Merge remote-tracking branch 'origin/master' into upcomingMaintenance

roles/slurm-common/tasks/main.yml

@@ -73,6 +73,7 @@
       - hwloc-devel
       - lua
       - lua-devel
+      - python3
   become: true
   when: ansible_os_family == "RedHat"
 
@@ -92,6 +93,7 @@
       - liblua5.2-dev
       - hwloc
       - libhwloc-dev
+      - python3
   become: true
   when: ansible_os_family == "Debian"
 
@@ -118,32 +120,12 @@
 
 - include: createSlurmDirectories.yml
 
-- name: check slurm generic resource
-  shell: "{{ slurm_gres_check }}"
-  register: slurm_generic_resource
-  ignore_errors: true
-  when: slurm_gres_check is defined
-  check_mode: no
-  changed_when: False
-
-- name: Gres - Test for Nvidia devices
-  script: scripts/nvidia-probe.py
-  register: probeOutput
-  check_mode: no
-  changed_when: False
-
 - name: get cpu count
   shell: 'lscpu | grep "On-line CPU" | cut -f 2 -d ":" | sed "s/\ *//g"'
   register: cpucount
   check_mode: no
   changed_when: False
 
-- name: "set nvidiaprobe slurm_gres_list"
-  set_fact: "slurm_gres_list={{  probeOutput.stdout }}"
-
-- name: template gres.conf file
-  template: src="gres.conf.j2" dest={{ slurm_dir }}/etc/gres.conf mode=644
-  become: true
 
 - name: make slurm prolog dir
   file: path=/opt/slurm/etc state=directory mode=755
@@ -158,15 +140,6 @@
   template: src=slurm.epilog.j2 dest=/opt/slurm/etc/slurm.epilog mode=755
   become: true
 
-- name: install slurm.conf
-  copy: src=files/slurm.conf dest={{ slurm_dir }}/etc/slurm.conf
-  become: true
-  when: slurm_use_vpn==False
-
-- name: install slurm.conf
-  template: src=slurm-vpn.conf.j2 dest={{ slurm_dir }}/etc/slurm.conf
-  become: true
-  when: slurm_use_vpn==True
 
 - name: setup envirnment variables
   template: src=slurm_setup.sh.j2 dest=/etc/profile.d/slurm_setup.sh

