> Out of curiosity, what am I setting myself up for? Just so that I can understand all the implications and make a good decision.
Are you using processor-hour limits? You may find that users adding time may cause some problems there (we don't, so I can't speak to how those limits work in Slurm). We have plans to allow users to extend their walltime, though we are looking to prevent gaming by having a tool/agent separate from Slurm to which users submit their request for walltime extension. That way we can put other limits on those requests. It'd be something that would limit the size and number of extensions, maybe even the number per day. Suppose we limited extensions to say 30% of the original WClimit: you can submit a job for 10 minutes, then when it starts you can only request an additional 3 minutes. However, a job that requested 10 days could get an additional 3 days. You'd only get so many extensions as well. Regrettably, other priorities have prevented us from getting on that task, but maybe you'd find value in this approach. Best M On Tue, Nov 17, 2015 at 12:33 PM, Jonathon A Anderson < [email protected]> wrote: > > > user[s will] game the system by submitting a job with a 1 minute time > limit, which will generally get it started very quickly because of > backfill, then they [will] increase it to whatever they wanted. > > > > > On Nov 17, 2015, at 12:10 PM, Jay Sullivan <[email protected]> > wrote: > > > > Thank you. > > > > Out of curiosity, what am I setting myself up for? Just so that I can > understand all the implications and make a good decision. > > > > Thank you once again. > > > > -Jay > > > > -----Original Message----- > > From: [email protected] [mailto:[email protected]] > > Sent: Tuesday, November 17, 2015 11:05 AM > > To: slurm-dev <[email protected]> > > Subject: [slurm-dev] Re: User Control of WallTime for running job > > > > > > Your users may well make you regret this, but here's a one-line patch to > do what you described: > > > > diff --git a/src/slurmctld/job_mgr.c b/src/slurmctld/job_mgr.c index > 47c3dd0..9518176 100644 > > --- a/src/slurmctld/job_mgr.c > > +++ b/src/slurmctld/job_mgr.c > > @@ -10409,7 +10409,8 @@ static int _update_job(struct job_record > *job_ptr, job_desc_msg_t * job_specs, > > else if (job_ptr->time_limit == job_specs->time_limit) { > > debug("sched: update_job: new time limit identical > to " > > "old time limit %u", job_ptr->job_id); > > - } else if (authorized || > > +// } else if (authorized || > > + } else if ( > > (job_ptr->time_limit > job_specs->time_limit)) { > > time_t old_time = job_ptr->time_limit; > > if (old_time == INFINITE) /* one year in > mins */ > > > > Quoting Jay Sullivan <[email protected]>: > >> Hello, > >> > >> I apologize if I missed the answer on how to do this, but I am hoping > >> there is a way. > >> > >> Scenario: A job is in the RUN state, and the job is taking longer than > >> expected. The user needs to increase the wall time of the job, to > >> allow it to complete. The user cannot increase the wall time, because > >> they do not have "operator" or "admin" privileges. > >> > >> For many reasons, I do not want to give even "operator" control to all > >> users, just to give them the ability to adjust their wall time. > >> > >> So a few questions: > >> > >> 1) Is there a way to do this with the stock configuration? > >> > >> 2) If 1 is not possible is there a way to add a custom > >> AdminLevel? One where I can set just the commands that users have > >> access to? > >> > >> 3) If neither of these are possible, can we file an RFE? > >> > >> Thanks, > >> -Jay > >> > >> Jay Sullivan > >> HPC Systems Administrator > >> Office: 310-970-3866 > >> Mobile: 424-255-2713 > > > > >
