I had seen the bit preempting and assumed that just by marking a job as production, its guaranteed whatever resources are allocated to it (by the job config) and non-producton jobs could be de-prioritized to meet the requirements.
It makes sense now. Thanks for the quick response, as always. On Thu, Feb 26, 2015 at 10:26 PM, Bill Farner <wfar...@apache.org> wrote: > This is, unfortunately, only sparsely documented [1]: > > Whether or not this is a production task backed by quota (Default: False). > > Production jobs may preempt any non-production job, and may only be > > preempted by production jobs in the same role and of higher priority. To > > run jobs at this level, the job role must have the appropriate quota. > > > What we neglect to mention is how one is granted quota. This is done with > the aurora_admin command. The idea here is that cluster administrators > make decisions about the amount of resources to guarantee to specific > users. > > $ aurora_admin set_quota > F0226 11:24:42.838977 19912 base.py:39] usage: set_quota cluster role cpu > ram[MGT] disk[MGT] > > [1] > > https://github.com/apache/incubator-aurora/blob/master/docs/configuration-reference.md#job-objects > > > > -=Bill > > On Thu, Feb 26, 2015 at 11:17 AM, Larry Weya <larryw...@gmail.com> wrote: > > > I have a pretty basic job > > > > prod_resources = Resources(cpu = 0.5, ram = 256*MB, disk=2000*MB) > > My cluster has 3 slaves, each with 4 cpu cores and 16GB or RAM but I'me > > getting > > > > Job creation failed due to error: > > > > Insufficient resource quota: CPU quota exceeded by 1.00 core(s); RAM > quota > > exceeded by 512.00 MB; DISK quota exceeded by 4000.00 MB > > This only happens when I set production to true in the job configuration, > > otherwise, the jobs run fine. > > > > -- > > Larry Weya > > > -- Larry Weya
