move restore to playbooks dir and attempt bugfix on setup-lxc
This commit is contained in:
139
playbooks/restore-backups.yml
Normal file
139
playbooks/restore-backups.yml
Normal file
@@ -0,0 +1,139 @@
|
||||
---
|
||||
- name: Centralized Restic Fleet Restore (Latest Snapshot)
|
||||
hosts: all
|
||||
become: true
|
||||
serial: 1 # Restores are riskier than backups; do one host at a time by default
|
||||
|
||||
# Same source/location mapping as run-backups.yaml, kept in sync intentionally.
|
||||
vars:
|
||||
confirm_restore: false
|
||||
|
||||
# Defaults to "latest" per host+tag. Override with -e "restore_snapshot_id=abcd1234"
|
||||
# to restore a specific snapshot instead. Applied to every target in the loop.
|
||||
restore_snapshot_id: "latest"
|
||||
|
||||
group_paths:
|
||||
app_data:
|
||||
label: apps_data
|
||||
path: /apps/data/
|
||||
app_docker:
|
||||
label: apps_docker
|
||||
path: /apps/docker/
|
||||
home_grail:
|
||||
label: home_grail
|
||||
path: /home/grail/
|
||||
|
||||
restore_targets: "{{ group_paths | dict2items | selectattr('key', 'in', group_names) | map(attribute='value') | list }}"
|
||||
|
||||
restic_env: &restic_env
|
||||
RESTIC_REPOSITORY: "s3:http://misamisa.duckdns.org:9000/resticbackup"
|
||||
RESTIC_PASSWORD: "{{ vault_restic_password }}"
|
||||
AWS_ACCESS_KEY_ID: "{{ vault_aws_access_key_id }}"
|
||||
AWS_SECRET_ACCESS_KEY: "{{ vault_aws_secret_access_key }}"
|
||||
|
||||
tasks:
|
||||
- name: Refuse to proceed without explicit confirmation
|
||||
ansible.builtin.assert:
|
||||
that:
|
||||
- confirm_restore | bool
|
||||
fail_msg: >-
|
||||
This playbook will overwrite backup_targets paths on {{ inventory_hostname }}
|
||||
with data from restic snapshots. Re-run with -e "confirm_restore=true" to proceed,
|
||||
optionally with -e "restore_snapshot_id=<id>" to pick a snapshot other than latest.
|
||||
success_msg: "Confirmed. Proceeding with restore on {{ inventory_hostname }}."
|
||||
|
||||
- name: Fail fast if this host matches no known group
|
||||
ansible.builtin.assert:
|
||||
that:
|
||||
- restore_targets | length > 0
|
||||
fail_msg: >-
|
||||
Host {{ inventory_hostname }} is not in any group listed in group_paths
|
||||
(app_data, app_docker, home_grail). Check group membership.
|
||||
|
||||
- name: List available snapshots (per target, for the log)
|
||||
community.docker.docker_container:
|
||||
name: "restic_snapshots_list_{{ item.label }}"
|
||||
image: restic/restic:latest
|
||||
state: started
|
||||
detach: false
|
||||
cleanup: true
|
||||
output_logs: true
|
||||
command: >-
|
||||
snapshots
|
||||
--host "{{ inventory_hostname }}"
|
||||
--tag "{{ item.label }}"
|
||||
env: *restic_env
|
||||
register: restic_snapshots_result
|
||||
loop: "{{ restore_targets }}"
|
||||
loop_control:
|
||||
label: "{{ item.label }}"
|
||||
|
||||
- name: Show available snapshots
|
||||
ansible.builtin.debug:
|
||||
var: restic_snapshots_result.results
|
||||
|
||||
- name: Restore Sequence with Safety Catch
|
||||
block:
|
||||
- name: Gracefully stop the application via Systemd
|
||||
ansible.builtin.systemd:
|
||||
name: app
|
||||
state: stopped
|
||||
|
||||
- name: Ensure restore target exists with correct ownership before mounting
|
||||
ansible.builtin.file:
|
||||
path: "{{ item.path }}"
|
||||
state: directory
|
||||
owner: docker
|
||||
group: docker
|
||||
mode: "0755"
|
||||
loop: "{{ restore_targets }}"
|
||||
loop_control:
|
||||
label: "{{ item.label }}"
|
||||
|
||||
- name: Check for and remove stale locks (per target)
|
||||
community.docker.docker_container:
|
||||
name: "restic_unlock_{{ item.label }}"
|
||||
image: restic/restic:latest
|
||||
state: started
|
||||
detach: false
|
||||
cleanup: true
|
||||
command: "unlock"
|
||||
env: *restic_env
|
||||
register: restic_unlock_result
|
||||
failed_when: restic_unlock_result.status != 0
|
||||
loop: "{{ restore_targets }}"
|
||||
loop_control:
|
||||
label: "{{ item.label }}"
|
||||
|
||||
- name: Restore latest snapshot into place (per target)
|
||||
community.docker.docker_container:
|
||||
name: "restic_restore_{{ item.label }}"
|
||||
image: restic/restic:latest
|
||||
state: started
|
||||
detach: false
|
||||
cleanup: true
|
||||
timeout: 600
|
||||
output_logs: true
|
||||
command: >-
|
||||
restore {{ restore_snapshot_id }}
|
||||
--host "{{ inventory_hostname }}"
|
||||
--tag "{{ item.label }}"
|
||||
--target /
|
||||
volumes:
|
||||
# NOTE: rw (no :ro) here — this is the whole point of a restore.
|
||||
- "{{ item.path }}:/backup_target/{{ item.label }}"
|
||||
env: *restic_env
|
||||
register: restic_restore_result
|
||||
failed_when: restic_restore_result.status != 0
|
||||
loop: "{{ restore_targets }}"
|
||||
loop_control:
|
||||
label: "{{ item.label }}"
|
||||
|
||||
- name: Show restore output
|
||||
ansible.builtin.debug:
|
||||
var: restic_restore_result.results
|
||||
always:
|
||||
- name: Ensure the application is started back up
|
||||
ansible.builtin.systemd:
|
||||
name: app
|
||||
state: started
|
||||
@@ -15,7 +15,21 @@
|
||||
|
||||
#- name: Setup-netbird.yml - setting up netbird mesh vpn
|
||||
# ansible.builtin.import_tasks: ./setup-netbird.yml
|
||||
- name: Capture inventory context before delegating
|
||||
ansible.builtin.set_fact:
|
||||
_inventory_file: "{{ inventory_file }}"
|
||||
_inventory_hostname: "{{ inventory_hostname }}"
|
||||
|
||||
- name: Run restore-backups.yaml runbook if either dir is missing
|
||||
ansible.builtin.command:
|
||||
cmd: >-
|
||||
ansible-playbook ./restore-backups.yml
|
||||
-i {{ playbook_dir }}/../inventory.yml
|
||||
--limit {{ inventory_hostname }}
|
||||
-e confirm_restore=true
|
||||
delegate_to: localhost
|
||||
when: not apps_docker_dir.stat.exists or not apps_data_dir.stat.exists
|
||||
|
||||
- name: Setup-restic.yml - setup restic
|
||||
ansible.builtin.import_tasks: ./setup-restic.yml
|
||||
|
||||
|
||||
Reference in New Issue
Block a user