Also.. just to add. If you are using Bright, the easiest way to fix your core count issue is to use the cmsh command, go to the node definition; roles; slurmclient and then edit the appropriate fields (details are in the manual).
This will populate the slurm.conf with the corrected node definitions. You may want to use pdsh and execute the following: pdsh -w node-[01-20] slurmd -C | grep -v UpTime |sort It should tell you the correct settings to use in the node- definitions above. Similarly, in Bright, you should set the default mem per cpu setting for your partitions / queue. Once you do that...your other problem should be sorted as well. Best Regards, MD Sharma(sent via a mobile device.. please blame the AI driven autocorrect algorithm for any errors) ________________________________ From: slurm-users <slurm-users-boun...@lists.schedmd.com> on behalf of Alex Chekholko <a...@calicolabs.com> Sent: Friday, April 5, 2019 12:35:42 AM To: Slurm User Community List Subject: Re: [slurm-users] Slurm 1 CPU Hi Chris, re: "can't run more than 1 job per node at a time. " try "scontrol show config" and grep for defmem IIRC by default the memory request for any job is all the memory in a node. Regards, Alex On Thu, Apr 4, 2019 at 4:01 PM Andy Riebs <andy.ri...@hpe.com<mailto:andy.ri...@hpe.com>> wrote: in slurm.conf, on the line(s) starting "NodeName=", you'll want to add specs for sockets, cores, and threads/core. ________________________________ From: Chris Bateson <cbate...@vt.edu><mailto:cbate...@vt.edu> Sent: Thursday, April 04, 2019 5:18PM To: Slurm-users <slurm-users@lists.schedmd.com><mailto:slurm-users@lists.schedmd.com> Cc: Subject: [slurm-users] Slurm 1 CPU I should start out by saying that I am extremely new to anything HPC. Our end users purchased a 20 node cluster which a vendor set up for us with Bright/Slurm. After our vendor said everything was complete and we started migrating our users workflow to the new cluster they discovered that they can't run more than 1 job per node at a time. We started researching enabling consumable resources which I believe we've done so however we're getting the same result. I've just discovered today that both scontrol show node and sinfo -lNe show that each of our nodes have 1 CPU. I'm guessing that's why we can't submit more than 1 job at a time. I'm trying to determine where is it getting this information and how can I get it to display the correct CPU information. Sample info: scontrol show node NodeName=cnode001 Arch=x86_64 CoresPerSocket=1 CPUAlloc=0 CPUErr=0 CPUTot=1 CPULoad=0.01 AvailableFeatures=(null) ActiveFeatures=(null) Gres=(null) NodeAddr=cnode001 NodeHostName=cnode001 Version=17.11 OS=Linux 3.10.0-693.el7.x86_64 #1 SMP Thu Jul 6 19:56:57 EDT 2017 RealMemory=192080 AllocMem=0 FreeMem=188798 Sockets=1 Boards=1 State=IDLE ThreadsPerCore=1 TmpDisk=2038 Weight=1 Owner=N/A MCS_label=N/A Partitions=defq BootTime=2019-03-26T14:28:24 SlurmdStartTime=2019-03-26T14:29:55 CfgTRES=cpu=1,mem=192080M,billing=1 AllocTRES= CapWatts=n/a CurrentWatts=0 LowestJoules=0 ConsumedJoules=0 ExtSensorsJoules=n/s ExtSensorsWatts=0 ExtSensorsTemp=n/s sinfo -lNe NODELIST NODES PARTITION STATE CPUS S:C:T MEMORY TMP_DISK WEIGHT AVAIL_FE REASON cnode001 1 defq* idle 1 1:1:1 192080 2038 1 (null) none lscpu Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 48 On-line CPU(s) list: 0-47 Thread(s) per core: 1 Core(s) per socket: 24 Socket(s): 2 NUMA node(s): 2 Vendor ID: GenuineIntel CPU family: 6 Model: 85 Model name: Intel(R) Xeon(R) Platinum 8168 CPU @ 2.70GHz Stepping: 4 CPU MHz: 2700.000 BogoMIPS: 5400.00 Virtualization: VT-x L1d cache: 32K L1i cache: 32K L2 cache: 1024K L3 cache: 33792K NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46 NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch epb cat_l3 cdp_l3 intel_pt tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts slrum.conf SelectType Configuration SelectType=select/cons_res SelectTypeParameters=CR_Core_Memory PartitionName=defq Default=YES MinNodes=1 AllowGroups=ALL PriorityJobFactor=1 PriorityTier=1 DisableRootJobs=NO RootOnly=NO Hidden=NO Shared=NO GraceTime=0 PreemptMode=OFF ReqResv=NO AllowAccounts=ALL AllowQos=ALL LLN=NO ExclusiveUser=NO OverSubscribe=YES OverTimeLimit=0 State=UP Nodes=cnode[001-020] I can provide other configs if you feel that it could help. Any ideas? I would have thought that slurm would grab the CPU information from the CPU instead of the configuration. Thanks Chris
