Merge branch 'master' into 'master'

Master

Mellanx drivers reboot did not work. Converted "command:" to "shell:"
Need new drivers for new kernel version

See merge request !6

roles/lustre-client/vars/CentOS_7.yaml

 ---
 lustre_pkgs:
-  - lustre-client-modules-2.7.0-3.10.0_229.14.1.el7.x86_64.x86_64.rpm
-  - lustre-client-2.7.0-3.10.0_229.14.1.el7.x86_64.x86_64.rpm
+  # old rmps for older kernel
+  #- lustre-client-modules-2.7.0-3.10.0_229.14.1.el7.x86_64.x86_64.rpm
+  #- lustre-client-2.7.0-3.10.0_229.14.1.el7.x86_64.x86_64.rpm
+  #simon comment out
+  #-  lustre-client-modules-2.7.0-3.10.0_229.20.1.el7.x86_64.x86_64.rpm
+  #-  lustre-client-2.7.0-3.10.0_229.20.1.el7.x86_64.x86_64.rpm
+  # shahahh mods
+  -  lustre-client-modules-2.7.65-3.10.0_327.4.4.el7.x86_64_gab38c3a.x86_64.rpm
+  -  lustre-client-2.7.65-3.10.0_327.4.4.el7.x86_64_gab38c3a.x86_64.rpm
 

roles/mellanox_drivers/tasks/main.yml

 ---
 - include_vars: mellanoxVars.yml
 
-
 - name: yum install dependencies
   yum:  name=perl,pciutils,gtk2,atk,cairo,gcc-gfortran,libxml2-python,tcsh,libnl,lsof,tcl,tk
   sudo: true
@@ -105,9 +104,9 @@
 # A REBOOT IS NEEDED AFTER SUCCESSFUL INSTALL
 #
 - name: restart machine
-  command: "sleep 5; sudo shutdown -r now"
+  shell: "sleep 5; sudo shutdown -r now"
   async: 2
-  poll: 0
+  poll: 1
   ignore_errors: true
   sudo: true
   when: ansible_os_family=="RedHat" and  drivers_installed|failed

roles/mellanox_drivers/vars/mellanoxVars.yml

 ---
  #note. do not add '.tgz' to driver src. done in playbook
- MELLANOX_DRIVER_SRC: "{% if ansible_os_family == 'RedHat'  %}MLNX_OFED_LINUX-3.1-1.0.3-rhel7.1-x86_64-ext{% elif ansible_os_family == 'Debian' %}MLNX_OFED_LINUX-3.1-1.0.3-ubuntu14.04-x86_64{% endif %}"
+ #MELLANOX_DRIVER_SRC: "{% if ansible_os_family == 'RedHat'  %}MLNX_OFED_LINUX-3.1-1.0.3-rhel7.1-x86_64-ext{% elif ansible_os_family == 'Debian' %}MLNX_OFED_LINUX-3.1-1.0.3-ubuntu14.04-x86_64{% endif %}"
+ MELLANOX_DRIVER_SRC: "{% if ansible_os_family == 'RedHat'  %}MLNX_OFED_LINUX-3.1-1.0.3-rhel7.2-x86_64-ext{% elif ansible_os_family == 'Debian' %}MLNX_OFED_LINUX-3.1-1.0.3-ubuntu14.04-x86_64{% endif %}"
  MELLANOX_DEVICE_NAME: "{% if ansible_os_family == 'RedHat'  %}ens6{% elif  ansible_os_family == 'Debian' %}eth1{% endif %}"

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

@@ -58,6 +58,9 @@ CompleteWait=10
 #UsePAM=
 #
 # TIMERS
+
+SlurmctldTimeout=3000 #added due to network failures causing jobs to be killed
+
 #SlurmctldTimeout=300
 #SlurmdTimeout=300
 #InactiveLimit=0
@@ -150,5 +153,5 @@ NodeName={{ node }} Procs={{ hostvars[node]['ansible_processor_vcpus'] }} RealMe
 {% endfor %}
 
 {% for queue in slurmqueues %}
-PartitionName={{ queue.name }} {% if queue.default %}Default=yes{% endif %} Nodes={{ groups[queue.group]|join(',') }} DefaultTime=24:00:00 State=UP 
+PartitionName={{ queue.name }} {% if queue.default %}Default=yes{% endif %} Nodes={{ groups[queue.group]|join(',') }} DefaultTime=72:00:00 State=UP 
 {% endfor %}