(We're running slurm 2.5.7.) We have set up the topology plugin on our cluster. I have a question about the semantics of the --switches option and max_switch_wait setting.
My interpretation is that - Slurm will first try to place a job on 1 switch, if that can be done without delaying the job start. If not, it will place it on more than one switch. - If a job specifies --switches=num@delay, slurm will not place the job on more than "num" switches, unless that would delay the job more than "delay". - The max_switch_wait setting is the maximum value a job can specify for "delay". So in order for allowing users to specify a delay of, say, 12 hours, one must set max_switch_wait in slurm.conf to something as large as 12 hours. Is this the right interpretation? -- Regards, Bjørn-Helge Mevik, dr. scient, Department for Research Computing, University of Oslo