BRC Cluster Workshop

chris@node0:~$ sinfo
PARTITION AVAIL  TIMELIMIT  NODES  STATE NODELIST
all*         up   infinite      4  drain node[5-7,11]
all*         up   infinite      9  alloc node[1-4,8,12-13,15,17]
all*         up   infinite      4   idle node[9-10,14,16]
short        up    2:00:00      4  drain node[5-7,11]
short        up    2:00:00     10  alloc node[1-4,8,12-13,15,17,91]
short        up    2:00:00      5   idle node[9-10,14,16,92]
gpu          up   infinite      1  alloc node91
bigmem       up   infinite      1   idle node92

chris@node0:~$ sinfo -o "%C"
CPUS(A/I/O/T)
78/178/64/320

chris@node0:~$ sinfo -p all -o "%C"
CPUS(A/I/O/T)
77/131/64/272

Note: drain means that the node is up but is not accepting new jobs. If jobs were running on the node it would say drng: the running jobs would be allowed to complete at which time the node enters drain state. A node is “drained” by the cluster administrators for maintenance or updates.

Note: on hyperion, which has a later version of SLURM, sinfo will print mix for a node which is partially allocated, and alloc for a node which is completely allocated.