Hello, Running slurm 15.08.11 on FreeBSD 10.3-RELEASE we're seeing issues with sacct not reporting CPU information correctly (or I am misunderstanding what should be reported). Where should we start digging to figure out why NCPUS and AllocCPUS are always 0?
Regards,
Joseph
% sacct --format=jobid,jobname,elapsed,ncpus,alloccpus,reqcpus,state | head
JobID JobName Elapsed NCPUS AllocCPUS ReqCPUS State
------------ ---------- ---------- ---------- ---------- -------- ----------
45069 NBS_1 00:02:30 0 0 1 COMPLETED
45070 NBS_2 00:01:27 0 0 1 COMPLETED
45071 NBS_3 00:03:15 0 0 1 COMPLETED
45072 NBS_4 00:02:44 0 0 1 COMPLETED
45073 NBS_5 00:02:56 0 0 1 COMPLETED
45074 NBS_6 00:04:08 0 0 1 COMPLETED
45075 NBS_7 00:01:43 0 0 1 COMPLETED
45076 NBS_8 00:02:24 0 0 1 COMPLETED
% head /var/log/slurm/slurm_accounting.log
45069 all 1464193197 1464193198 1001 0 - - 0 NBS_1 1 4294895959 1 awarnach1
(null)
45070 all 1464193197 1464193198 1001 0 - - 0 NBS_2 1 4294895958 1 awarnach1
(null)
45071 all 1464193197 1464193198 1001 0 - - 0 NBS_3 1 4294895957 1 awarnach1
(null)
45072 all 1464193197 1464193198 1001 0 - - 0 NBS_4 1 4294895956 1 awarnach1
(null)
45073 all 1464193197 1464193198 1001 0 - - 0 NBS_5 1 4294895955 1 awarnach2
(null)
45074 all 1464193197 1464193198 1001 0 - - 0 NBS_6 1 4294895954 1 awarnach2
(null)
45075 all 1464193197 1464193198 1001 0 - - 0 NBS_7 1 4294895953 1 awarnach2
(null)
45076 all 1464193197 1464193198 1001 0 - - 0 NBS_8 1 4294895952 1 awarnach2
(null)
45077 all 1464193197 1464193198 1001 0 - - 0 NBS_9 1 4294895951 1 awarnach3
(null)
45078 all 1464193197 1464193198 1001 0 - - 0 NBS_10 1 4294895950 1 awarnach3
(null)
% sinfo -Nel
Wed May 25 13:30:45 2016
NODELIST NODES PARTITION STATE CPUS S:C:T MEMORY TMP_DISK WEIGHT
FEATURES REASON
awarnach1 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach2 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach3 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach4 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach5 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach6 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach7 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach8 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach9 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach10 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach11 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach12 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach13 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach14 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach15 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach16 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach18 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach19 1 all* idle 4 4:1:1 1 0 1
(null) none
awarnach20 1 all* idle 48 48:1:1 1 0 1
(null) none
Log of the slurm build via the FreeBSD port:
http://pkg.awarnach.mathstat.dal.ca/data/10amd64-default/2016-05-18_00h20m52s/logs/slurm-wlm-15.08.11.log
signature.asc
Description: PGP signature
