Commit graph

1474 commits

Author SHA1 Message Date
Slavi Pantaleev 66cdc7bf5a Clean up worker.yaml generation a bit and make it more flexible 2021-01-25 13:02:01 +02:00
Slavi Pantaleev 1462409b34 Fix worker listening addresses
Not specifying bind addresses for the worker resulted in this warning:

> synapse.app - 47 - WARNING - None - Failed to listen on 0.0.0.0, continuing because listening on [::]

Additionally, metrics listening only on 127.0.0.1 seems like a no-op.
Only having it accessible from within the container is likely not what
we intend. Changed that to all interfaces as well.

Whether it actually gets exposed or not depends on the systemd service
and `matrix_synapse_workers_container_host_bind_address`.
2021-01-25 12:29:47 +02:00
Slavi Pantaleev 01747c8cc4 Prevent Synapse warning about enabling metric listeners with enable_metrics: false
> synapse.app.generic_worker - 606 - WARNING - None - Metrics listener configured, but enable_metrics is not True!
2021-01-25 12:24:12 +02:00
Slavi Pantaleev 70796703d3 Run Synapse workers in their own containers
This switches the `docker exec` method of spawning
Synapse workers inside the `matrix-synapse` container with
dedicated containers for each worker.

We also have dedicated systemd services for each worker,
so this are now:
- more consistent with everything else (we don't use systemd
instantiated services anywhere)
- we don't need the "parse systemd instance name into worker name +
port" part
- we don't need to keep track of PIDs manually
- we don't need jq (less depenendencies)
- workers dying would be restarted by systemd correctly, like any other
service
- `docker ps` shows each worker separately and we can observe resource
usage
2021-01-25 12:14:46 +02:00
Slavi Pantaleev da50fb27a0 Whitelist /_matrix/key requests for going to generic workers on the federation port 2021-01-25 09:46:50 +02:00
Slavi Pantaleev 4d62a75f6f Get matrix-corporal to play nicely with a Synapse worker setup
We do this by creating one more layer of indirection.

First we reach some generic vhost handling matrix.DOMAIN.
A bunch of override rules are added there (capturing traffic to send to
ma1sd, etc). nginx-status and similar generic things also live there.

We then proxy to the homeserver on some other vhost (only Synapse being
available right now, but repointing this to Dendrite or other will be
possible in the future).
Then that homeserver-specific vhost does its thing to proxy to the
homeserver. It may or may not use workers, etc.

Without matrix-corporal, the flow is now:
1. matrix.DOMAIN (matrix-nginx-proxy/matrix-domain.conf)
2. matrix-nginx-proxy/matrix-synapse.conf
3. matrix-synapse

With matrix-corporal enabled, it becomes:
1. matrix.DOMAIN (matrix-nginx-proxy/matrix-domain.conf)
2. matrix-corporal
3. matrix-nginx-proxy/matrix-synapse.conf
4. matrix-synapse

(matrix-corporal gets injected at step 2).
2021-01-25 09:46:41 +02:00
Slavi Pantaleev c05d3d09bd Disable systemd services while stopping them
This removes some `multi-target.wants` symlinks as well, etc.

But despite systemd saying:

> Removed symlink /etc/systemd/system/matrix-synapse.service.wants/matrix-synapse-worker@appservice:0.service

.. I still see such symlinks tehre for me for some reason, so keeping the
code (below) to find & delete them still seems like a good idea.
2021-01-25 08:58:23 +02:00
Slavi Pantaleev 63301b0ef1 Improvements around Synapse worker/metrics ports exposure
There was a `matrix_nginx_proxy_enabled|default(False)` check, but:
- it didn't seem to work reliably for some reason (hmm)
- referring to a `matrix_nginx_proxy_*` variable from within the
  `matrix-synapse` role is not ideal
- exposing always happened on `127.0.0.1`, which may not be good enough
  for some rarer setups (where the own webserver is external to the host)
2021-01-25 08:25:43 +02:00
Slavi Pantaleev f66a6b066b Be more specific with the Redis version being used 2021-01-25 01:34:58 +02:00
Slavi Pantaleev 5ca68210cd Do not handle /_matrix/federation on client-server port, nor /_matrix/client stuff on federation port
I guess it didn't hurt to do it until now, but it's not great serving
federation APIs on the client-server API port, etc.

matrix-corporal doesn't work yet (still something to be solved in the
future), but its firewalling operations will also be sabotaged
by Client-Server APIs being served on the federation port (it's a way to get around its firewalling).
2021-01-24 22:22:57 +02:00
Slavi Pantaleev cc5cf0d725 Load roles/matrix-synapse/vars/workers.yml earlier to not break --tags=setup-nginx-proxy
If we load it at runtime, during matrix-synapse role execution,
it's good enough for matrix-synapse and all roles after that,
but.. it breaks when someone uses `--tags=setup-nginx-proxy` alone.

