Make munge daemon start depend on cloud-init final finishing

meta/main.yml

@@ -11,8 +11,9 @@ galaxy_info:
   min_ansible_version: 2.5.0
   platforms:
     - name: EL
-      versions:
+      release:
         - 7
+        - 8
   galaxy_tags:
     - openhpc
     - slurm

tasks/drain.yml

@@ -20,5 +20,5 @@
   register: drained_nodes
   until: "inventory_hostname in drained_nodes.stdout_lines"
   delay: "{{ openhpc_retry_delay }}"
-  retries: "{{ (openhpc_drain_timeout/openhpc_retry_delay) | int }}"
+  retries: "{{ (openhpc_drain_timeout / openhpc_retry_delay) | int }}"
   changed_when: false

tasks/install.yml

@@ -46,6 +46,25 @@
     name: "{{ openhpc_slurm_pkglist | reject('eq', '') }}"
   when: openhpc_slurm_pkglist | default(false, true)
 
+- name: Create drop-in directory for munge service
+  ansible.builtin.file:
+    path: /etc/systemd/system/munge.service.d
+    state: directory
+    owner: root
+    group: root
+    mode: 0644
+
+- name: Delay munge start till after cloud-init
+  # in case mungekey is injected by cloud-init
+  copy:
+    dest: /etc/systemd/system/munge.service.d/slurmapp.conf
+    content: |
+      [Unit]
+      After=cloud-final.service
+    owner: root
+    group: root
+    mode: 0644
+
 - name: Install packages from openhpc_packages variable
   yum:
     name: "{{ openhpc_packages }}"

tasks/resume.yml

@@ -20,5 +20,5 @@
   register: resumed_nodes
   until: "inventory_hostname in resumed_nodes.stdout_lines"
   delay: "{{ openhpc_retry_delay }}"
-  retries: "{{ (openhpc_resume_timeout/openhpc_retry_delay) | int }}"
+  retries: "{{ (openhpc_resume_timeout / openhpc_retry_delay) | int }}"
   changed_when: false

tasks/runtime.yml

@@ -164,7 +164,7 @@
     enabled: "{{ openhpc_slurm_service_enabled | bool }}"
     state: "{{ 'started' if openhpc_slurm_service_started | bool else 'stopped' }}"
 
-- name: flush handler
+- name: Flush handler
   meta: flush_handlers # as then subsequent "ensure" is a no-op if slurm services bounced
 
 - name: Ensure slurmdbd is started and running