roles/slurm-common/templates/slurm.conf.j2

-#
-# Example slurm.conf file. Please run configurator.html
-# (in doc/html) to build a configuration file customized
-# for your environment.
-#
-#
-# slurm.conf file generated by configurator.html.
-#
-# See the slurm.conf man page for more information.
-#
-ClusterName={{ clustername }}
-ControlMachine={{ slurmctrl }}
-#ControlAddr=
-#BackupController=
-#BackupAddr=
-#
-SlurmUser=slurm
-SlurmdUser=root
-SlurmctldPort=6817
-SlurmdPort=6818
-AuthType=auth/munge
-#JobCredentialPrivateKey=
-#JobCredentialPublicCertificate=
-StateSaveLocation={{ slurmstatedir }}
-SlurmdSpoolDir={{ slurmdatadir }}
-SwitchType=switch/none
-MpiDefault=pmi2
-SlurmctldPidFile={{ slurmpiddir }}/slurmctld.pid
-SlurmdPidFile={{ slurmpiddir }}/slurmd.pid
-ProctrackType=proctrack/linuxproc
-#PluginDir=
-CacheGroups=0
-#FirstJobId=
-ReturnToService=1
-#MaxJobCount=
-#PlugStackConfig=
-#PropagatePrioProcess=
-#PropagateResourceLimits=
-#PropagateResourceLimitsExcept=
-#Prolog=
-#Epilog=
-#SrunProlog=
-#SrunEpilog=
-#TaskProlog=
-#TaskEpilog=
-TaskPlugin=task/cgroup
-#TaskPlugin=task/affinity
-#TaskPlugin=task/affinity,task/cgroup
-{% if slurm_lua is defined %}
-JobSubmitPlugins=lua
-{% endif %}
-OverTimeLimit=1
-CompleteWait=10
-
-#TrackWCKey=no
-#TreeWidth=50
-#TmpFS=
-#UsePAM=
-#
-# TIMERS
-
-SlurmctldTimeout=3000 #added due to network failures causing jobs to be killed
-
-#SlurmctldTimeout=300
-#SlurmdTimeout=300
-#InactiveLimit=0
-#MinJobAge=300
-KillWait=10
-#Waittime=0
-#
-# SCHEDULING
-SchedulerType={{ slurmschedulertype }}
-#SchedulerAuth=
-#SchedulerPort=
-#SchedulerRootFilter=
-SelectType={{ slurmselecttype }}
-{% if slurmselecttype.find("cons_res") > 0 %}
-SelectTypeParameters=CR_Core_Memory
-{% endif %}
-FastSchedule={{ slurmfastschedule }}
-#PriorityType=priority/multifactor
-#PriorityFlags=Ticket_Based
-#PriorityCalcPeriod=5
-#PriorityDecayHalfLife=0
-#PriorityUsageResetPeriod=14-0
-##PriorityWeightFairshare=10000
-#PriorityWeightAge=10000
-#PriorityWeightPartition=10000
-#PriorityWeightJobSize=10000
-#PriorityMaxAge=14-0
-#
-# LOGGING
-{% if slurmctlddebug %}
-SlurmctldDebug={{ slurmctlddebug.level }}
-SlurmctldLogFile={{ slurmctlddebug.log }}
-{% else %}
-#SlurmctldDebug=
-#SlurmctldLogFile=
-{% endif %}
-{% if slurmddebug %}
-SlurmdDebug={{ slurmddebug.level }}
-SlurmdLogFile={{ slurmddebug.log }}
-{% else %}
-#SlurmdDebug=
-#SlurmdLogFile=
-{% endif %}
-{% if slurmschedlog %}
-SlurmSchedlogLevel={{ slurmschedlog.level }}
-SlurmSchedLogFile={{ slurmschedlog.log }}
-{% else %}
-#SlurmSchedlogLevel=
-#SlurmSchedLogFile=
-{% endif %}
-JobCompType=jobcomp/none
-#JobCompLoc=
-#
-{% if slurmjob is defined %}
-Prolog={{ slurmjob.prolog }}
-Epilog={{ slurmjob.epilog }}
-{% endif %}
-#
-# ACCOUNTING
-#JobAcctGatherType=jobacct_gather/linux
-#JobAcctGatherFrequency=30
-#
-AccountingStorageType=accounting_storage/slurmdbd
-AccountingStorageHost={{ slurmctrl }}
-#AccountingStorageEnforce=limits,safe
-#AccountingStorageLoc=
-#AccountingStoragePass=
-#AccountingStorageUser=
-#
-#GRES
-GresTypes=gpu
-
-# Fair share
-{% if slurmfairshare.def %}
-PriorityWeightFairshare={{ slurmfairshare.val }}
-{% endif %}
-
-DisableRootJobs=YES
-MpiParams=ports=12000-12999
-# COMPUTE NODES
-{% set nodelist = [] %}
-{% for queue in slurmqueues %}
-{% for node in groups[queue.group] %}
-{% if nodelist.append(node) %}
-{% endif %}
-{% endfor %}
-{% endfor %}
-{% for node in nodelist|unique %}
-NodeName={{ node }} Procs={{ hostvars[node]['ansible_processor_vcpus'] }} RealMemory={{ hostvars[node].ansible_memory_mb.real.total }} Sockets={{ hostvars[node]['ansible_processor_vcpus'] }} CoresPerSocket=1 ThreadsPerCore={{ hostvars[node].ansible_processor_threads_per_core }} {% if hostvars[node].ansible_hostname.find('vis') != -1 %}Gres=gpu:1{% endif %} {% if hostvars[node]['ansible_processor_vcpus'] == 1 %}Weight=1{% endif %}{% if hostvars[node]['ansible_processor_vcpus'] > 1 and hostvars[node]['ansible_processor_vcpus'] <= 16 %}Weight=3{% endif %}{% if hostvars[node]['ansible_processor_vcpus'] > 16 and hostvars[node]['ansible_processor_vcpus'] <= 20 %}Weight=5{% endif %}{% if hostvars[node]['ansible_processor_vcpus'] > 20 and hostvars[node]['ansible_processor_vcpus'] <= 40 %}Weight=7{% endif %}{% if hostvars[node]['ansible_processor_vcpus'] > 40 and hostvars[node]['ansible_processor_vcpus'] <= 64 %}Weight=8{% endif %}{% if hostvars[node]['ansible_processor_vcpus'] > 64 and hostvars[node]['ansible_processor_vcpus'] <= 128 %}Weight=9{% endif %}{% if hostvars[node]['ansible_processor_vcpus'] > 128 %}Weight=10{% endif %} Feature=stage1 State=UNKNOWN  
-{% endfor %}
-
-{% for queue in slurmqueues %}
-PartitionName={{ queue.name }} {% if queue.default %}Default=yes{% endif %} Nodes={{ groups[queue.group]|join(',') }} DefaultTime=72:00:00 State=UP 
-{% endfor %}

roles/slurm-start/templates/slurmd.service.j2

 [Unit]
 Description=Slurm node daemon
 After=network.target
-ConditionPathExists={{ slurm_dir }}/etc/slurm.conf
+# After Slurm-20.02 (configless), this is not needed
+# ConditionPathExists={{ slurm_dir }}/etc/slurm.conf
 
 [Service]
 Type=forking
 KillMode=process
 LimitMEMLOCK=infinity
 #EnvironmentFile=/etc/default/slurmd
-ExecStart={{ slurm_dir }}/sbin/slurmd $SLURMD_OPTIONS
+ExecStart={{ slurm_dir }}/sbin/slurmd --conf-server {{ slurmctrl }}:6817
 PIDFile={{ slurmpiddir }}/slurmd.pid
 
 [Install]