Mirrors
/
synapse-old
mirror of https://github.com/matrix-org/synapse.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
synapse-old/docker/configure_workers_and_start.py

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

1032 lines
41 KiB
Python
Raw Normal View History

Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
#!/usr/bin/env python
# Copyright 2021 The Matrix.org Foundation C.I.C.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# This script reads environment variables and generates a shared Synapse worker,
# nginx and supervisord configs depending on the workers requested.
#
# The environment variables it reads are:
# * SYNAPSE_SERVER_NAME: The desired server_name of the homeserver.
# * SYNAPSE_REPORT_STATS: Whether to report stats.
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# * SYNAPSE_WORKER_TYPES: A comma separated list of worker names as specified in WORKERS_CONFIG
# below. Leave empty for no workers. Add a ':' and a number at the end to
# multiply that worker. Append multiple worker types with '+' to merge the
# worker types into a single worker. Add a name and a '=' to the front of a
# worker type to give this instance a name in logs and nginx.
# Examples:
# SYNAPSE_WORKER_TYPES='event_persister, federation_sender, client_reader'
# SYNAPSE_WORKER_TYPES='event_persister:2, federation_sender:2, client_reader'
# SYNAPSE_WORKER_TYPES='stream_writers=account_data+presence+typing'
Support registering Application Services when running with workers under Complement. (#12826) Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
2022-05-23 07:11:06 -06:00
# * SYNAPSE_AS_REGISTRATION_DIR: If specified, a directory in which .yaml and .yml files
# will be treated as Application Service registration files.
Remove Caddy from the Synapse workers image used in Complement. (#12818)
2022-05-23 03:29:24 -06:00
# * SYNAPSE_TLS_CERT: Path to a TLS certificate in PEM format.
# * SYNAPSE_TLS_KEY: Path to a TLS key. If this and SYNAPSE_TLS_CERT are specified,
# Nginx will be configured to serve TLS on port 8448.
Improve startup times in Complement test runs against workers, particularly in CPU-constrained environments. (#13127) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-30 05:58:12 -06:00
# * SYNAPSE_USE_EXPERIMENTAL_FORKING_LAUNCHER: Whether to use the forking launcher,
# only intended for usage in Complement at the moment.
# No stability guarantees are provided.
Add the ability to set the log level using the `SYNAPSE_TEST_LOG_LEVEL` environment when using `complement.sh`. (#13152)
2022-07-05 03:46:20 -06:00
# * SYNAPSE_LOG_LEVEL: Set this to DEBUG, INFO, WARNING or ERROR to change the
# log level. INFO is the default.
# * SYNAPSE_LOG_SENSITIVE: If unset, SQL and SQL values won't be logged,
# regardless of the SYNAPSE_LOG_LEVEL setting.
Log when events are (unexpectedly) filtered out of responses in tests (#14213) See https://github.com/matrix-org/synapse/pull/14095#discussion_r990335492 This is useful because when see that a relevant event is an `outlier` or `soft-failed`, then that's a good unexpected indicator explaining why it's not showing up. `filter_events_for_client` is used in `/sync`, `/messages`, `/context` which are all common end-to-end assertion touch points (also notifications, relations).
2023-06-01 20:27:18 -06:00
# * SYNAPSE_LOG_TESTING: if set, Synapse will log additional information useful
# for testing.
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
#
# NOTE: According to Complement's ENTRYPOINT expectations for a homeserver image (as defined
# in the project's README), this script may be run multiple times, and functionality should
# continue to work if so.
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
import dataclasses
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
import os
Set LD_PRELOAD to load jemalloc in Dockerfile-workers. (#14182)
2022-10-14 06:29:49 -06:00
import platform
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
import re
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
import subprocess
import sys
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
from collections import defaultdict
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
from dataclasses import dataclass, field
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
from itertools import chain
Support registering Application Services when running with workers under Complement. (#12826) Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
2022-05-23 07:11:06 -06:00
from pathlib import Path
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
from typing import (
Any,
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
Callable,
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
Dict,
List,
Mapping,
MutableMapping,
NoReturn,
Optional,
Set,
SupportsIndex,
)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
import yaml
Merge the Complement testing Docker images into a single, multi-purpose image. (#12881) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-08 03:57:05 -06:00
from jinja2 import Environment, FileSystemLoader
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
MAIN_PROCESS_HTTP_LISTENER_PORT = 8080
Remove `worker_replication_*` settings (#15491) * Add master to the instance_map as part of Complement, have ReplicationEndpoint look at instance_map for master. * Fix typo in drive by. * Remove unnecessary worker_replication_* bits from unit tests and add master to instance_map(hopefully in the right place) * Several updates: 1. Switch from master to main for naming the main process in the instance_map. Add useful constants for easier adjustment of names in the future. 2. Add backwards compatibility for worker_replication_* to allow time to transition to new style. Make sure to prioritize declaring main directly on the instance_map. 3. Clean up old comments/commented out code. 4. Adjust unit tests to match with new code. 5. Adjust Complement setup infrastructure to only add main to the instance_map if workers are used and remove now unused options from the worker.yaml template. * Initial Docs upload * Changelog * Missed some commented out code that can go now * Remove TODO comment that no longer holds true. * Fix links in docs * More docs * Remove debug logging * Apply suggestions from code review Co-authored-by: reivilibre <olivier@librepush.net> * Apply suggestions from code review Co-authored-by: reivilibre <olivier@librepush.net> * Update version to latest, include completeish before/after examples in upgrade notes. * Fix up and docs too --------- Co-authored-by: reivilibre <olivier@librepush.net>
2023-05-11 04:30:56 -06:00
MAIN_PROCESS_INSTANCE_NAME = "main"
MAIN_PROCESS_LOCALHOST_ADDRESS = "127.0.0.1"
MAIN_PROCESS_REPLICATION_PORT = 9093
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
# Obviously, these would only be used with the UNIX socket option
MAIN_PROCESS_UNIX_SOCKET_PUBLIC_PATH = "/run/main_public.sock"
MAIN_PROCESS_UNIX_SOCKET_PRIVATE_PATH = "/run/main_private.sock"
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
@dataclass
class WorkerTemplate:
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources: Set[str] = field(default_factory=set)
endpoint_patterns: Set[str] = field(default_factory=set)
Tweak comments
2023-11-16 08:26:11 -07:00
# (worker_name) -> {config}
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
shared_extra_conf: Callable[[str], Dict[str, Any]] = lambda _worker_name: {}
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
worker_extra_conf: str = ""
