Tasks:
- Enable docker_compose_flush_handlers=true so services come up immediately.
- Add DB readiness guard via maintenance/sql.php (SELECT 1).
- Run maintenance/install.php on empty schema with robust changed_when/failed_when (merge stdout+stderr); keep secrets hidden.
- Run maintenance/update.php for migrations with neutral changed_when unless work is done.
- Make admin creation idempotent: tolerate 'already exists' and 'Account exists', keep async+no_log.
Config changes:
- Remove LocalSettings.php template and its host bind-mount from compose.
- Drop MediaWiki settings path variables and META namespace variable (unused after switch).
Result: First boot is fully automated (schema + admin), subsequent runs are cleanly idempotent.
Ref: ChatGPT conversation (Aug 28, 2025, Europe/Berlin) — https://chatgpt.com/share/68b0d2e1-9bc0-800f-81a5-db03ce0b81e3.
- Removed outdated Installation.md (manual plugin instructions no longer needed)
- Added schema/main.yml with validation for secret_pin
- Added config.php.j2 template to manage DB + admin config
- Refactored tasks/main.yml to deploy config.php instead of legacy docker-compose
- Removed setup-domain.yml (TLS/domain handling moved to core roles)
- Updated docker-compose.yml.j2 to mount config.php and use new vars
- Cleaned up vars/main.yml: standardized MYBB_* variable names, added MYBB_SECRET_PIN, config paths, and container port
See ChatGPT conversation: https://chatgpt.com/share/68b0ae26-93ec-800f-8785-0da7c9303090
- Converted group_vars/all/13_ldap.yml from lower-case to ALL-CAPS nested keys.
- Updated all roles, tasks, templates, and filter_plugins to reference LDAP.* instead of ldap.*.
- Fixed Keycloak JSON templates to properly quote Jinja variables.
- Adjusted svc-db-openldap filter plugins and unit tests to handle new LDAP structure.
- Updated integration test to only check uniqueness of TOP-LEVEL ALL-CAPS constants, ignoring nested keys.
See: https://chatgpt.com/share/68b01017-efe0-800f-a508-7d7e2f1c8c8d
- Allow kc_object_kind='realm'
- Map endpoint to 'realms' and default lookup_field to 'id'
- Use realm-specific kcadm GET/UPDATE (no -r flag)
- Preserve immutables: id, realm
- Guard query-based ID resolution to non-realm objects
Context: fixing failure in 'Update REALM mail settings' task.
See: https://chatgpt.com/share/68affdb8-3d28-800f-8480-aa6a74000bf8
Add roles/web-app-taiga/tasks/01_administrator.yml to handle admin creation via 'createsuperuser' and, on failure, an upsert fallback using 'manage.py shell'. Ensures email, is_staff, is_superuser, is_active are set and password is updated when needed; emits CHANGED marker for idempotence.
Update roles/web-app-taiga/tasks/main.yml to include the new 01_administrator.yml task file, removing the inline admin logic for better separation of concerns.
Uses taiga-manage helper service and composes docker-compose.yml with docker-compose-inits.yml to inherit env/networks/volumes consistently.
Chat reference: https://chatgpt.com/share/68af7637-225c-800f-b670-2b948f5dea54
- Introduced dedicated variables in vars/main.yml:
* TAIGA_FLAVOR_TAIGAIO
* TAIGA_TAIGAIO_ENABLED
- Replaced inline Jinja2 get_app_conf checks with TAIGA_TAIGAIO_ENABLED for
consistency in tasks, docker-compose template and env file.
- Adjusted env.j2 to use TAIGA_TAIGAIO_ENABLED instead of direct flavor checks.
- Enabled css by default (true instead of false).
- Cleaned up spacing/indentation in config and env.
This improves readability, reduces duplicated logic, and makes it easier to
maintain both OIDC flavors (robrotheram, taigaio).
Conversation: https://chatgpt.com/share/68af65b3-27c0-800f-964f-ff4f2d96ff5d
- Introduce WEBSOCKET_PROTOCOL derived from WEB_PROTOCOL (wss if https, else ws).
- Replace hardcoded websocket URLs in EspoCRM, Nextcloud and Taiga with {{ WEBSOCKET_PROTOCOL }}.
- Fix mautrix-imessage to use ws:// for internal synapse:8008.
- Standardize Pixelfed OIDC env spacing.
- Refactor Taiga variables to TAIGA_* naming convention and clean up EMAIL_BACKEND definition.
See: https://chatgpt.com/share/68af62fa-4dcc-800f-9aaf-cff746daab1e
- Move image definition from images.joomla to docker.services.joomla
- Add container name, container_port variable, and healthcheck
- Introduce JOOMLA_IMAGE, JOOMLA_VERSION, JOOMLA_CONTAINER, JOOMLA_VOLUME in vars
- Use volume mapping via docker.volumes.data
See: https://chatgpt.com/share/68af55a9-6514-800f-b6f7-1dc86356936e
- Extend certbundle task to ignore 'The service is down for maintenance or had an internal error'
as a fatal failure.
