qa.yml

+#- hosts: 'all'
+  #gather_facts: false # not sure if false is clever here
+  #tasks:
+  #- include_vars: vars/ldapConfig.yml
+  #- include_vars: vars/filesystems.yml
+  #- include_vars: vars/slurm.yml
+  #- include_vars: vars/vars.yml
+  #- { name: set use shared state, set_fact: usesharedstatedir=False }
+  #tags: [ always ]
+
+# this playbook is roughly sorted by
+# - hostgroupstopics like ComputeNodes or ComputeNodes,LoginNodes, last VisNodes
+# - "tag_groups" each starting after a #comment see #misc or misc tag
+- hosts: 'ComputeNodes'
+  gather_facts: false
+  tasks:
+  # these are just templates.
+  #Note the tag never! Everything with never is only executed if called explicitly aka ansible-playbook --tags=foo,bar OR -tags=tag_group
+  - { name: template_shell, shell: ls, tags: [never,tag_group,uniquetag_foo] }
+  - { name: template_command, command: uname chdir=/bin, tags: [never,tag_group,uniquetag_bar] }
+  - { name: template_scipt, script: ./scripts/qa/test.sh, tags: [never,tag_group,uniquetag_script] }
+#mpi stuff
+  - { name: run mpi on one computenode, command: ls, args: {chdir: "/tmp"} , failed_when: "TODO is TRUE", tags: [never,mpi,mpi_local,TODO] }
+  - { name: run mpi on two computenode, command: ls, args: {chdir: "/tmp"} , failed_when: "TODO is TRUE", tags: [never,mpi,mpi_local_two,TODO] }
+  #- { name: run mpi via sbatch, command: cmd=ls chdir="/tmp" , failed_when: "TODO is TRUE", tags: [never,mpi,slurm_mpi,TODO] }
+  #- { name: mpi_pinging,        command: cmd=ls chdir="/tmp" , failed_when: "TODO is TRUE", tags: [never,mpi,mpi_ping,TODO] }
+   #module load openmpi/3.1.6-ucx;mpirun --mca btl self --mca pml ucx -x UCX_TLS=mm  -n 24 /projects/pMOSP/mpi/parallel_mandelbrot/parallel/mandelbrot
+   #module load openmpi/3.1.6-ucx;srun mpirun --mca btl self --mca pml ucx -x UCX_TLS=mm  -n 24 /projects/pMOSP/mpi/parallel_mandelbrot/parallel/mandelbrot
+
+#slurm
+  - { name: slurmd should be running, service: name=slurmd state=started, tags: [never,slurm,slurmd] }
+  - { name: munged should be running, service: name=munged state=started, tags: [never,slurm,munged] }
+  - { name: ensure connectivity to the controller, shell: scontrol ping, tags: [never,slurm,scontrol_ping] }
+  - { name: the most simple srun test, shell: srun --reservation=AWX hostname, tags: [never,slurm,srun_hostname] }
+#nhc, manually run nhc because it contains many tests
+  - { name: run nhc explicitly, command: /opt/nhc-1.4.2/sbin/nhc -c /opt/nhc-1.4.2/etc/nhc/nhc.conf, become: true , tags: [never,slurm,nhc] }
+# networking
+  - { name: ping license server, shell: ls, tags: [never,network,ping_license] }
+  - { name: ping something outside monash, command: ping -c 1 8.8.8.8, tags: [never,network,ping_external] }
+#mounts
+- hosts: 'ComputeNodes,LoginNodes'
+  gather_facts: false
+  tasks:
+  - { name: check mount for usr_local, shell: "mount | grep -q local", tags: [never,mountpoints,mountpoints_local] }
+  - { name: check mount for projects, shell: "lfs df -h", tags: [never,mountpoints_projects] }
+  - { name: check mount for home, shell: "mount | grep -q home", tags: [never,mountpoints,mountpoints_home] }
+  - { name: check mount for scratch, shell: "mount | grep -q scratch" , tags: [never,mountpoints_scratch] }
+#misc
+  - { name: check singularity, shell: module load octave && octave --version, tags: [never,misc,singularity3] }
+  - { name: module test, shell: cmd="module load gcc" executable="/bin/bash", tags: [never,misc,modulecmd] }
+  - { name: contact ldap, shell: maybe test ldapsearch, failed_when: "TODO is TRUE", tags: [never,misc,ldap,TODO] }
+#gpu
+- hosts: 'VisNodes'
+  gather_facts: false
+  tasks:
+  - { name: run nvida-smi to see if a gpu driver is present, command: "/bin/nvidia-smi", tags: [never,gpu,smi] }
+  - { name: run gpu burn defaults to 30 seconds, command: "/usr/local/gpu_burn/1.0/run_silent.sh", tags: [never,gpu,long,gpuburn] }
+
+
+# extended time-consuming tests
+# relion see https://docs.massive.org.au/communities/cryo-em/tuning/tuning.html
+# linpack
+
+#module load openmpi/1.10.7-mlx;ldd /usr/local/openmpi/1.10.7-mlx/bin/* | grep -ic found

roles/MonashBioinformaticsPlatform_node_allocation/files/get_node.py

+#!/usr/bin/python
+import subprocess
+import sys
+
+def getTime():
+    print "How long do you think you need this computer for?"
+    print "If you need the computer for 2 days and 12 hours please enter as 2-12 or 2-12:00:00"
+    time=sys.stdin.readline().strip()
+    try:
+        (days,hours)=time.split('-')
+    except:
+        days=0
+        hours=time
+    try:
+        (hours,minues) = time.split(':')
+    except:
+        pass
+    return (days,hours)
+
+def getNCPUs():
+    print "How many CPUs would you like?"
+    cpus=None
+    while cpus==None:
+        cpustr=sys.stdin.readline().strip()
+        try:
+            cpus=int(cpustr)
+        except:
+            print "Sorry I can't interpret %s as a number"%cpustr
+            print "How many CPUs would you like?"
+
+    return cpus
+
+def getRAM():
+    print "How much RAM would you like (press enter for the default)?"
+    ramstr= sys.stdin.readline().strip()
+    while ramstr!=None and ramstr!="":
+            try:
+                ram=int(ramstr)
+                return ram
+            except:
+                print "Sorry I can't interpret %s as a number"%ramstr
+                print "How much RAM would you like?"
+                ramstr= sys.stdin.readline()
+    return None
+
+def subjob(time,cpus,ram):
+    if ram==None:
+        ram=cpus*2000
+    import subprocess
+    scriptpath='/home/chines'
+    p=subprocess.Popen(['sbatch','--time=%s-%s'%(time[0],time[1]),'--nodes=1','--mincpu=%s'%cpus,'--mem=%s'%ram,'%s/mbpjob.sh'%scriptpath],stdout=subprocess.PIPE,stderr=subprocess.PIPE)
+    (stdout,stderr)=p.communicate()
+    import re
+    m=re.match('Submitted batch job (?P<jobid>[0-9]+)',stdout)
+    if m:
+        return m.groupdict()['jobid']
+
+def isState(jobid,state='RUNNING'):
+    import re
+    p=subprocess.Popen(['scontrol','show','job','-d',jobid],stdout=subprocess.PIPE,stderr=subprocess.PIPE)
+    (stdout,stderr)=p.communicate()
+    jobidre=re.compile('JobId=(?P<jobid>[0-9]+)\s')
+    statere=re.compile('^\s+JobState=(?P<state>\S+)\s')
+    currentjobid=None
+    for l in stdout.splitlines():
+        m=jobidre.match(l)
+        if m:
+            currentjobid=m.groupdict()['jobid']
+        m=statere.match(l)
+        if m:
+            if m.groupdict()['state']==state:
+                if jobid==currentjobid:
+                    return True
+            else:
+                if jobid==currentjobid:
+                    return False
+    return False
+
+def waitjob(jobid):
+    import time
+    while True:
+        if isState(jobid,'RUNNING'):
+            return
+        else:
+            print "job %s not running"%jobid
+            time.sleep(1)
+
+def listJobs():
+    import re
+    r=[]
+    user = subprocess.check_output(['whoami'])
+    jobs = subprocess.check_output(['squeue','-u',user,'-h','-o','"%i %L %j %c"'])
+    jobre=re.compile("(?P<jobid>(?P<jobidNumber>[0-9]+)) (?P<time>\S+ (?P<jobname>\S+) (?P<cpus>[0-9]+))$"
+    for l in jobs.splitlines():
+        m=jobidre.search(l)
+        if m:
+            r.append(m.groupdict())
+    return r
+
+def getNode(jobid):
+    import re
+    stdout=subprocess.check_output(['scontrol','show','job','-d',jobid])
+    for l in stdout.splitlines():
+        m=re.search('^\s+Nodes=(?P<nodelist>\S+)\s',l)
+        if m:
+            nodes=m.groupdict()['nodelist'].split(',')
+            return nodes[0]
+
+def createJob(*args,**kwargs):
+    time=getTime()
+    #cpus=getNCPUs()
+    cpus=1
+    #ram=getRAM()
+    ram=None
+    subjob(time,cpus,ram)
+
+def selectJob(jobidlist):
+    if len(jobidlist)==1:
+        return jobidlist[0]['jobid']
+    else:
+        print "Please select a job (or press enter to cancel)"
+        i=1
+        print "\tJob name\tNum CPUs\tRemaining Time"
+        for j in jobidlist:
+            print "%s\t%s\t%s\t%s"%(i,j['jobname'],j['numcpus'],j['time'])
+        try:
+            jobnum=int(sys.stdin.readline().strip())
+            if (jobnum>0 and jobnum<=jobidlist):
+                return jobidlist[jobnum-1]['jobid']
+        except:
+            pass
+    return None
+
+
+def connect(*args,**kwargs):
+    jobidlist=listJobs()
+    jobid=selectJob(jobidlist)
+    if jobid!=None:
+        waitjob(jobid)
+        node=getNode(jobid)
+        print node
+
+def stop(*args,**kwargs):
+    jobidlist=listJobs()
+    jobid=selectJob(jobidlist)
+    if jobid!=None:
+        stopjob(jobid)
+
+
+
+def main():
+    import argparse
+    parser = argparse.ArgumentParser()
+    subparser = parser.add_subparsers()
+
+    start = subparser.add_parser('start', help='alloate a node to the user')
+    start.set_defaults(func=createJob)
+
+    connect = subparser.add_parser('connect')
+    start.set_defaults(func=connect)
+
+    stop = subparser.add_parser('stop')
+    start.set_defaults(func=stop)
+        
+    args = parser.parse_args()
+    args.func(args)
+
+try:
+    jobidlist=listJobs()
+    if len(jobidlist)>1:
+        print "cancel all jobs here"
+    jobidlist=listJobs()
+    if len(jobidlist)==0:
+        time=getTime()
+        #cpus=getNCPUs()
+        cpus=1
+        #ram=getRAM()
+        ram=None
+        subjob(time,cpus,ram)
+    jobidlist=listJobs()
+    if len(jobidlist)==1:
+        jobid=jobidlist[0]
+        waitjob(jobid)
+        node=getNode(jobid)
+        print node
+        sys.exit(0)
+except Exception as e:
+    print e
+    import traceback
+    print traceback.format_exc()
+    sys.exit(1)
+
+main()

roles/MonashBioinformaticsPlatform_node_allocation/files/mbp_node

+#!/bin/bash 
+mpbctrl='/home/hines/mbp_script/get_node.py'
+node=$( $mbpctrl $1 )
+if [[ $node ]]; then
+    ssh -t $node tmux attach-session
+fi

