Kayobe 14.1, OpenStack 2023.1
Had a test environment spun up, and wanted to wipe it again. Ran service destroy, and it failed at this point:
TASK [inspection-store : Check whether inspection store volumes are present] ***********************************************************************************************************************************************************
task path: /home/ubuntu/venvs/kayobe/share/kayobe/ansible/roles/inspection-store/tasks/destroy.yml:8
Monday 22 January 2024 13:06:20 +0100 (0:00:00.629) 0:00:02.088 ********
skipping: [control0] => (item=[{'container_name': 'inspection_store', 'enabled': False, 'image': 'nginx:stable', 'network_mode': 'host'}, '/etc/localtime:/etc/localtime:ro']) => changed=false
ansible_loop_var: item
item:
- container_name: inspection_store
enabled: false
image: nginx:stable
network_mode: host
- /etc/localtime:/etc/localtime:ro
skip_reason: Conditional result was False
skipping: [control0] => (item=[{'container_name': 'inspection_store', 'enabled': False, 'image': 'nginx:stable', 'network_mode': 'host'}, '/home/ubuntu/etc/inspection-store/nginx.conf:/etc/nginx/nginx.conf:ro']) => changed=false
ansible_loop_var: item
item:
- container_name: inspection_store
enabled: false
image: nginx:stable
network_mode: host
- /home/ubuntu/etc/inspection-store/nginx.conf:/etc/nginx/nginx.conf:ro
skip_reason: Conditional result was False
failed: [control0] (item=[{'container_name': 'inspection_store', 'enabled': False, 'image': 'nginx:stable', 'network_mode': 'host'}, 'inspection_store:/data']) => changed=false
ansible_loop_var: item
cmd:
- docker
- volume
- inspect
- inspection_store
delta: '0:00:00.017810'
end: '2024-01-22 13:06:37.246449'
failed_when_result: true
item:
- container_name: inspection_store
enabled: false
image: nginx:stable
network_mode: host
- inspection_store:/data
msg: non-zero return code
rc: 1
start: '2024-01-22 13:06:37.228639'
stderr: 'Error response from daemon: get inspection_store: no such volume'
stderr_lines: <omitted>
stdout: '[]'
stdout_lines: <omitted>
When I ssh to control0 and run `docker ps` or `docker volume ls` I see that no containers are running and no volumes are shown. Since this task is supposed to see what volumes are present, it really should not be failing if the answer is 'none'.
Yikes, good catch again. I guess we've only tested this when overcloud service deploy has run to completion. That said, this is one of the final playbooks and I imagine everything will have been destroyed by the point of this failure (we should definitely fix the bug though)