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
# True if and only if multiple of this worker type are allowed.
sharding_allowed: bool = True
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
Tweak comments
2023-11-16 08:26:11 -07:00
# Workers with exposed endpoints needs either "client", "federation", or "media" listener_resources
# Watching /_matrix/client needs a "client" listener
# Watching /_matrix/federation needs a "federation" listener
# Watching /_matrix/media and related needs a "media" listener
# Stream Writers require "client" and "replication" listeners because they
# have to attach by instance_map to the master process and have client endpoints.
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
WORKERS_CONFIG: Dict[str, WorkerTemplate] = {
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
"pusher": WorkerTemplate(
shared_extra_conf=lambda worker_name: {
"pusher_instances": [worker_name],
}
),
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
"user_dir": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client"},
endpoint_patterns={
Add support for `/_matrix/media/v3` APIs (#11371) * Add support for `/_matrix/media/v3` APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Update `workers.md` to use v3 client and media APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Add changelog Signed-off-by: Aaron Raimist <aaron@raim.ist>
2021-11-17 08:30:24 -07:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/user_directory/search$"
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
shared_extra_conf=lambda worker_name: {
"update_user_directory_from_worker": worker_name
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"media_repository": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"media"},
endpoint_patterns={
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"^/_matrix/media/",
"^/_synapse/admin/v1/purge_media_cache$",
"^/_synapse/admin/v1/room/.*/media.*$",
"^/_synapse/admin/v1/user/.*/media.*$",
"^/_synapse/admin/v1/media/.*$",
"^/_synapse/admin/v1/quarantine_media/.*$",
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Modernize configure_workers_and_start.py bootstrapping script for Dockerfile-workers. (#14294)
2022-11-09 05:02:15 -07:00
# The first configured media worker will run the media background jobs
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
shared_extra_conf=lambda worker_name: {
Modernize configure_workers_and_start.py bootstrapping script for Dockerfile-workers. (#14294)
2022-11-09 05:02:15 -07:00
"enable_media_repo": False,
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
"media_instance_running_background_jobs": worker_name,
Modernize configure_workers_and_start.py bootstrapping script for Dockerfile-workers. (#14294)
2022-11-09 05:02:15 -07:00
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
worker_extra_conf="enable_media_repo: true",
),
"appservice": WorkerTemplate(
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
shared_extra_conf=lambda worker_name: {
"notify_appservices_from_worker": worker_name
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
"federation_sender": WorkerTemplate(
shared_extra_conf=lambda worker_name: {
"federation_sender_instances": [worker_name],
}
),
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
"synchrotron": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client"},
endpoint_patterns={
Add support for `/_matrix/media/v3` APIs (#11371) * Add support for `/_matrix/media/v3` APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Update `workers.md` to use v3 client and media APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Add changelog Signed-off-by: Aaron Raimist <aaron@raim.ist>
2021-11-17 08:30:24 -07:00
"^/_matrix/client/(v2_alpha|r0|v3)/sync$",
"^/_matrix/client/(api/v1|v2_alpha|r0|v3)/events$",
"^/_matrix/client/(api/v1|r0|v3)/initialSync$",
"^/_matrix/client/(api/v1|r0|v3)/rooms/[^/]+/initialSync$",
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"client_reader": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client"},
endpoint_patterns={
Mark /relations endpoint as usable on workers. (#14028) Co-authored-by: Eric Eastwood <erice@element.io>
2022-10-12 04:46:13 -06:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/publicRooms$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/joined_members$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/context/.*$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/members$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/state$",
"^/_matrix/client/v1/rooms/.*/hierarchy$",
"^/_matrix/client/(v1|unstable)/rooms/.*/relations/",
Stabilize the threads API. (#14175) Stabilize the threads API (MSC3856) by supporting (only) the v1 path for the endpoint. This also marks the API as safe for workers since it is a read-only API.
