Merge branch 'slurmtestfix' into 'master'

Slurmtestfix

See merge request !326

.gitlab-ci.yml

@@ -141,12 +141,14 @@ tests:
     - ansible -m ping -i files/inventory.$STACKNAME --key-file ../gc_key.pem all
     - ansible -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "sudo ls" all
     - echo -e '[defaults]\r\nallow_world_readable_tmpfiles = True' > ansible.cfg
+    - ansible-playbook -i files/inventory.$STACKNAME --key-file ../gc_key.pem ./tests/mockSlurmData.yml
     # Need to find a better check for sinfo
     #- ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "find /opt/ -name sinfo -type f" ManagementNodes
     #- ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "find /opt/ -name squeue -type f" ManagementNodes
     - ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "systemctl is-active --quiet mariadb" SQLNodes
     - ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "systemctl is-active --quiet slurmctld" ManagementNodes
     - ansible -B 30 -i files/inventory.$STACKNAME --key-file ../gc_key.pem -a "systemctl is-active --quiet slurmdbd" ManagementNodes
+    
 
     - bash -e ./tests/run_tests.sh all "files/inventory.$STACKNAME" "../gc_key.pem"
     - bash -e ./tests/run_tests.sh ComputeNodes "files/inventory.$STACKNAME" "../gc_key.pem"

CICD/plays/allnodes.yml

@@ -26,6 +26,8 @@
   - { role: config_repos, tags: [ repos ] }
   - { role: upgrade, tags: [ upgrade ]}
   - { role: set_password }
+  - { role: ntp }
+  - { role: set_timezone }
 
 
 - hosts: 'DesktopNodes,ComputeNodes,LoginNodes,ManagementNodes'
@@ -46,5 +48,4 @@
   - { role: calculateKnownHosts, tags: [ calculateKnownHosts ] }
   - { role: SSHKnownHosts, tags: [ known_hosts ] }
   - { role: jasons_ssh_ca, tags: [ ssh_ca ] }
-  - { role: ntp }
-  - { role: set_timezone }
+

CICD/plays/computenodes.yml

@@ -57,20 +57,6 @@
   roles:
   - { role: slurm_config, tags: [slurm, slurm_config] }
 
-- hosts: 'DesktopNodes,ComputeNodes'
-  gather_facts: False
-  vars_files:
-  - vars/passwords.yml
-  - vars/names.yml
-  - vars/ldapConfig.yml
-  - vars/filesystems.yml
-  - vars/slurm.yml
-  - vars/vars.yml
-  strategy: free
-  roles:
-  - { role: slurm-start, start_slurmd: True, tags: [ slurm, slurmstart ] }
-  #- { role: mate-de-install, tags: [ mate-de-install ] }   # TODO this crashes for everything except cmca
-
 - hosts: 'K1Nodes'
   tasks:
   - { name: set nvidia driver version, set_fact: nvidia_version='367.130' }
@@ -81,7 +67,7 @@
   - { name: set cuda monitoring, set_fact: cudamonitor=true }
   tags: [ always ]
 
-- hosts: 'ComputeNodes'
+- hosts: 'ComputeNodes,DesktopNodes'
   vars_files:
   - vars/slurm.yml
   roles:

CICD/tests/mockSlurmData.yml

+---
+- hosts: ManagementNodes,LoginNodes,ComputeNodes
+  gather_facts: false
+  tasks:
+  - name: add user hpctest
+    user:
+      name: hpctest
+      shell: /bin/bash
+    become: true
+
+- hosts: ManagementNodes
+  gather_facts: false
+  tasks:
+  - name: Create a parent account
+    command: ./sacctmgr -i add account parentAccount cluster=cicd Description="Test parent account" Organization="Monash"
+    args:
+      chdir: '/opt/slurm-latest/bin'
+    become: true
+    register: result 
+    failed_when: result.rc != 0 and result.stdout != " Nothing new added."
+    
+  - name: Create a project associated with a given parent
+    command: ./sacctmgr -i add account testProject parent=parentAccount cluster=cicd Organization="Monash"
+    args:
+      chdir: '/opt/slurm-latest/bin'
+    become: true
+    register: result 
+    failed_when: result.rc != 0 and result.stdout != " Nothing new added."
+    
+  - name: Create a user and associate them with a project
+    command: ./sacctmgr -i create user hpctest cluster=cicd account=testProject partition=batch
+    args:
+      chdir: '/opt/slurm-latest/bin'
+    become: true 
+    register: result 
+    failed_when: result.rc != 0 and result.stdout != " Nothing new added."
+
+  - name: restart slurmctld
+    service:
+      name: slurmctld
+      state: restarted
+    become: true
+
+  #- name: reconfigure scontrol
+  #  command: ./scontrol reconfigure 
+  #  args:
+  #    chdir: '/opt/slurm-latest/bin'
+  #  become: true
+  #  become_user: slurm
+    
+#sudo `which sacctmgr` modify user where name=hpctest set maxjobs=200
+##  18  sudo `which sacctmgr` update account hpctest set qos=normal
+#   22  sudo `which sacctmgr` update account testProject set qos=normal
+
+- hosts: LoginNodes,ComputeNodes,ManagementNodes
+  gather_facts: false
+  tasks:
+  - name: make sure munge is running
+    service:
+      name: munge
+      state: restarted
+    become: true

CICD/tests/slurm/srunHostname.yml

 ---
-- hosts: ManagementNodes,LoginNodes,ComputeNodes
-  gather_facts: false
-  tasks:
-  - name: add user hpctest
-    user:
-      name: hpctest
-      shell: /bin/bash
-    become: true
-
-- hosts: ManagementNodes
-  gather_facts: false
-  tasks:
-  - name: Create a parent account
-    command: ./sacctmgr -i add account parentAccount cluster=cicd Description="Test parent account" Organization="Monash"
-    args:
-      chdir: '/opt/slurm-latest/bin'
-    become: true
-    register: result 
-    failed_when: result.rc != 0 and result.stdout != " Nothing new added."
-    
-  - name: Create a project associated with a given parent
-    command: ./sacctmgr -i add account testProject parent=parentAccount cluster=cicd Organization="Monash"
-    args:
-      chdir: '/opt/slurm-latest/bin'
-    become: true
-    register: result 
-    failed_when: result.rc != 0 and result.stdout != " Nothing new added."
-    
-  - name: Create a user and associate them with a project
-    command: ./sacctmgr -i create user hpctest cluster=cicd account=testProject partition=batch
-    args:
-      chdir: '/opt/slurm-latest/bin'
-    become: true 
-    register: result 
-    failed_when: result.rc != 0 and result.stdout != " Nothing new added."
-
-#sudo `which sacctmgr` modify user where name=hpctest set maxjobs=200
-##  18  sudo `which sacctmgr` update account hpctest set qos=normal
-#   22  sudo `which sacctmgr` update account testProject set qos=normal
-
 - hosts: LoginNodes
   gather_facts: false
   tasks:
-  - name: make sure munge is running
-    service:
-      name: munge
-      state: started
-    become: true
   - name: simple srun test
     command: ./srun --ntasks=1  --partition=batch  hostname
     args: