Commit Graph
12 Commits
Author SHA1 Message Date
Slavi PantaleevandClaude Opus 5.5 bff3f54d53 Harden pre-start container cleanup against name-release races
Containers are created with `--rm`, so Docker removes them asynchronously after they stop.
The pre-start `docker rm` returns immediately while that removal is still in progress
(its error is suppressed), and the following `docker create --name` can fail with
"The container name ... is already in use".

This is easy to hit since systemd_service_manager switched its default restart mode to `all-at-once`:
many containers stop and start at the same time and Docker falls behind on removals.
The playbook then fails at "Restart systemd services needing restart at once",
even though the services recover on their own (`Restart=always`) a little later.

The external roles (MASH, devture) were hardened against this in 2026-03
(see mother-of-all-self-hosting/ansible-role-valkey@0ba79c0), but the roles in `roles/custom` never were.
This ports the same bounded cleanup loop to them: stop, force-remove,
and wait until Docker releases the name, failing explicitly after a timeout.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-10-16 19:04:04 +03:00
bd6e08a93e Allow host network mode for MDAD-managed containers (#5310)
* feat: support container_network=host across all roles + systemd templates

Mirror the pattern Slavi introduced for matrix-coturn (aafa8f0) across the
fork: every 'Ensure X container network is created' task gets a
'when: <var> not in ["", "host"]' guard so MDAD does not try to
docker_network create a network literally named 'host' (returns 403,
since host is a pre-defined Docker network).

Mirror the same guard in every systemd unit template that does
'ExecStartPre=docker network connect <addnet> <container>' loops over
matrix_<role>_container_additional_networks: skip the connects when the
container is on host networking (where additional --network attaches
are invalid).

Unblocks DiD setups where MDAD-managed containers share their host's
network namespace (matrix-mdad outer compose service joined to central
postgres/openldap networks) to reach external services on the outer
Docker daemon.

* Simplify container network guards (!= 'host') and fix duplicate when

Guarding on the empty string ('') as well was misleading: systemd unit
templates still render an unconditional --network= flag, so an empty
network value produces a broken docker create command. Only 'host' is
actually supported, so only guard on that. This also matches the
existing convention in the Traefik role
(when: traefik_container_network != 'host').

Also fix a duplicate when key in the meshtastic-relay role, where the
network-creation task already had a when condition - the two are now
combined into a list.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Slavi Pantaleev <slavi@devture.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-06-10 23:48:03 +03:00
Slavi Pantaleev 51ab9ed9ff Adjust lstrip_blocks to use a bool value for compatibility with ansible-core>=2.19.0
Related to https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/4445
2025-07-26 17:58:12 +03:00
Slavi Pantaleev e8c9a0e410 Randomize matrix-synapse-auto-compressor.timer start with a RandomizedDelaySec (defaults to 6h) 2025-03-29 10:34:37 +02:00
Suguru Hirahara a039174f96 Add license information to files for matrix-synapse-auto-compressor
Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>
2025-03-03 00:17:47 +09:00
Slavi Pantaleev d12643eb91 Do not use docker stop --time (deprecated in Docker v28 in favor of --timeout or just -t)
`-t` works on both old and new Docker versions, so it's best to use that.
2025-02-22 21:42:18 +02:00
Suguru Hirahara 08e34a4a6f Rename "matrix_synapse_auto_compressor_calendar" to "matrix_synapse_auto_compressor_schedule" for consistency
Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>
2025-02-05 00:25:11 +09:00
70cbf3d5ae add synapse-auto-compressor workaround, fixes #3397 (#3473)
* add synapse-auto-compressor workaround, fixes #3397

* Clarify what the PG-prefixed variables are for

---------

Co-authored-by: Slavi Pantaleev <slavi@devture.com>
2024-08-15 11:36:48 +03:00
Slavi Pantaleev 2511b34a7c Stop containers gracefully, instead of outright killing them 2023-12-06 11:52:23 +02:00
Slavi Pantaleev 328d0d8a5f Move synapse-auto-compressor Postgres argument to an environment variable
This provides an additional security benefit. The password won't leak in
the process list anymore.
2023-03-12 10:18:33 +02:00
Slavi Pantaleev 26d5719df4 Make matrix-synapse-auto-compressor live in its own container network
It will, additionally, be connected to the devture-postgres network, if
devture-postgres is enabled.
2023-03-12 10:18:33 +02:00
Aine bf3677a9b5 add synapse_auto_compressor 2023-03-10 23:48:43 +02:00