The downside of including this vars file like this in `setup.yml`
is that the variables contained in it cannot be overriden by the user
(in their inventory's `vars.yml`).
... but it's not like overriding these variables was possible anyway
when including them at runtime.
2021-01-24 20:19:55 +02:00
Slavi Pantaleev 92ee3d78a0 Fix matrix-remove-all for when Synapse workers are enabled 2021-01-24 19:42:32 +02:00
Marcel Partap edc21f15e5 Restrict publishing worker (metrics) ports to localhost 2021-01-24 08:53:09 +01:00
Marcel Partap 183adec3d8 Merge remote-tracking branch 'origin/master' into synapse-workers 2021-01-23 15:04:11 +01:00
Marcel Partap c8f051a42d Track workers endpoint list in repo instead of regenerating on user side 2021-01-23 14:44:36 +01:00
Marcel Partap f2c7d79238 Drop probably incorrect comment from synapse homeserver.yaml.j2 2021-01-23 14:44:36 +01:00
Slavi Pantaleev a56cb34850 Notify people if /matrix/postgres/data-auto-upgrade-backup exists 2021-01-23 14:14:45 +02:00
Slavi Pantaleev a2422c458a Notify of remaining matrix-postgres local data in a better way 2021-01-23 14:04:51 +02:00
Slavi Pantaleev 1cd251ed78 Don't delete Docker images which may have been pulled by another
Some people run Coturn or Jitsi, etc., by themselves and disable it
in the playbook.

Because the playbook is trying to be nice and clean up after itself,
it was deleting these Docker images.

However, people wish to pull and use them separately and would rather
they don't get deleted.

We could make this configurable for the sake of this special case, but
it's simpler to just avoid deleting these images.
It's not like this "cleaning things up" thing works anyway.
As time goes on, the playbook gets updated with newer image tags
and we leave so many images behind. If one doesn't run
`docker system prune -a` manually once in a while, they'd get swamped
with images anyway. Whether we leave a few images behind due to the lack
of this cleanup now is pretty much irrelevant.
2021-01-23 14:01:31 +02:00
Slavi Pantaleev f085362149 Fix some Postgres CLI scripts to target the correct database
Fixes a regression introduced in 95346f3117.

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/814

Using `matrix_synapse_` variables in the `matrix-postgres` role is not
ideal, but.. this script belongs neither here, nor there.
We'll have it be like that for now.
2021-01-23 11:38:34 +02:00
Slavi Pantaleev 3051655d21
Ensure matrix_appservice_irc_docker_src_files_path created when self-building
The git module will create it anyway, but that would likely use `root:root`.
2021-01-22 22:42:40 +02:00
Panagiotis Georgiadis f10e3fef0d
Merge branch 'master' into irc 2021-01-22 20:30:24 +00:00
Panagiotis Georgiadis e502ee33da
Selfbuild appservice-irc bridge 2021-01-22 21:28:53 +01:00
Slavi Pantaleev f9968b6981 Fix matrix_postgres_connection_password length check 2021-01-22 21:22:58 +02:00
Slavi Pantaleev 2997a7fc3e Make mx-puppet-* bridges not log to files
We log everything in systemd/journald for every service already,
so there's no need for double-logging, bridges rotating log files
manually and other such nonsense.
2021-01-22 19:22:26 +02:00
Slavi Pantaleev f3dd346724 Try to tighten Signal bridge security 2021-01-22 18:56:08 +02:00
Slavi Pantaleev 8ec975e3c8 Use matrix:matrix for Signal bridge (not root) 2021-01-22 18:52:20 +02:00
Slavi Pantaleev 37909aa7a9 Create signald/{avatars,attachments,data} and rename config dir 2021-01-22 18:40:51 +02:00
Slavi Pantaleev 88addd71fc Fix Postgres imports going to the matrix DB by default
Well, they still do go to that DB by default,
but our docs give a better command to users, which would do the right
thing.
2021-01-22 17:39:08 +02:00
Slavi Pantaleev bef0702fea Wait some more when starting Postgres during setup on ARM 2021-01-22 16:21:30 +02:00
Slavi Pantaleev f9c1d62435 Fix Postgres database (-alpine) failing to start on ARM32 2021-01-22 13:52:55 +02:00
Slavi Pantaleev 95346f3117 Reorganize Postgres access (breaking change)
In short, this makes Synapse a 2nd class citizen,
preparing for a future where it's just one-of-many homeserver software
options.

We also no longer have a default Postgres superuser password,
which improves security.

