight_department a high value compared to the other one and weight_user. Assign each project or department an fshare proportionate to the share of the cluster you want them to have. Assign users to the appropriate project or department. Users can be in multiple projects (and select which one to use for each job when they submit it) but only one department. William
signature.asc
Description: Digital signature
_______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