2022-10-14 05:16:50 -06:00
"^/_matrix/client/v1/rooms/.*/threads$",
Mark /relations endpoint as usable on workers. (#14028) Co-authored-by: Eric Eastwood <erice@element.io>
2022-10-12 04:46:13 -06:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/login$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/account/3pid$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/account/whoami$",
"^/_matrix/client/versions$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/voip/turnServer$",
"^/_matrix/client/(r0|v3|unstable)/register$",
Load `/register/available` endpoint on workers (#15268)
2023-03-17 07:50:31 -06:00
"^/_matrix/client/(r0|v3|unstable)/register/available$",
Mark /relations endpoint as usable on workers. (#14028) Co-authored-by: Eric Eastwood <erice@element.io>
2022-10-12 04:46:13 -06:00
"^/_matrix/client/(r0|v3|unstable)/auth/.*/fallback/web$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/messages$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/event",
"^/_matrix/client/(api/v1|r0|v3|unstable)/joined_rooms",
"^/_matrix/client/(api/v1|r0|v3|unstable/.*)/rooms/.*/aliases",
Move MSC3030 `/timestamp_to_event` endpoint to stable v1 location (#14471) Fix https://github.com/matrix-org/synapse/issues/14390 - Client API: `/_matrix/client/unstable/org.matrix.msc3030/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>` -> `/_matrix/client/v1/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>` - Federation API: `/_matrix/federation/unstable/org.matrix.msc3030/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>` -> `/_matrix/federation/v1/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>` Complement test changes: https://github.com/matrix-org/complement/pull/559
2022-11-28 14:54:18 -07:00
"^/_matrix/client/v1/rooms/.*/timestamp_to_event$",
Mark /relations endpoint as usable on workers. (#14028) Co-authored-by: Eric Eastwood <erice@element.io>
2022-10-12 04:46:13 -06:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/search",
Allow use of the `/filter` Client-Server APIs on workers. (#15134)
2023-02-28 09:37:19 -07:00
"^/_matrix/client/(r0|v3|unstable)/user/.*/filter(/|$)",
Load `/password_policy` endpoint on workers. (#15331)
2023-03-27 05:37:17 -06:00
"^/_matrix/client/(r0|v3|unstable)/password_policy$",
Load `/directory/room/{roomAlias}` endpoint on workers (#15333) * Enable `directory` * move to worker store * newsfile * disable `ClientDirectoryListServer` and `ClientAppserviceDirectoryListServer` for workers
2023-04-14 03:24:06 -06:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/directory/room/.*$",
Load `/capabilities` endpoint on workers (#15436)
2023-04-14 10:26:07 -06:00
"^/_matrix/client/(r0|v3|unstable)/capabilities$",
Add `/notifications` endpoint to workers (#16265)
2023-09-07 03:26:07 -06:00
"^/_matrix/client/(r0|v3|unstable)/notifications$",
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"federation_reader": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"federation"},
endpoint_patterns={
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"^/_matrix/federation/(v1|v2)/event/",
"^/_matrix/federation/(v1|v2)/state/",
"^/_matrix/federation/(v1|v2)/state_ids/",
"^/_matrix/federation/(v1|v2)/backfill/",
"^/_matrix/federation/(v1|v2)/get_missing_events/",
"^/_matrix/federation/(v1|v2)/publicRooms",
"^/_matrix/federation/(v1|v2)/query/",
"^/_matrix/federation/(v1|v2)/make_join/",
"^/_matrix/federation/(v1|v2)/make_leave/",
"^/_matrix/federation/(v1|v2)/send_join/",
"^/_matrix/federation/(v1|v2)/send_leave/",
"^/_matrix/federation/(v1|v2)/invite/",
"^/_matrix/federation/(v1|v2)/query_auth/",
"^/_matrix/federation/(v1|v2)/event_auth/",
Move MSC3030 `/timestamp_to_event` endpoint to stable v1 location (#14471) Fix https://github.com/matrix-org/synapse/issues/14390 - Client API: `/_matrix/client/unstable/org.matrix.msc3030/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>` -> `/_matrix/client/v1/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>` - Federation API: `/_matrix/federation/unstable/org.matrix.msc3030/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>` -> `/_matrix/federation/v1/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>` Complement test changes: https://github.com/matrix-org/complement/pull/559
2022-11-28 14:54:18 -07:00
"^/_matrix/federation/v1/timestamp_to_event/",
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"^/_matrix/federation/(v1|v2)/exchange_third_party_invite/",
"^/_matrix/federation/(v1|v2)/user/devices/",
"^/_matrix/federation/(v1|v2)/get_groups_publicised$",
"^/_matrix/key/v2/query",
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"federation_inbound": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"federation"},
endpoint_patterns={"/_matrix/federation/(v1|v2)/send/"},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"event_persister": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"replication"},
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
shared_extra_conf=lambda worker_name: {
"stream_writers": {"events": [worker_name]}
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"background_worker": WorkerTemplate(
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# This worker cannot be sharded. Therefore, there should only ever be one
# background worker. This is enforced for the safety of your database.
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
shared_extra_conf=lambda worker_name: {"run_background_tasks_on": worker_name},
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
sharding_allowed=False,
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"event_creator": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client"},
endpoint_patterns={
Add support for `/_matrix/media/v3` APIs (#11371) * Add support for `/_matrix/media/v3` APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Update `workers.md` to use v3 client and media APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Add changelog Signed-off-by: Aaron Raimist <aaron@raim.ist>
2021-11-17 08:30:24 -07:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/redact",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/send",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/(join|invite|leave|ban|unban|kick)$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/join/",
Add support for knocking to workers. (#15133)
2023-03-02 10:59:53 -07:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/knock/",
Add support for `/_matrix/media/v3` APIs (#11371) * Add support for `/_matrix/media/v3` APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Update `workers.md` to use v3 client and media APIs Signed-off-by: Aaron Raimist <aaron@raim.ist> * Add changelog Signed-off-by: Aaron Raimist <aaron@raim.ist>
2021-11-17 08:30:24 -07:00
"^/_matrix/client/(api/v1|r0|v3|unstable)/profile/",
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"frontend_proxy": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client", "replication"},
endpoint_patterns={"^/_matrix/client/(api/v1|r0|v3|unstable)/keys/upload"},
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"account_data": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client", "replication"},
endpoint_patterns={
Add all Stream Writer worker types to configure_workers_and_start.py (#14197) Co-authored-by: reivilibre <oliverw@matrix.org>
2022-11-08 06:14:00 -07:00
"^/_matrix/client/(r0|v3|unstable)/.*/tags",
"^/_matrix/client/(r0|v3|unstable)/.*/account_data",
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
shared_extra_conf=lambda worker_name: {
"stream_writers": {"account_data": [worker_name]}
},
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
sharding_allowed=False,
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"presence": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client", "replication"},
endpoint_patterns={"^/_matrix/client/(api/v1|r0|v3|unstable)/presence/"},
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
shared_extra_conf=lambda worker_name: {
"stream_writers": {"presence": [worker_name]}
},
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
sharding_allowed=False,
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"receipts": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client", "replication"},
endpoint_patterns={
Add all Stream Writer worker types to configure_workers_and_start.py (#14197) Co-authored-by: reivilibre <oliverw@matrix.org>
2022-11-08 06:14:00 -07:00
"^/_matrix/client/(r0|v3|unstable)/rooms/.*/receipt",
"^/_matrix/client/(r0|v3|unstable)/rooms/.*/read_markers",
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
},
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
shared_extra_conf=lambda worker_name: {
"stream_writers": {"receipts": [worker_name]}
},
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
sharding_allowed=False,
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"to_device": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client", "replication"},
endpoint_patterns={"^/_matrix/client/(r0|v3|unstable)/sendToDevice/"},
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
shared_extra_conf=lambda worker_name: {
"stream_writers": {"to_device": [worker_name]}
},
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
sharding_allowed=False,
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
"typing": WorkerTemplate(
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
listener_resources={"client", "replication"},
endpoint_patterns={"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/typing"},
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
shared_extra_conf=lambda worker_name: {
"stream_writers": {"typing": [worker_name]}
},
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
sharding_allowed=False,
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
),
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
}
# Templates for sections that may be inserted multiple times in config files
NGINX_LOCATION_CONFIG_BLOCK = """
Run complement with Synapse workers manually. (#10039) Adds an option to complement.sh to run Synapse in worker mode (instead of the default monolith mode).
2021-05-24 13:32:45 -06:00
location ~* {endpoint} {{
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
proxy_pass {upstream};
proxy_set_header X-Forwarded-For $remote_addr;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header Host $host;
Run complement with Synapse workers manually. (#10039) Adds an option to complement.sh to run Synapse in worker mode (instead of the default monolith mode).
2021-05-24 13:32:45 -06:00
}}
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"""
NGINX_UPSTREAM_CONFIG_BLOCK = """
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
upstream {upstream_worker_base_name} {{
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
{body}
Run complement with Synapse workers manually. (#10039) Adds an option to complement.sh to run Synapse in worker mode (instead of the default monolith mode).
2021-05-24 13:32:45 -06:00
}}
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"""
# Utility functions
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
def log(txt: str) -> None:
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
print(txt)
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
def error(txt: str) -> NoReturn:
Flush stdout/err in Dockerfile-workers before replacing the current process (#14195) Also update `subprocess.check_output` to the slightly newer `subprocess.run`. Signed-off-by: Jason Little <realtyem@gmail.com>
2022-10-18 05:56:20 -06:00
print(txt, file=sys.stderr)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
sys.exit(2)
Flush stdout/err in Dockerfile-workers before replacing the current process (#14195) Also update `subprocess.check_output` to the slightly newer `subprocess.run`. Signed-off-by: Jason Little <realtyem@gmail.com>
2022-10-18 05:56:20 -06:00
def flush_buffers() -> None:
sys.stdout.flush()
sys.stderr.flush()
Add `merge_into`
2023-11-16 08:26:17 -07:00
def merge_into(dest: Any, new: Any) -> None:
"""
Merges `new` into `dest` with the following rules:
- dicts: values with the same key will be merged recursively
- lists: `new` will be appended to `dest`
- primitives: they will be checked for equality and inequality will result
in a ValueError
It is an error for `dest` and `new` to be of different types.
"""
if isinstance(dest, dict) and isinstance(new, dict):
for k, v in new.items():
if k in dest:
merge_into(dest[k], v)
else:
dest[k] = v
elif isinstance(dest, list) and isinstance(new, list):
dest.extend(new)
elif type(dest) != type(new):
raise TypeError(f"Cannot merge {type(dest).__name__} and {type(new).__name__}")
elif dest != new:
raise ValueError(f"Cannot merge primitive values: {dest!r} != {new!r}")
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
def merged(a: Dict[str, Any], b: Dict[str, Any]) -> Dict[str, Any]:
"""
Merges `b` into `a` and returns `a`.
"""
merge_into(a, b)
return a
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
def convert(src: str, dst: str, **template_vars: object) -> None:
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"""Generate a file from a template
Args:
src: Path to the input file.
dst: Path to write to.
template_vars: The arguments to replace placeholder variables in the template with.
"""
# Read the template file
Merge the Complement testing Docker images into a single, multi-purpose image. (#12881) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-08 03:57:05 -06:00
# We disable autoescape to prevent template variables from being escaped,
# as we're not using HTML.
env = Environment(loader=FileSystemLoader(os.path.dirname(src)), autoescape=False)
template = env.get_template(os.path.basename(src))
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Merge the Complement testing Docker images into a single, multi-purpose image. (#12881) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-08 03:57:05 -06:00
# Generate a string from the template.
rendered = template.render(**template_vars)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Write the generated contents to a file
#
# We use append mode in case the files have already been written to by something else
# (for instance, as part of the instructions in a dockerfile).
with open(dst, "a") as outfile:
# In case the existing file doesn't end with a newline
outfile.write("\n")
outfile.write(rendered)
Rename function to add_worker_to_instance_map given reduction of scope
2023-11-16 08:49:20 -07:00
def add_worker_to_instance_map(
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
shared_config: dict,
worker_name: str,
worker_port: int,
) -> None:
Rename function to add_worker_to_instance_map given reduction of scope
2023-11-16 08:49:20 -07:00
"""
Update the shared config map to add the worker in the instance_map.
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Args:
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
shared_config: The config dict that all worker instances share (after being
converted to YAML)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
worker_name: The name of the worker instance.
worker_port: The HTTP replication port that the worker instance is listening on.
"""
instance_map = shared_config.setdefault("instance_map", {})
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
if os.environ.get("SYNAPSE_USE_UNIX_SOCKET", False):
instance_map[worker_name] = {
"path": f"/run/worker.{worker_port}",
}
else:
instance_map[worker_name] = {
"host": "localhost",
"port": worker_port,
}
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
def merge_worker_template_configs(
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
existing_template: WorkerTemplate,
to_be_merged_template: WorkerTemplate,
) -> WorkerTemplate:
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
"""When given an existing dict of worker template configuration consisting with both
dicts and lists, merge new template data from WORKERS_CONFIG(or create) and
return new dict.
Add all Stream Writer worker types to configure_workers_and_start.py (#14197) Co-authored-by: reivilibre <oliverw@matrix.org>
2022-11-08 06:14:00 -07:00
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
Args:
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
existing_template: Either an existing worker template or a fresh blank one.
to_be_merged_template: The template from WORKERS_CONFIGS to be merged into
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
existing_dict.
Returns: The newly merged together dict values.
"""
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
# copy existing_template without any replacements
new_template: WorkerTemplate = dataclasses.replace(existing_template)
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
# add listener resources from the other set
new_template.listener_resources |= to_be_merged_template.listener_resources
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
# add endpoint patterns from the other set
new_template.endpoint_patterns |= to_be_merged_template.endpoint_patterns
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
# merge dictionaries; the worker name will be replaced later
Remove special logic for adding stream_writers: just make it part of the extra config template
2023-11-16 08:47:28 -07:00
new_template.shared_extra_conf = lambda worker_name: merged(
existing_template.shared_extra_conf(worker_name),
to_be_merged_template.shared_extra_conf(worker_name),
)
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
# There is only one worker type that has a 'worker_extra_conf' and it is
# the media_repo. Since duplicate worker types on the same worker don't
# work, this is fine.
new_template.worker_extra_conf = (
new_template.worker_extra_conf + to_be_merged_template.worker_extra_conf
)
return new_template
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
def insert_worker_name_for_worker_config(
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
existing_template: WorkerTemplate, worker_name: str
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
) -> Dict[str, Any]:
"""Insert a given worker name into the worker's configuration dict.
Args:
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
existing_template: The WorkerTemplate that is imported into shared_config.
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
worker_name: The name of the worker to insert.
Returns: Copy of the dict with newly inserted worker name
"""
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
dict_to_edit = dataclasses.asdict(existing_template)
Use a lambda for the worker name rather than search and replace later
2023-11-16 08:05:39 -07:00
dict_to_edit["shared_extra_conf"] = existing_template.shared_extra_conf(worker_name)
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
dict_to_edit["endpoint_patterns"] = sorted(existing_template.endpoint_patterns)
dict_to_edit["listener_resources"] = sorted(existing_template.listener_resources)
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
return dict_to_edit
def apply_requested_multiplier_for_worker(worker_types: List[str]) -> List[str]:
"""
Apply multiplier(if found) by returning a new expanded list with some basic error
checking.
Args:
worker_types: The unprocessed List of requested workers
Returns:
A new list with all requested workers expanded.
"""
# Checking performed:
# 1. if worker:2 or more is declared, it will create additional workers up to number
# 2. if worker:1, it will create a single copy of this worker as if no number was
# given
# 3. if worker:0 is declared, this worker will be ignored. This is to allow for
# scripting and automated expansion and is intended behaviour.
# 4. if worker:NaN or is a negative number, it will error and log it.
new_worker_types = []
for worker_type in worker_types:
if ":" in worker_type:
worker_type_components = split_and_strip_string(worker_type, ":", 1)
worker_count = 0
# Should only be 2 components, a type of worker(s) and an integer as a
# string. Cast the number as an int then it can be used as a counter.
try:
worker_count = int(worker_type_components[1])
except ValueError:
error(
f"Bad number in worker count for '{worker_type}': "
f"'{worker_type_components[1]}' is not an integer"
)
# As long as there are more than 0, we add one to the list to make below.
for _ in range(worker_count):
new_worker_types.append(worker_type_components[0])
else:
# If it's not a real worker_type, it will error out later.
new_worker_types.append(worker_type)
return new_worker_types
def split_and_strip_string(
given_string: str, split_char: str, max_split: SupportsIndex = -1
) -> List[str]:
"""
Helper to split a string on split_char and strip whitespace from each end of each
element.
Args:
given_string: The string to split
split_char: The character to split the string on
max_split: kwarg for split() to limit how many times the split() happens
Returns:
A List of strings
"""
# Removes whitespace from ends of result strings before adding to list. Allow for
# overriding 'maxsplit' kwarg, default being -1 to signify no maximum.
return [x.strip() for x in given_string.split(split_char, maxsplit=max_split)]
Add all Stream Writer worker types to configure_workers_and_start.py (#14197) Co-authored-by: reivilibre <oliverw@matrix.org>
2022-11-08 06:14:00 -07:00
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
def generate_base_homeserver_config() -> None:
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"""Starts Synapse and generates a basic homeserver config, which will later be
modified for worker support.
Raises: CalledProcessError if calling start.py returned a non-zero exit code.
"""
# start.py already does this for us, so just call that.
# note that this script is copied in in the official, monolith dockerfile
os.environ["SYNAPSE_HTTP_PORT"] = str(MAIN_PROCESS_HTTP_LISTENER_PORT)
Flush stdout/err in Dockerfile-workers before replacing the current process (#14195) Also update `subprocess.check_output` to the slightly newer `subprocess.run`. Signed-off-by: Jason Little <realtyem@gmail.com>
2022-10-18 05:56:20 -06:00
subprocess.run(["/usr/local/bin/python", "/start.py", "migrate_config"], check=True)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
def parse_worker_types(
requested_worker_types: List[str],
) -> Dict[str, Set[str]]:
"""Read the desired list of requested workers and prepare the data for use in
generating worker config files while also checking for potential gotchas.
Args:
requested_worker_types: The list formed from the split environment variable
containing the unprocessed requests for workers.