roles/MonashBioinformaticsPlatform_node_allocation/tasks/main.yml

+---
+- name: make sure /usr/local/bin exists
+  file: path=/usr/local/bin state=directory mode=755 owner=root
+  become: true
+
+- name: install get_node.py
+  copy: src=get_node.py dest=/usr/local/bin/get_node.py mode=755 owner=root
+  become: true
+
+- name: install mbp_node
+  copy: src=mbp_node dest=/usr/local/bin/mbp_node mode=755 owner=root
+  become: true

roles/OSImageBugfix/tasks/main.yml

+---
+# This role is to  fix a misconfiguration of some OpenStack Base images at Monash University. 
+# the misconfiguration is dev/vdb mounted in fstab of the Image and the Openstack Flavour not providing a second disk.
+- name: unmount vdb if absent
+  mount:
+    path: "/mnt"
+    src: "/dev/vdb"
+    state: absent
+  become: true
+  when: 'hostvars[inventory_hostname]["ansible_devices"]["vdb"] is not defined'
+
+- name: keep mnt present
+  file:
+    path: "/mnt"
+    owner: root
+    group: root
+    mode: "u=rwx,g=rx,o=rx"
+    state: directory
+  become: true
+  when: 'hostvars[inventory_hostname]["ansible_devices"]["vdb"] is not defined'

roles/OpenVPN-Client/handlers/main.yml

 ---
   - name: restart openvpn
     service: name=openvpn state=restarted
-    sudo: true
+    become: true

roles/OpenVPN-Client/tasks/copyCerts.yml

---- 
-- 
+---
+-
   copy: "src=/tmp/{{ inventory_hostname }}/ca.crt dest=/etc/openvpn/ca.crt  mode=644 owner=root group=root"
   name: "Copying CA certificate"
   when: "client_ca_cert.stat.exists == false"
-- 
+-
   copy: "src=/tmp/{{ inventory_hostname }}/{{ inventory_hostname }}.crt dest=/etc/openvpn/{{ inventory_hostname }}.crt mode=644 owner=root group=root"
   name: "Copying Client certificate"
   when: "client_sign_cert.stat.exists == false"
-- 
+-
   copy: "src=/tmp/{{ inventory_hostname }}/{{ inventory_hostname }}.key dest=/etc/openvpn/{{ inventory_hostname }}.key  mode=600 owner=root group=root"
   name: "Copying Client key"
   when: "client_key.stat.exists == false"

roles/OpenVPN-Client/tasks/installOpenVPN.yml

---- 
+---
 - name: "Install OpenVPN"
   yum: "name=openvpn state=present"
-  sudo: true
+  become: true
   notify: restart openvpn
 
 - name: "Copying client.conf to the OpenVPN client"
   template: "src=client.conf.j2 dest=/etc/openvpn/client.conf"
-  sudo: true
+  become: true
   notify: restart openvpn
 

roles/OpenVPN-Client/tasks/main.yml

@@ -3,6 +3,6 @@
  include: installOpenVPN.yml
 
 - name: "Start OpenVPN"
-  service: name=openvpn state=started
-  sudo: true
+  service: name=openvpn state=started enabled=yes
+  become: true
 

roles/OpenVPN-Client/vars/main.yml

+readme.txt
\ No newline at end of file

roles/OpenVPN-Server/handlers/main.yml

 ---
   - name: restart openvpn
     service: name=openvpn state=restarted
-    sudo: true
+    become: true

roles/OpenVPN-Server/tasks/installOpenVPN.yml

---- 
+---
 - name: "Install OpenVPN"
   yum: "name=openvpn state=present"
   notify: "restart openvpn"
-  sudo: true
+  become: true
 
 - name: Create path
   shell: mkdir -p {{ dhparms_file | dirname }}
   args:
     creates: "{{ dhparms_file | dirname }}"
-  sudo: true
+  become: true
 
 - name: "Generate DH parameters"
   shell: openssl dhparam -out {{ dhparms_file }} 512
   args:
     creates: "{{ dhparms_file }}"
-  sudo: true
+  become: true
 
