This leads to much easier management and potential safety
features (validation). In the future, we could try to avoid port
conflicts as well, but it didn't seem worth the effort to do it now.
Our port ranges seem large enough.
This can also pave the way for a "presets" feature
(similar to `matrix_nginx_proxy_ssl_presets`) which makes it even easier
for people to configure worker counts.
These are just defensive cleanup tasks that we run.
In the good case, there's nothing to kill or remove, so they trigger an
error like this:
> Error response from daemon: Cannot kill container: something: No such container: something
and:
> Error: No such container: something
People often ask us if this is a problem, so instead of always having to
answer with "no, this is to be expected", we'd rather eliminate it now
and make logs cleaner.
In the event that:
- a container is really stuck and needs cleanup using kill/rm
- and cleanup fails, and we fail to report it because of error
suppression (`2>/dev/null`)
.. we'd still get an error when launching ("container name already in use .."),
so it shouldn't be too hard to investigate.
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`.
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
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)
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).
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.
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.
Since we'll likely have generic SQLite database importing
via [pgloader](https://pgloader.io/) for migrating bridge
databases from SQLite to Postgres, we'd rather avoid
calling the "import Synapse SQLite database" command
as just `--tags=import-sqlite-db`.
Similarly, for the media store, we'd like to mention that it's
related to Synapse as well.
We'd like to be more explicit, so as to be less confusing,
especially in light of other homeserver implementations
coming in the future.
The Docker 19.04 -> 20.10 upgrade contains the following change
in `/usr/lib/systemd/system/docker.service`:
```
-BindsTo=containerd.service
-After=network-online.target firewalld.service containerd.service
+After=network-online.target firewalld.service containerd.service multi-user.target
-Requires=docker.socket
+Requires=docker.socket containerd.service
Wants=network-online.target
```
The `multi-user.target` requirement in `After` seems to be in conflict
with our `WantedBy=multi-user.target` and `After=docker.service` /
`Requires=docker.service` definitions, causing the following error on
startup for all of our systemd services:
> Job matrix-synapse.service/start deleted to break ordering cycle starting with multi-user.target/start
A workaround which appears to work is to add `DefaultDependencies=no`
to all of our services.
ma1sd requires the openid endpoints for certain functionality.
Example: 90b2b5301c/src/main/java/io/kamax/mxisd/auth/AccountManager.java (L67-L99)
If federation is disabled, we still need to expose these openid APIs on the
federation port.
Previously, we were doing similar magic for Dimension.
As per its documentation, when running unfederated, one is to enable
the openid listener as well. As per their recommendation, people
are advised to do enable it on the Client-Server API port
and use the `federationUrl` variable to override where the federation
port is (making federation requests go to the Client-Server API).
Because ma1sd always uses the federation port (unless you do some
DNS overwriting magic using its configuration -- which we'd rather not
do), it's better if we just default to putting the `openid` listener
where it belongs - on the federation port.
With this commit, we retain the "automatically enable openid APIs" thing
we've been doing for Dimension, but move it to the federation port instead.
We also now do the same thing when ma1sd is enabled.
Many people probably didn't even know this - that ansible can be
quite a bit picky about what it will be willing to work with remotely.
Thanks @maxklenk !
`-v` magically creates the source destination as a directory,
if it doesn't exist already. We'd like to avoid this magic
and the potential breakage that it might cause.
We'd rather fail while Docker tries to find things to `--mount`
than have it automatically create directories and fail anyway,
while having contaminated the filesystem.
There's a lot more `-v` instances remaining to be fixed later on.
This is just some start.
Things like `matrix_synapse_container_additional_volumes` and
`matrix_nginx_proxy_container_additional_volumes` were not changed to
use `--mount`, as options for each one are passed differently
(`ro` is `ro`, but `rw` doesn't exist and `slave` is `bind-propagation=slave`).
To avoid breaking people's custom volume mounts, we keep it as it is for now.
A deficiency with `--mount` is that it lacks the `z` option (SELinux
ownership changes), and some of our `-v` instances use that. I'm not
sure how supported SELinux is for us right now, but it might be,
and breaking that would not be a good idea.
Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/716
This patch makes us use more fully-qualified container image names
(either prefixed with docker.io/ or with localhost/).
The latter happens when self-building is enabled.
We've recently had issues where if an image was removed manually
and the service was restarted (making `docker run` fetch it from Docker Hub, etc.),
we'd end up with a pulled image, even though we're aiming for a self-built one.
Re-running the playbook would then not do a rebuild, because:
- the image with that name already exists (even though it's something
else)
- we sometimes had conditional logic where we'd build only if the git
repo changed
By explicitly changing the name of the images (prefixing with localhost/),
we avoid such confusion and the possibility that we'd automatically pul something
which is not what we expect.