Returns: A dict of worker names to set of worker types. Format:
{'worker_name':
{'worker_type', 'worker_type2'}
}
"""
# A counter of worker_base_name -> int. Used for determining the name for a given
# worker when generating its config file, as each worker's name is just
# worker_base_name followed by instance number
worker_base_name_counter: Dict[str, int] = defaultdict(int)
# Similar to above, but more finely grained. This is used to determine we don't have
# more than a single worker for cases where multiples would be bad(e.g. presence).
worker_type_shard_counter: Dict[str, int] = defaultdict(int)
# The final result of all this processing
dict_to_return: Dict[str, Set[str]] = {}
# Handle any multipliers requested for given workers.
multiple_processed_worker_types = apply_requested_multiplier_for_worker(
requested_worker_types
)
# Process each worker_type_string
# Examples of expected formats:
# - requested_name=type1+type2+type3
# - synchrotron
# - event_creator+event_persister
for worker_type_string in multiple_processed_worker_types:
# First, if a name is requested, use that — otherwise generate one.
worker_base_name: str = ""
if "=" in worker_type_string:
# Split on "=", remove extra whitespace from ends then make list
worker_type_split = split_and_strip_string(worker_type_string, "=")
if len(worker_type_split) > 2:
error(
"There should only be one '=' in the worker type string. "
f"Please fix: {worker_type_string}"
)
# Assign the name
worker_base_name = worker_type_split[0]
if not re.match(r"^[a-zA-Z0-9_+-]*[a-zA-Z_+-]$", worker_base_name):
# Apply a fairly narrow regex to the worker names. Some characters
# aren't safe for use in file paths or nginx configurations.
# Don't allow to end with a number because we'll add a number
# ourselves in a moment.
error(
"Invalid worker name; please choose a name consisting of "
"alphanumeric letters, _ + -, but not ending with a digit: "
f"{worker_base_name!r}"
)
# Continue processing the remainder of the worker_type string
# with the name override removed.
worker_type_string = worker_type_split[1]
# Split the worker_type_string on "+", remove whitespace from ends then make
# the list a set so it's deduplicated.
worker_types_set: Set[str] = set(
split_and_strip_string(worker_type_string, "+")
)
if not worker_base_name:
# No base name specified: generate one deterministically from set of
# types
worker_base_name = "+".join(sorted(worker_types_set))
# At this point, we have:
# worker_base_name which is the name for the worker, without counter.
# worker_types_set which is the set of worker types for this worker.
# Validate worker_type and make sure we don't allow sharding for a worker type
# that doesn't support it. Will error and stop if it is a problem,
# e.g. 'background_worker'.
for worker_type in worker_types_set:
# Verify this is a real defined worker type. If it's not, stop everything so
# it can be fixed.
if worker_type not in WORKERS_CONFIG:
error(
f"{worker_type} is an unknown worker type! Was found in "
f"'{worker_type_string}'. Please fix!"
)
if worker_type in worker_type_shard_counter:
Add `sharding_allowed` to the WorkerTemplate rather than having a separate function for that
2023-11-16 08:52:52 -07:00
if not WORKERS_CONFIG[worker_type].sharding_allowed:
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
error(
f"There can be only a single worker with {worker_type} "
"type. Please recount and remove."
)
# Not in shard counter, must not have seen it yet, add it.
worker_type_shard_counter[worker_type] += 1
# Generate the number for the worker using incrementing counter
worker_base_name_counter[worker_base_name] += 1
worker_number = worker_base_name_counter[worker_base_name]
worker_name = f"{worker_base_name}{worker_number}"
if worker_number > 1:
# If this isn't the first worker, check that we don't have a confusing
# mixture of worker types with the same base name.
first_worker_with_base_name = dict_to_return[f"{worker_base_name}1"]
if first_worker_with_base_name != worker_types_set:
error(
f"Can not use worker_name: '{worker_name}' for worker_type(s): "
f"{worker_types_set!r}. It is already in use by "
f"worker_type(s): {first_worker_with_base_name!r}"
)
dict_to_return[worker_name] = worker_types_set
return dict_to_return
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
def generate_worker_files(
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
environ: Mapping[str, str],
config_path: str,
data_dir: str,
requested_worker_types: Dict[str, Set[str]],
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
) -> None:
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
"""Read the desired workers(if any) that is passed in and generate shared
homeserver, nginx and supervisord configs.
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Args:
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
environ: os.environ instance.
Fix up healthcheck generation for workers docker image (#12405) This wasn't quite generating the right thing.
2022-04-11 07:38:58 -06:00
config_path: The location of the generated Synapse main worker config file.
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
data_dir: The location of the synapse data directory. Where log and
user-facing config files live.
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
requested_worker_types: A Dict containing requested workers in the format of
{'worker_name1': {'worker_type', ...}}
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
"""
# Note that yaml cares about indentation, so care should be taken to insert lines
# into files at the correct indentation below.
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
# Convenience helper for if using unix sockets instead of host:port
using_unix_sockets = environ.get("SYNAPSE_USE_UNIX_SOCKET", False)
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# First read the original config file and extract the listeners block. Then we'll
# add another listener for replication. Later we'll write out the result to the
# shared config file.
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
listeners: List[Any]
if using_unix_sockets:
listeners = [
{
"path": MAIN_PROCESS_UNIX_SOCKET_PRIVATE_PATH,
"type": "http",
"resources": [{"names": ["replication"]}],
}
]
else:
listeners = [
{
"port": MAIN_PROCESS_REPLICATION_PORT,
"bind_address": MAIN_PROCESS_LOCALHOST_ADDRESS,
"type": "http",
"resources": [{"names": ["replication"]}],
}
]
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
with open(config_path) as file_stream:
original_config = yaml.safe_load(file_stream)
original_listeners = original_config.get("listeners")
if original_listeners:
listeners += original_listeners
# The shared homeserver config. The contents of which will be inserted into the
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# base shared worker jinja2 template. This config file will be passed to all
# workers, included Synapse's main process. It is intended mainly for disabling
# functionality when certain workers are spun up, and adding a replication listener.
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
shared_config: Dict[str, Any] = {"listeners": listeners}
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Refactor the Dockerfile-workers configuration script to use Jinja2 templates in Synapse workers' Supervisord blocks. (#13054) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-27 04:43:20 -06:00
