Simon nfs fix

Fix whitespace issue with nfs client (not sure why this should be a bug)
Fix nfs enable but in nfs-server

See merge request !43

see  https://bugzilla.redhat.com/show_bug.cgi?id=1178720
I have split enable and start commands for redhat

fixed keynames i.e. ansible_host_key_ecdsa_public => ansible_ssh_host_key_ecdsa_public

fixed keynames i.e. ansible_host_key_ecdsa_public => ansible_ssh_host_key_ecdsa_public
ditto for ansible_host_key_rsa_public.
Removed superflous check "and hostvars[node]['ansible_host_key_rsa_public'] "

See merge request !42

diit for ansible_host_key_rsa_public.
Removed superflous check "and hostvars[node]['ansible_host_key_rsa_public'] "

alter the ownership on the directory where slurm stores its pid file

See merge request !41

set the password for the default user enabling console login without keys

See merge request !40

Disable selinux

See merge request !39

Merge branch 'master' of gitlab.erc.monash.edu.au:hpc-team/ansible_cluster_in_a_box into disable_selinux

add a role that just configures gluster and epel on centos to the upstream repos

See merge request !38

allow the playbook to configure if slurmd is started on boot, default to yes

See merge request !37

switch variables from ansible_hostname to inventory_hostname incase ansible is a…

…lso configuring the hostname based on the inventory

See merge request !36

switch variables from ansible_hostname to inventory_hostname incase ansible is also configuring the hostname based on the inventory

Updatelustre

modified lustre.yml and vars/Centos_7.yml for latest INTEL provided centos7 rpm
added file/ibdev2netdev.sh to get Mellanox device name (ie..ens6 or eth1)
template lustre.conf to contain Mellanox device name
modified lustre.yml to push lustre.conf to host OS

See merge request !35

added file/ibdev2netdev.sh to get Mellanox device name (ie..ens6 or eth1)
template lustre.conf to contain Mellanox device name

included latest mellanox driver name in mellanoxVars.yml

Multiple fixes

See merge request !33

speed up the default slurm provisioing by running groups of users rather than adding each user individually

Simons updates

See merge request !32

Conflicts:
	roles/slurm-common/tasks/installNhc.yml

conflicts resolved manually