Also, I've removed that condition where building would happen on git
changes only. We now always build (unless an image with that name
already exists). We just force-build when the git repo changes.
also, worker.yaml.j2:
- hone worker_name
- remove worker_pid_file entry (would only be used if worker_daemonize
set to true; also, synapse only knows about the container namespace
and thus can not provide the required host-view PID)
- cherry-pick "Ensure worker config exists in systemd service (#7528)"
from synapse d74cdc1a42e8b487d74c214b1d0ca575429d546a:
"check that the worker config file exists instead of silently failing."
There's no change in the source code. Just a release bump for packaing
reasons. It doesn't matter much for us here, but let's be on the latest
tag anyway.
Depending on the distro, common commands like sleep and chown may either
be located in /bin or /usr/bin.
Systemd added path lookup to ExecStart in v239, allowing only the
command name to be put in unit files and not the full path as
historically required. At least Ubuntu 18.04 LTS is however still on
v237 so we should maintain portability for a while longer.
Well, actually 8cd9cde won't work, unless we put the
`|to_nice_yaml` thing on a new line.
We can, but that takes more lines and makes things look uglier.
Using `|to_json` seems good enough.
The whole file is parsed as YAML later on and merged with the
`_extension` variable before being dumped as YAML again in the end.
This keeps the roles cleaner and more independent of matrix-base,
which may be important for people building their own playbook
out of the individual roles and not using the matrix-base role.
Synapse v1.9.0 changed some things which made the REST Auth Password
Provider break.
The ma1uta/matrix-synapse-rest-password-provider implements some
workarounds for now and will likely deliver a proper fix in the future.
Not much has changed between the 2 projects, so this should be a
painless transition.
Can you double check that the way I have this set only exposes it locally? It is important that the manhole is not available to the outside world since it is quite powerful and the password is hard coded.
Prompted by: https://matrix.org/blog/2019/11/09/avoiding-unwelcome-visitors-on-private-matrix-servers
This is a bit controversial, because.. the Synapse default remains open,
while the general advice (as per the blog post) is to make it more private.
I'm not sure exactly what kind of server people set up and whether they
want to make the room directory public. Our general goal is to favor
privacy and security when running personal (family & friends) and corporate
homeservers, both of which likely benefit from having a more secure default.
Don't mention systemd-journald adjustment anymore, because
we've changed log levels to WARNING and Synapse is not chatty by default
anymore.
The "excessive log messages may get dropped on CentOS" issue no longer
applies to most users and we shouldn't bother them with it.
matrix_synapse_storage_path is already defined in matrix-synapse/defaults/main.yml (with a default of "{{ matrix_synapse_base_path }}/storage"), but was not being used for its presumed purpose in matrix-synapse.service.j2. As a result, if matrix_synapse_storage_path was overridden (in a vars.yml), the synapse service failed to start.
This only gets triggered if:
- the Synapse role is used standalone and the default values are used
- the whole playbook is used, with `matrix_mxisd_enabled: false`
Also discussed previously in #213 (Github Pull Request).
shared-secret-auth and rest-auth logging is still at `INFO`
intentionally, as user login events seem more important to keep.
Those modules typically don't spam as much.
ef5e4ad061 intentionally makes us conform to
the logging format suggested by the official Docker image.
Reverting this part, because it's uglier.
This likely should be fixed upstream as well though.
Somewhat related to #213 (Github Pull Request).
We've been moving in the opposite direction for quite a long time.
All services should just leave logging to systemd's journald.
Fixes a regression introduced during the upgrade to
Synapse v1.1.0 (in 2b3865ceea).
Since Synapse v1.1.0 upgraded to Python 3.7
(https://github.com/matrix-org/synapse/pull/5546),
we need to use a different modules directory when mounting
password provider modules.
Using `|to_json` on a string is expected to correctly wrap it in quotes (e.g. `"4"`).
Wrapping it explicitly in double-quotes results in undesirable double-quoting (`""4""`).
We do use some `:latest` images by default for the following services:
- matrix-dimension
- Goofys (in the matrix-synapse role)
- matrix-bridge-appservice-irc
- matrix-bridge-appservice-discord
- matrix-bridge-mautrix-facebook
- matrix-bridge-mautrix-whatsapp
It's terribly unfortunate that those software projects don't release
anything other than `:latest`, but that's how it is for now.
Updating that software requires that users manually do `docker pull`
on the server. The playbook didn't force-repull images that it already
had.
With this patch, it starts doing so. Any image tagged `:latest` will be
force re-pulled by the playbook every time it's executed.
It should be noted that even though we ask the `docker_image` module to
force-pull, it only reports "changed" when it actually pulls something
new. This is nice, because it lets people know exactly when something
gets updated, as opposed to giving the indication that it's always
updating the images (even though it isn't).