The changelog explains more as to why this was done
and how to proceed from here.
2021-01-22 13:26:12 +02:00
throwawayay a30ef0cc29
Update element-web (1.7.16 -> 1.7.17) 2021-01-20 08:35:07 -05:00
Slavi Pantaleev 024a23ed17 Upgrade mautrix-facebook to the new Postgres-only version
I had intentionally held it back in 39ea3496a4
until:
- it received more testing (there were a few bugs during the
migration, but now it seems OK)
- this migration guide was written
2021-01-20 10:12:51 +02:00
pushytoxin d51ea25219 When validating LE certs, do not wait for a random time
While administering we will occasionally invoke this script interactively with the "non-interactive" switch still there, yet still sit at the desk waiting for 300 seconds for this timer to run out.

The systemd-timer already uses a 3h randomized delay for automatic renewals, which serves this purpose well.
2021-01-19 18:41:45 +01:00
Slavi Pantaleev 39ea3496a4 Downgrade/lock mautrix-facebook to pre-mobile times
The `mobile` branch got merged to `master`, which ends up becoming
`:latest`. It's a "rewrite" of the bridge's backend and only
supports a Postgres database.

We'd like to go back (well, forward) to `:latest`, but that will take
a little longer, because:
- we need to handle and document things for people still on SQLite
(especially those with external Postgres, who are likely on SQLite for
bridges)
- I'd rather test the new builds (and migration) a bit before
releasing it to others and possibly breaking their bridge

Brave ones who are already using the bridge with Postgres
can jump on `:latest` and report their experience.
2021-01-19 18:44:15 +02:00
Slavi Pantaleev c9d96d8135 Fix mautrix-telegram paths creation bug 2021-01-19 09:15:34 +02:00
Slavi Pantaleev 56c54d5cc7 Upgrade matrix-corporal (2.0.1 -> 2.1.0) 2021-01-18 18:23:17 +02:00
Slavi Pantaleev c1008fde44 Upgrade matrix-coturn (4.5.1.3 -> 4.5.2) 2021-01-18 00:41:47 +02:00
Slavi Pantaleev cf06f84608 Upgrade matrix-corporal (2.0.0 -> 2.0.1) 2021-01-17 22:05:26 +02:00
Slavi Pantaleev d95cbe38d7 Rename configuration setting 2021-01-17 18:29:26 +02:00
Slavi Pantaleev 28d86e3aaa Initial work on support for matrix-corporal v2 2021-01-16 23:47:14 +02:00
Slavi Pantaleev 8549926395 Attempt to fix mautrix-whatsapp DB migration user table conflict
Discussed in https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/791
2021-01-15 17:13:47 +02:00
Slavi Pantaleev 1692a28fe4 Work around annoying Docker warning about undefined $HOME
> WARNING: Error loading config file: .dockercfg: $HOME is not defined

.. which appeared in Docker 20.10.
2021-01-15 00:23:01 +02:00
Slavi Pantaleev 26f0bbfdef Fix self-building for matrix-ma1sd on non-version tag/branch
Building `master` or something like this was failing.
2021-01-14 23:57:38 +02:00
Slavi Pantaleev 9e936e45ad Use BuildKit for ma1sd Docker building
Newer versions (`master`) use things like `--platform=...`,
which are not supported unless we enable the new BuildKit building
backend.
2021-01-14 23:48:30 +02:00
Slavi Pantaleev e1690722f7 Replace cronjobs with systemd timers
Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/756

Related to https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/737

I feel like timers are somewhat more complicated and dirty (compared to
cronjobs), but they come with these benefits:

- log output goes to journald
- on newer systemd distros, you can see when the timer fired, when it
will fire, etc.
- we don't need to rely on cron (reducing our dependencies to just
systemd + Docker)

Cronjobs work well, but it's one more dependency that needs to be
installed. We were even asking people to install it manually
(in `docs/prerequisites.md`), which could have gone unnoticed.

Once in a while someone says "my SSL certificates didn't renew"
and it's likely because they forgot to install a cron daemon.

Switching to systemd timers means that installation is simpler
and more unified.
2021-01-14 23:35:50 +02:00
Slavi Pantaleev 05ca9357a8 Add .service suffix to systemd units list
We'll be adding `.timer` units later on, so it's good to be
more explicit.
2021-01-14 23:02:10 +02:00
Slavi Pantaleev 653d1d7924 Revert "Don't self-build ma1sd every time unless git sources changed"
This reverts commit 2a25b63bb6.

Looking at other roles, we trigger building regardless of this.
It's better to always trigger it, because it's less fragile.
If the build fails and we only trigger it on "git changes"
then we won't trigger it for a while. That's not good.

Triggering it each and every time may seem like a waste,
but it supposedly runs quickly due to Docker caching.
2021-01-14 22:20:51 +02:00
Slavi Pantaleev 6f5aaad48d Split install/uninstall tasks in matrix-coturn 2021-01-14 22:11:38 +02:00