This role sets up a slurm cluster along with a database for slurm accounting. If a database exists in your environment then it will try to use that.
ONLY MYSQL/MARIADB or compliant DBs are compatible.
- Have a "slurm" section to your ansible inventory.
- Mark one of the slurm nodes with the variable
slurm_builder=True
, so that the role knows which node to build slurm on. This is also used for other things so ensure it's set in the slurm group. - Mark the headnode with the variable
headnode=True
, so as to avoid deploying the slurm daemon there. THIS ROLE ONLY SUPPORTS ON HEADNODE AS OF NOW, THINGS WILL BREAK WITH MORE - Create a playbook that includes this role. E.g.:
- hosts: all become: true gather_facts: false pre_tasks: - name: Install python2 for Ansible raw: bash -c "test -e /usr/bin/python || (apt -qqy update && apt install -qqy python-minimal)" register: output changed_when: output.stdout != "" - name: Gathering Facts setup: roles: - role: ansible-role-slurm