load hostvars

Author: sjpb

ansible/roles/compute_init/files/compute-init.yml

@@ -8,40 +8,10 @@
     server_node_ip: "{{ os_metadata.meta.k3s_server }}"
     compute_groups: "{{ os_metadata.meta.compute_groups | default([]) }}"
 
-    # TODO: "role defaults"
+    # TODO: "= role defaults" - could be moved to a vars_file: on play with similar precedence effects
+    # this is a good example: common environment actually defines this (non-functional w/o compute groups), but role default is empty
     resolv_conf_nameservers: []
-
-    # nfs_configurations:
-    #   - nfs_export: "/exports/home"
-    #     nfs_client_mnt_options:
-    #     nfs_client_mnt_point: "/home"
-    #     nfs_client_mnt_state: mounted
-    #     nfs_server: "{{ server_node_ip }}"
-
-    # os_manila_mount_state: mounted
-    # os_manila_mount_opts:
-    #   - x-systemd.device-timeout=30
-    #   - x-systemd.mount-timeout=30
-    #   - noatime
-    #   - _netdev # prevents mount blocking early boot before networking available
-    #   - rw
-    # os_manila_mount_ceph_conf_path: /etc/ceph
-
-    # basic_users_manage_homedir: false
-    # basic_users_userdefaults:
-    #   state: present
-    #   create_home: "{{ basic_users_manage_homedir }}"
-    #   generate_ssh_key: "{{ basic_users_manage_homedir }}"
-    #   ssh_key_comment: "{{ item.name }}"
-    # test_user_password: "zXpcWyGQL7jtZnqylQra4g=="
-    # basic_users_users:
-    #   - name: testuser # can't use rocky as $HOME isn't shared!
-    #     password: "{{ test_user_password | password_hash('sha512', 65534 | random(seed=inventory_hostname) | string) }}" # idempotent
-    #     uid: 1005
-    # basic_users_groups: []
-
-    # openhpc_conf_server: "{{ server_node_ip }}"
-
+    
   tasks:
     - block:
         - name: Report skipping initialization if not compute node
@@ -80,3 +50,12 @@
 
         - meta: end_play
       when: _mount_mnt_cluster.failed
+
+    - name: Load hostvars from NFS
+      # this is higher priority than vars block = normal ansible's hostvars
+      include_vars:
+        file: "/mnt/cluster/hostvars/{{ ansible_hostname }}/hostvars.yml" # can't use inventory_hostname
+
+    - name: Demonstrate hostvars have loaded
+      debug:
+        var: prometheus_version

docs/experimental/compute-init.md

@@ -7,8 +7,7 @@ On deploy host:
 
 On compute node:
 
-    [root@rl9-compute-0 rocky]# rm /var/lib/ansible-init.done
-    [root@rl9-compute-0 rocky]# systemctl restart ansible-init
+    [root@rl9-compute-0 rocky]# rm -f /var/lib/ansible-init.done && systemctl restart ansible-init
     [root@rl9-compute-0 rocky]# systemctl status ansible-init
 
 
@@ -97,4 +96,26 @@ Now run site.yml, then restart ansible-init again:
     [root@rl9-compute-0 rocky]# ls /mnt/cluster/hostvars/rl9-compute-
     rl9-compute-0/ rl9-compute-1/ 
     [root@rl9-compute-0 rocky]# ls /mnt/cluster/hostvars/rl9-compute-0/
-    hostvars.yml
+    hostvars.yml
+
+This commit - shows that hostvars have loaded:
+
+    [root@rl9-compute-0 rocky]# systemctl status ansible-init
+    ● ansible-init.service
+        Loaded: loaded (/etc/systemd/system/ansible-init.service; enabled; preset: disabled)
+        Active: active (exited) since Fri 2024-12-13 21:06:20 UTC; 5s ago
+        Process: 27585 ExecStart=/usr/bin/ansible-init (code=exited, status=0/SUCCESS)
+    Main PID: 27585 (code=exited, status=0/SUCCESS)
+            CPU: 8.161s
+
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27769]: TASK [Demonstrate hostvars have loaded] ****************************************
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27769]: ok: [127.0.0.1] => {
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27769]:     "prometheus_version": "2.27.0"
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27769]: }
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27769]: PLAY RECAP *********************************************************************
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27769]: 127.0.0.1                  : ok=5    changed=0    unreachable=0    failed=0    skipped=2    rescued=0    ignored=0
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27585]: [INFO] executing remote playbooks for stage - post
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27585]: [INFO] writing sentinel file /var/lib/ansible-init.done
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid ansible-init[27585]: [INFO] ansible-init completed successfully
+    Dec 13 21:06:20 rl9-compute-0.rl9.invalid systemd[1]: Finished ansible-init.service.
+