stackhpc / ansible-slurm-appliance

A Slurm-based HPC workload management environment, driven by Ansible.
36 stars 15 forks source link

Use project manila share user from Azimuth #359

Closed sjpb closed 4 months ago

sjpb commented 4 months ago

Changes CaaS project manila shares to use the share user provided by Azimuth (in https://github.com/stackhpc/azimuth/pull/148/files), rather assuming there is only a single access rule. Fixes cases where multiple access rules are defined on the project share.

sjpb commented 4 months ago

With azimuth-config on f06894dc56f3f09d76cb7ac3e59f0e3c5c240042 and ci environment configured as shown below this failed to deploy in rcp-cloud-portal-demo with

Failed to create platform. To retry please click patch. Possible reason for the failure was: non-zero return code

config:

# Set the volume-type name for fast volumes
arcus_fast_volume_type: arcus-ceph01-nvme-rbd

# Enable the project Manila share in Azimuth
azimuth_openstack_manila_project_share_gb: 40

# Enable the fast-volume-type variants of the SSH and standard workstation platforms
azimuth_caas_stackhpc_workstation_metadata_url: >-
  https://raw.githubusercontent.com/stackhpc/caas-workstation/{{ azimuth_caas_stackhpc_workstation_git_version }}/ui-meta/workstation-fast-volume-type.yml
azimuth_caas_stackhpc_workstation_ssh_metadata_url: >-
  https://raw.githubusercontent.com/stackhpc/caas-workstation/{{ azimuth_caas_stackhpc_workstation_git_version }}/ui-meta/workstation-external-ssh-fast-volume-type.yml
azimuth_caas_stackhpc_workstation_extra_vars_overrides:
  cluster_volume_type_fast: "{{ arcus_fast_volume_type }}"

# Slurm appliance: home on fast volume
azimuth_caas_stackhpc_slurm_appliance_git_version: feat/caas-manila-share-user
azimuth_caas_stackhpc_slurm_appliance_metadata_url: >-
  https://raw.githubusercontent.com/stackhpc/ansible-slurm-appliance/{{ azimuth_caas_stackhpc_slurm_appliance_git_version }}/environments/.caas/ui-meta/slurm-infra-fast-volume-type.yml
azimuth_caas_stackhpc_slurm_appliance_extra_vars_overrides:
  home_volume_type_fast: "{{ arcus_fast_volume_type }}"
  cluster_home_manila_share: false

# Slurm appliance: home on manila
azimuth_caas_cluster_templates_overrides:
  slurm-project-manila:
    gitUrl: "{{ azimuth_caas_stackhpc_slurm_appliance_git_url }}"
    gitVersion: "{{ azimuth_caas_stackhpc_slurm_appliance_git_version }}"
    uiMetaUrl: >-
      https://raw.githubusercontent.com/stackhpc/ansible-slurm-appliance/{{ azimuth_caas_stackhpc_slurm_appliance_git_version }}/environments/.caas/ui-meta/slurm-infra-manila-home.yml
    playbook: "{{ azimuth_caas_stackhpc_slurm_appliance_playbook }}"
    jobTimeout: "{{ azimuth_caas_stackhpc_slurm_appliance_job_timeout_seconds }}"
    extraVars:
      cluster_image: "{{ community_images_image_ids.openhpc }}"
      cluster_home_manila_share: true # is default if cluster_project_manila_share anyway
      cluster_home_manila_share_type: ceph01_cephfs
    jobTimeout: "{{ azimuth_caas_stackhpc_slurm_appliance_job_timeout_seconds }}"
    envVars:
      ANSIBLE_INVENTORY: environments/common/inventory,environments/.caas/inventory