# List of dicts that describe workers.
# We pass this to the Supervisor template later to generate the appropriate
# program blocks.
worker_descriptors: List[Dict[str, Any]] = []
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# Upstreams for load-balancing purposes. This dict takes the form of the worker
# type to the ports of each worker. For example:
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# {
# worker_type: {1234, 1235, ...}}
# }
# and will be used to construct 'upstream' nginx directives.
Add missing type definitions for scripts in docker folder (#12280) Signed-off-by: Jorge Florian <jafn28@gmail.com>
2022-04-08 04:10:58 -06:00
nginx_upstreams: Dict[str, Set[int]] = {}
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# A map of: {"endpoint": "upstream"}, where "upstream" is a str representing what
# will be placed after the proxy_pass directive. The main benefit to representing
# this data as a dict over a str is that we can easily deduplicate endpoints
# across multiple instances of the same worker. The final rendering will be combined
# with nginx_upstreams and placed in /etc/nginx/conf.d.
nginx_locations: Dict[str, str] = {}
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Create the worker configuration directory if it doesn't already exist
os.makedirs("/conf/workers", exist_ok=True)
# Start worker ports from this arbitrary port
worker_port = 18009
Fix up healthcheck generation for workers docker image (#12405) This wasn't quite generating the right thing.
2022-04-11 07:38:58 -06:00
# A list of internal endpoints to healthcheck, starting with the main process
# which exists even if no workers do.
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
# This list ends up being part of the command line to curl, (curl added support for
# Unix sockets in version 7.40).
if using_unix_sockets:
healthcheck_urls = [
f"--unix-socket {MAIN_PROCESS_UNIX_SOCKET_PUBLIC_PATH} "
# The scheme and hostname from the following URL are ignored.
# The only thing that matters is the path `/health`
"http://localhost/health"
]
else:
healthcheck_urls = ["http://localhost:8080/health"]
Fix up healthcheck generation for workers docker image (#12405) This wasn't quite generating the right thing.
2022-04-11 07:38:58 -06:00
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# Get the set of all worker types that we have configured
all_worker_types_in_use = set(chain(*requested_worker_types.values()))
# Map locations to upstreams (corresponding to worker types) in Nginx
# but only if we use the appropriate worker type
for worker_type in all_worker_types_in_use:
Convert listener_resources and endpoint_patterns to Set[str]
2023-11-16 08:12:18 -07:00
for endpoint_pattern in sorted(WORKERS_CONFIG[worker_type].endpoint_patterns):
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
nginx_locations[endpoint_pattern] = f"http://{worker_type}"
# For each worker type specified by the user, create config values and write it's
# yaml config file
for worker_name, worker_types_set in requested_worker_types.items():
# The collected and processed data will live here.
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
worker_template: WorkerTemplate = WorkerTemplate()
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# Merge all worker config templates for this worker into a single config
for worker_type in worker_types_set:
# Merge worker type template configuration data. It's a combination of lists
# and dicts, so use this helper.
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
worker_template = merge_worker_template_configs(
worker_template, WORKERS_CONFIG[worker_type]
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
)
# Replace placeholder names in the config template with the actual worker name.
Convert worker templates into dataclass
2023-11-16 08:00:48 -07:00
worker_config: Dict[str, Any] = insert_worker_name_for_worker_config(
worker_template, worker_name
)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
worker_config.update(
Add missing type definitions for scripts in docker folder (#12280) Signed-off-by: Jorge Florian <jafn28@gmail.com>
2022-04-08 04:10:58 -06:00
{"name": worker_name, "port": str(worker_port), "config_path": config_path}
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
)
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# Update the shared config with any worker_type specific options. The first of a
# given worker_type needs to stay assigned and not be replaced.
worker_config["shared_extra_conf"].update(shared_config)
shared_config = worker_config["shared_extra_conf"]
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
if using_unix_sockets:
healthcheck_urls.append(
f"--unix-socket /run/worker.{worker_port} http://localhost/health"
)
else:
healthcheck_urls.append("http://localhost:%d/health" % (worker_port,))
Fix up healthcheck generation for workers docker image (#12405) This wasn't quite generating the right thing.
2022-04-11 07:38:58 -06:00
Rename function to add_worker_to_instance_map given reduction of scope
2023-11-16 08:49:20 -07:00
# Add all workers to the `instance_map`
# Technically only certain types of workers, such as stream writers, are needed
# here but it is simpler just to be consistent.
add_worker_to_instance_map(shared_config, worker_name, worker_port)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Enable the worker in supervisord
Refactor the Dockerfile-workers configuration script to use Jinja2 templates in Synapse workers' Supervisord blocks. (#13054) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-27 04:43:20 -06:00
worker_descriptors.append(worker_config)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Write out the worker's logging config file
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
log_config_filepath = generate_worker_log_config(environ, worker_name, data_dir)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Then a worker config file
convert(
"/conf/worker.yaml.j2",
Run pyupgrade for python 3.7 & 3.8. (#16110)
2023-08-15 06:11:20 -06:00
f"/conf/workers/{worker_name}.yaml",
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
**worker_config,
worker_log_config_filepath=log_config_filepath,
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
using_unix_sockets=using_unix_sockets,
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
)
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# Save this worker's port number to the correct nginx upstreams
for worker_type in worker_types_set:
nginx_upstreams.setdefault(worker_type, set()).add(worker_port)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
worker_port += 1
# Build the nginx location config blocks
nginx_location_config = ""
for endpoint, upstream in nginx_locations.items():
nginx_location_config += NGINX_LOCATION_CONFIG_BLOCK.format(
endpoint=endpoint,
upstream=upstream,
)
# Determine the load-balancing upstreams to configure
nginx_upstream_config = ""
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
for upstream_worker_base_name, upstream_worker_ports in nginx_upstreams.items():
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
body = ""
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
if using_unix_sockets:
for port in upstream_worker_ports:
body += f" server unix:/run/worker.{port};\n"
else:
for port in upstream_worker_ports:
body += f" server localhost:{port};\n"
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Add to the list of configured upstreams
nginx_upstream_config += NGINX_UPSTREAM_CONFIG_BLOCK.format(
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
upstream_worker_base_name=upstream_worker_base_name,
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
body=body,
)