- Add debug/warning output when this error occurs, so playbook does not stop but logs the issue.
- Ensure changed_when does not mark run as changed if only maintenance error was hit.
Ref: https://chatgpt.com/share/68af4e15-24cc-800f-b1dd-6a5f2380e35a
- Removed COPY of msmtp configuration from Dockerfile to avoid baking secrets/config into the image
- Added volume mount for host-side msmtp config ({{ WORDPRESS_HOST_MSMTP_CONF }}) in docker-compose.yml
- Keeps PHP upload.ini handling inside the image, but externalizes sensitive mail configuration
- Increases flexibility and avoids rebuilds when msmtp config changes
Ref: https://chatgpt.com/share/68af3c51-0544-800f-b76f-b2660c43addb
- Replaced all lowercase wordpress_* variables with uppercase WORDPRESS_* equivalents
- Ensured consistency across tasks, templates, and vars
- Improves readability and aligns with naming conventions
Conversation: https://chatgpt.com/share/68af29b5-8e7c-800f-bd12-48cc5956311c
- Fixed incorrect filter usage in docker-compose handler (proper use of | path_join).
- Improved LetsEncrypt template by joining paths with filenames instead of appending manually.
- Enhanced sys-svc-msmtp task with an additional condition to only run if no-reply mailu_token exists.
- Updated Keycloak meta to depend on Mailu (ensuring token generation before setup).
- Refactored Keycloak import path variables to use path_join consistently.
- Adjusted Mailu meta dependency to run after Matomo instead of Keycloak.
See: https://chatgpt.com/share/68af13e6-edc0-800f-b76a-a5f427837173
- Improved CSP filter to properly include web-svc-cdn and use protocol-aware domains
- Added Todo.md with redis and OIDC notes
- Enhanced Akaunting role config with CSP flags and redis option
- Updated schema to include app_key validation
- Reworked tasks to handle first-run marker logic cleanly
- Fixed docker-compose template (marker, healthcheck, setup flag)
- Expanded env.j2 with cache, email, proxy, and redis options
- Added javascript.js.j2 template for SSO warning
- Introduced structured vars for Akaunting role
- Removed deprecated update-repository-with-files.yml task
See conversation: https://chatgpt.com/share/68af00df-2c74-800f-90b6-6ac5b29acdcb
- update cli/create/credentials.py to handle vault literals correctly:
* strip 'vault |' headers and keep only ANSIBLE_VAULT body
* skip reprocessing keys added in same run (no duplicate confirmation prompts)
* detect both 'vault' and 'ANSIBLE_VAULT' as already encrypted
Refs: https://chatgpt.com/share/68aed780-ad4c-800f-877d-aa4c40a47755
- Introduced a generic sso_warning.js.j2 template under
templates/roles/web-app/templates/javascripts/
- Included this template in web-app-baserow/templates/javascript.js.j2
- Added new variable js_application_name in
roles/web-app-baserow/vars/main.yml to make the warning
application-specific
- Implemented cookie-based logic so the warning is only shown once
per user (default: 365 days)
Reference: https://chatgpt.com/share/68aecdae-82d0-800f-b05e-f2cb680664f1
- Extended kc_force_attrs in tasks/main.yml to source 'publicClient',
'serviceAccountsEnabled' and 'frontchannelLogout' directly from
KEYCLOAK_DICTIONARY_REALM for consistency with import definitions.
- Updated default.json.j2 import template to set 'publicClient' to true.
- Public client mode is required so the frontend API of role 'web-app-desktop'
can handle login/logout flows without client secret.
Ref: https://chatgpt.com/share/68ae0060-4fac-800f-9f02-22592a4087d3
Replaced over-escaped '{{{{.Names}}}}' with proper '{{.Names}}'
in docker ps commands. This resolves 'failed to parse template:
unexpected "{" in command' errors during unhealthy/exited
container detection.
Reference: https://chatgpt.com/share/68addfd9-fa78-800f-abda-49161699e673
- Added detect_env_file() to both sys-ctl-rpr-docker-soft and sys-ctl-rpr-docker-hard
* prefer .env, fallback to .env/env
* append --env-file parameter automatically
- Refactored soft script to use compose_cmd() for consistent command building
- Adjusted error recovery path in soft script to also respect env-file
- Extended unit tests for soft script to cover env-file priority and restart commands
- Added new unit tests for hard script verifying env-file priority, cwd handling,
and --only filter logic
Ref: https://chatgpt.com/share/68ad7b30-7510-800f-8172-56f03a2f40f5
- extended generic systemctl template to support ExecStartPost
- health-docker-volumes: run main script with whitelist, trigger both compose alarm and cleanup on failure
- repair-docker-hard: added ExecStartPre lock, ExecStart, and ExecStartPost to trigger compose alarm always, plus cleanup on failure
- removed obsolete role-specific systemctl.service.j2 templates
- improved consistency across vars and defaults
See: https://chatgpt.com/share/68ad6cb8-c164-800f-96b6-a45c6c7779b3
