Added a conditional test to the compute_create script to exit if the user does not provide a compute node name as the argument

Added code to download the ipxe.iso file if missing

See merge request jpr/ohpc_vagrant!15

Copr repo support for custom RPMS
Remove explict slurm patch for 17.11.11 now accessible via copr

Update readme for use with ood

See merge request jpr/ohpc_vagrant!14

Simplify the project cloning instructions
Clarify motivation of project providing links to upstream
Add instructions to create Open OnDemand node

Fix to force the slurm upgrade during provisioning
since it's required for working ood dev cluster
Enhance wwgetfiles to not delay during ood provision.

Get the latest project hash that fixes EB bootstraping
on ohpc master.

* Support use of CentOS maintained Vagrant boxes (kernel and interface name changes)
* Fix broken cluster interactive desktop on OOD due to broken slurm-17.11.10
in OHPC 1.3.6 from November 2018 by replacing with slurm-17.11.11
* Refactor OOD roles to separate one-time setup from ops-like tasks
* Fix compute image support for http proxy port selection (nc)
* Enable EasyBuild install for applications testing

Feat change vagrant box to pristine centos to simplify tracking upstream releases by using upstream maintained vagrant box.

See merge request !12

Move insert_key directive before VM create to avoid custom key create
Add comments motivating the config choices for ssh config steps
Expand user key authz logic to append key rather than just replace all
existing keys in vagrant user authorized_keys file.

Patch easybuild installation fix

See merge request jpr/ohpc_vagrant!10

Set insert_key to false, prevent vagrant from generating new key
then set our own ssh private key to be used, first priority
finally, copy our own ssh public key into vm as authorized key

'Dont run bootstrap as root' problem
  while running eb install ansible task
Simply put -b option(become) to let ansible run the bootstrap script as build user

* Update the ohpc cluster submodule for support of VNC on the
cluster compute node vnfs image and activates interactive desktops in
the ood UI.
* Fix #7 hang of ood during reboot due to invalid selinux policy

Resolve "post ansible staging of slurm.conf"

Closes #6

See merge request jpr/ohpc_vagrant!8

This configures the ood node to be a part of the slurm cluster.
It enables job execution, control and shell access to the headnode.

It turned out to be easy enough to add warewulf synchronization to
the ood node in cluster rather than use Vagrant to move files. This
also ensure future slurm.conf and other registered file changes are
maintained across cluster nodes after changes.

Note: `vagrant ssh ood "mount -a"` must be manually run after provision
to sync /home across cluster.

Fixes #6

Update submodule for upstream post merge commit.

Resolve "feature: add support for ood"

Closes #3

See merge request jpr/ohpc_vagrant!6

This introduces permissive selinux policy state and
avoids a reboot.

We need to be able to test both http and https configurations
since https with be the default for ood (http is only for dev).
Also remove the unused host rename config.

This is needed so that the ansible playbook is correctly
selected based on hostname.

Chose high number IP address to avoid DHCP address conflict with
warewulf. This range should be marked off eventually.

The existing ansible line works for both ohpc and ood based on the
hostname and role configuration.

Resolve "ansible for create slurm cluster fails"

See merge request jpr/ohpc_vagrant!4

Resolve "ansible for create slurm cluster fails"

Closes #1

See merge request !3

Signficant changes:
- Change cluster name to match vagrant hostname
- Fix sacct configuration to use site vars and template file

The default name for the first node in this cluster is c0.
Would c1 be a more natural start?