From cafb8d5e1af143c83ca3dfe239e50d8bbc5f6ac5 Mon Sep 17 00:00:00 2001
From: Pierre Riteau <pierre@stackhpc.com>
Date: Fri, 29 Nov 2024 23:34:41 +0100
Subject: Fix nova-compute startup after reprovisioning

We need to restore the /var/lib/nova/compute_id file to avoid
nova-compute failing to start with the following error:

    nova.exception.InvalidConfiguration: No local node identity found,
    but this is not our first startup on this host. Refusing to start
    after potentially having lost that state!

K-A now query IDs of hypervisors by their hostname and create necessary
compute_id file under node config directory for nova-compute before
starting nova-compute containers.
The compute_id file gets copied under /var/lib/nova/ on container start
up.

Co-Authored-By: bjorn lofdahl <bjorn.lofdahl@gmail.com>
Co-Authored-By: Seunghun Lee <seunghun@stackhpc.com>
Co-Authored-By: Owen Jones <owen@stackhpc.com>
Change-Id: If635577fa240f0953ad02d2c28473480d9ad011d
Closes-Bug: #2051011
Signed-off-by: Owen Jones <owen@stackhpc.com>
Signed-off-by: Maksim Malchuk <maksim.malchuk@gmail.com>
(cherry picked from commit 3e4cfd49da38c7db2ac3ad47a75dc89b14cbb513)
---
 ansible/roles/nova-cell/tasks/config.yml           | 60 ++++++++++++++++++++++
 .../roles/nova-cell/templates/nova-compute.json.j2 |  6 +++
 .../notes/bug-2051011-e78146d498a5875a.yaml        |  6 +++
 3 files changed, 72 insertions(+)
 create mode 100644 releasenotes/notes/bug-2051011-e78146d498a5875a.yaml

diff --git a/ansible/roles/nova-cell/tasks/config.yml b/ansible/roles/nova-cell/tasks/config.yml
index f9e9c10ee..83569eedd 100644
--- a/ansible/roles/nova-cell/tasks/config.yml
+++ b/ansible/roles/nova-cell/tasks/config.yml
@@ -54,6 +54,66 @@
   when:
     - vendordata_file.stat.exists
 
+- name: Ensure compute_id exists in nova_compute volume
+  vars:
+    service: "{{ nova_cell_services['nova-compute'] }}"
+    hypervisor_nodename: "{{ hostvars[inventory_hostname].ansible_facts.nodename }}"
+  when: service | service_enabled_and_mapped_to_host
+  block:
+    - name: List all hypervisors
+      kolla_toolbox:
+        container_engine: "{{ kolla_container_engine }}"
+        module_name: openstack.cloud.resources
+        module_args:
+          auth: "{{ openstack_auth }}"
+          cloud: "{{ openstack_auth_cloud }}"
+          ca_cert: "{{ openstack_cacert }}"
+          region_name: "{{ openstack_region_name }}"
+          endpoint_type: "{{ openstack_interface }}"
+          service: "compute"
+          type: "hypervisor"
+      register: hypervisors
+      check_mode: false
+      become: true
+      delegate_to: "{{ groups[nova_cell_conductor_group][0] }}"
+
+    - name: Match current hypervisor
+      ansible.builtin.set_fact:
+        matched_hypervisors: >-
+          {{ hypervisors.resources | selectattr('name', 'equalto', hypervisor_nodename) }}
+      when: hypervisors.resources | length > 0
+
+    - name: Assert that only one hypervisor was matched
+      ansible.builtin.assert:
+        that:
+          - matched_hypervisors | length == 1
+        fail_msg: |
+          "Found multiple compute services with same host name, please check your Nova "
+          "compute agents list and fix it. The query with hypervisor name "
+          "'{{ hypervisor_nodename }}' returned following: {{ matched_hypervisors }}"
+      when:
+        - hypervisors.resources | length > 0
+        - matched_hypervisors | length > 0
+
+    - name: Copy the matched hypervisor UUID to compute_id
+      ansible.builtin.copy:
+        content: "{{ matched_hypervisors[0].id }}"
+        dest: "{{ node_config_directory }}/nova-compute/compute_id"
+        mode: "0660"
+      become: true
+      when:
+        - hypervisors.resources | length > 0
+        - matched_hypervisors | length == 1
+
+    # If the host was never registered, create a new hypervisor id and copy as compute_id
+    - name: Generate a new UUID then copy to compute_id
+      ansible.builtin.copy:
+        content: "{{ hypervisor_nodename | to_uuid }}"
+        dest: "{{ node_config_directory }}/nova-compute/compute_id"
+        mode: "0660"
+      become: true
+      when: hypervisors.resources | length == 0 or matched_hypervisors | length == 0
+
 - name: Copying over config.json files for services
   become: true
   ansible.builtin.template:
diff --git a/ansible/roles/nova-cell/templates/nova-compute.json.j2 b/ansible/roles/nova-cell/templates/nova-compute.json.j2
index 400301d21..b90367a42 100644
--- a/ansible/roles/nova-cell/templates/nova-compute.json.j2
+++ b/ansible/roles/nova-cell/templates/nova-compute.json.j2
@@ -6,6 +6,12 @@
             "dest": "/etc/nova/nova.conf",
             "owner": "nova",
             "perm": "0600"
+        },
+        {
+            "source": "{{ container_config_directory }}/compute_id",
+            "dest": "/var/lib/nova/compute_id",
+            "owner": "nova",
+            "perm": "0600"
         }{% if nova_policy_file is defined %},
         {
             "source": "{{ container_config_directory }}/{{ nova_policy_file }}",
diff --git a/releasenotes/notes/bug-2051011-e78146d498a5875a.yaml b/releasenotes/notes/bug-2051011-e78146d498a5875a.yaml
new file mode 100644
index 000000000..1195b2d6e
--- /dev/null
+++ b/releasenotes/notes/bug-2051011-e78146d498a5875a.yaml
@@ -0,0 +1,6 @@
+---
+fixes:
+  - |
+    Fixes ``nova-compute`` failing to start after ``/var/lib/nova`` content has
+    been wiped, for example following a reprovisioning.
+    `LP#2051011 <https://launchpad.net/bugs/2051011>`__
-- 
cgit v1.2.3