 - name: "Configure OpenVPN Server"
   template: "src=server.conf.j2 dest=/etc/openvpn/server.conf"
   notify: "restart openvpn"
-  sudo: true
+  become: true

roles/OpenVPN-Server/tasks/main.yml

@@ -3,5 +3,5 @@
  include: installOpenVPN.yml
 
 - name: "Start OpenVPN"
-  service: name=openvpn state=started
-  sudo: true
+  service: name=openvpn state=started enabled=yes
+  become: true

roles/OpenVPN-Server/vars/main.yml

+readme.txt
\ No newline at end of file

roles/SSHKnownHosts/tasks/main.yml

+- name: install known hosts file
+  copy: src=files/ssh_known_hosts dest=/etc/ssh/ssh_known_hosts owner=root mode=644
+  become: true
+  become_user: root

roles/additional_paths/tasks/main.yml

+- name: setup additiona PATHs in /etc/profile.d
+  template:
+    src: additional_paths.sh.j2
+    dest: /etc/profile.d/additional_paths.sh
+  become: true
+  when: additional_paths is defined

roles/additional_paths/templates/additional_paths.sh.j2

+export PATH=$PATH:{{ additional_paths|join(":") }}

roles/allow_stale_nfs/tasks/main.yml

+---
+- name: place /usr/local/ last in the PATH in /etc/profile
+  lineinfile:
+  args:
+    dest: "/etc/profile"
+    insertbefore: BOF
+    line: "PATH=/bin:/usr/bin:/usr/local/bin"
+  become: true
+  become_user: root
+
+- name: remove old line
+  lineinfile:
+  args:
+    dest: "/etc/profile"
+    regexp: "^PATH=/usr/local/bin:/bin:/usr/bin$"
+    state: absent
+  become: true
+  become_user: root
+
+- name: remove /usr/local/ from the PATH in /etc/profile
+  lineinfile:
+  args:
+    dest: "/etc/profile"
+    regexp: ".*pathmunge /usr/local.*"
+    state: absent
+  become: true
+  become_user: root
+
+- name: dont execute abrt-cli on login
+  file: path=/etc/profile.d/abrt-console-notification.sh state=absent
+  become: true
+  become_user: root

roles/apache2/tasks/apacheDebian.yml

+---
+-
+ name: "Install Apache2"
+ apt: name={{ item }} state=present
+ with_items:
+  - apache2
+  - apache2-dev
+ become: true
+
+-
+ name: "Templating default-ssl site"
+ template: src=default-ssl.j2 dest=/etc/apache2/sites-available/default-ssl.conf owner=www-data group=www-data
+ become: true
+-
+ name: "Templating default site"
+ template: src=default.j2 dest=/etc/apache2/sites-available/000-default.conf owner=www-data group=www-data
+ become: true
+
+-
+ name: "Enable ssl module"
+ apache2_module: state=present name=ssl
+ become: true
+
+-
+ name: "Enable default-ssl site"
+ shell: a2ensite default-ssl
+ become: true
+ notify: restart apache2
+
+

roles/apache2/tasks/apacheRedHat.yml

+---
+
+-
+ name: "Installing Apache"
+ become: true
+ yum: name={{ item }} state=present
+ with_items:
+  - mod_ssl
+  - mod_wsgi
+  - openssl
+  - httpd
+  - httpd-devel
+-
+ name: Setting httpd.conf
+ become: true
+ replace: dest=/etc/httpd/conf/httpd.conf regexp="^#ServerName www.example.com:80" replace="ServerName {{ ansible_fqdn }}"
+
+-
+ name: "Templating default-ssl site"
+ template: src=default-ssl.j2 dest=/etc/httpd/conf.d/ssl.conf owner=apache group=apache
+ become: true
+
+-
+ name: Templating wsgi.conf
+ become: true
+ template: src=wsgi.conf.j2 dest=/etc/httpd/conf.d/wsgi.conf owner=root group=root
+-
+ name: Restarting Apache
+ become: true
+ service: name=httpd state=restarted