hi All：
my job runing at a low priority partition can not be requeue when it preempted by a job at a hight priority partition。 my slurm.conf as below: # ----------------------------------------------------- PreemptType=preempt/partition_prio PreemptMode=REQUEUE JobRequeue=1 NodeName=cp01-sys-hic-gpu-01.cp01.baidu.com,cp01-sys-hic-gpu-00.cp01.baidu.com CPUs=12 RealMemory=500 Sockets=2 CoresPerSocket=6 ThreadsPerCore=1 Procs=1 State=IDLE Gres=gpu:2 PartitionName=hig Default=NO Priority=65533 Nodes=cp01-sys-hic-gpu-00.cp01.baidu.com MaxTime=INFINITE State=UP PartitionName=low AllowGroups=ALL Default=yes Priority=100 Nodes=cp01-sys-hic-gpu-01.cp01.baidu.com,cp01-sys-hic-gpu-00.cp01.baidu.com MaxTime=INFINITE State=UP # ----------------------------------------------------- first I run a job(job1) at Partition low, and the job run at cp01-sys-hic-gpu-00.cp01.baidu.com. then I run a job(job2) at Partition hig, and the job also run at cp01-sys-hic-gpu-00.cp01.baidu.com, now job1 will be preempted and have a state of PD. after job2 finished, job1 will be rerun. how can I let job1 running at cp01-sys-hic-gpu-01.cp01.baidu.com after it preempted (after job1 preempted, job2 running at cp01-sys-hic-gpu-00.cp01.baidu.com in a partition hig, but now job1 have an idle node in partition low, In my mind, job1 should be requeue and rerun at cp01-sys-hic-gpu-01.cp01.baidu.com). Best, shengzhao wen