# Finally, we'll write out the config files.
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
# log config for the master process
master_log_config = generate_worker_log_config(environ, "master", data_dir)
shared_config["log_config"] = master_log_config
Support registering Application Services when running with workers under Complement. (#12826) Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
2022-05-23 07:11:06 -06:00
# Find application service registrations
appservice_registrations = None
appservice_registration_dir = os.environ.get("SYNAPSE_AS_REGISTRATION_DIR")
if appservice_registration_dir:
# Scan for all YAML files that should be application service registrations.
appservice_registrations = [
str(reg_path.resolve())
for reg_path in Path(appservice_registration_dir).iterdir()
if reg_path.suffix.lower() in (".yaml", ".yml")
]
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
workers_in_use = len(requested_worker_types) > 0
Merge the Complement testing Docker images into a single, multi-purpose image. (#12881) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-08 03:57:05 -06:00
Remove `worker_replication_*` settings (#15491) * Add master to the instance_map as part of Complement, have ReplicationEndpoint look at instance_map for master. * Fix typo in drive by. * Remove unnecessary worker_replication_* bits from unit tests and add master to instance_map(hopefully in the right place) * Several updates: 1. Switch from master to main for naming the main process in the instance_map. Add useful constants for easier adjustment of names in the future. 2. Add backwards compatibility for worker_replication_* to allow time to transition to new style. Make sure to prioritize declaring main directly on the instance_map. 3. Clean up old comments/commented out code. 4. Adjust unit tests to match with new code. 5. Adjust Complement setup infrastructure to only add main to the instance_map if workers are used and remove now unused options from the worker.yaml template. * Initial Docs upload * Changelog * Missed some commented out code that can go now * Remove TODO comment that no longer holds true. * Fix links in docs * More docs * Remove debug logging * Apply suggestions from code review Co-authored-by: reivilibre <olivier@librepush.net> * Apply suggestions from code review Co-authored-by: reivilibre <olivier@librepush.net> * Update version to latest, include completeish before/after examples in upgrade notes. * Fix up and docs too --------- Co-authored-by: reivilibre <olivier@librepush.net>
2023-05-11 04:30:56 -06:00
# If there are workers, add the main process to the instance_map too.
if workers_in_use:
instance_map = shared_config.setdefault("instance_map", {})
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
if using_unix_sockets:
instance_map[MAIN_PROCESS_INSTANCE_NAME] = {
"path": MAIN_PROCESS_UNIX_SOCKET_PRIVATE_PATH,
}
else:
instance_map[MAIN_PROCESS_INSTANCE_NAME] = {
"host": MAIN_PROCESS_LOCALHOST_ADDRESS,
"port": MAIN_PROCESS_REPLICATION_PORT,
}
Remove `worker_replication_*` settings (#15491) * Add master to the instance_map as part of Complement, have ReplicationEndpoint look at instance_map for master. * Fix typo in drive by. * Remove unnecessary worker_replication_* bits from unit tests and add master to instance_map(hopefully in the right place) * Several updates: 1. Switch from master to main for naming the main process in the instance_map. Add useful constants for easier adjustment of names in the future. 2. Add backwards compatibility for worker_replication_* to allow time to transition to new style. Make sure to prioritize declaring main directly on the instance_map. 3. Clean up old comments/commented out code. 4. Adjust unit tests to match with new code. 5. Adjust Complement setup infrastructure to only add main to the instance_map if workers are used and remove now unused options from the worker.yaml template. * Initial Docs upload * Changelog * Missed some commented out code that can go now * Remove TODO comment that no longer holds true. * Fix links in docs * More docs * Remove debug logging * Apply suggestions from code review Co-authored-by: reivilibre <olivier@librepush.net> * Apply suggestions from code review Co-authored-by: reivilibre <olivier@librepush.net> * Update version to latest, include completeish before/after examples in upgrade notes. * Fix up and docs too --------- Co-authored-by: reivilibre <olivier@librepush.net>
2023-05-11 04:30:56 -06:00
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Shared homeserver config
convert(
"/conf/shared.yaml.j2",
"/conf/workers/shared.yaml",
shared_worker_config=yaml.dump(shared_config),
Support registering Application Services when running with workers under Complement. (#12826) Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
2022-05-23 07:11:06 -06:00
appservice_registrations=appservice_registrations,
Merge the Complement testing Docker images into a single, multi-purpose image. (#12881) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-08 03:57:05 -06:00
enable_redis=workers_in_use,
workers_in_use=workers_in_use,
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
using_unix_sockets=using_unix_sockets,
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
)
# Nginx config
convert(
"/conf/nginx.conf.j2",
"/etc/nginx/conf.d/matrix-synapse.conf",
worker_locations=nginx_location_config,
upstream_directives=nginx_upstream_config,
Remove Caddy from the Synapse workers image used in Complement. (#12818)
2022-05-23 03:29:24 -06:00
tls_cert_path=os.environ.get("SYNAPSE_TLS_CERT"),
tls_key_path=os.environ.get("SYNAPSE_TLS_KEY"),
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
using_unix_sockets=using_unix_sockets,
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
)
# Supervisord config
Dockerfile-workers: reduce the amount we install (#12464) This is an attempt to reduce the rebuild time. In short, we reduce the amount of stuff that the dockerfile installs, so as to give a faster startup.
2022-04-14 08:36:49 -06:00
os.makedirs("/etc/supervisor", exist_ok=True)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
convert(
"/conf/supervisord.conf.j2",
Dockerfile-workers: reduce the amount we install (#12464) This is an attempt to reduce the rebuild time. In short, we reduce the amount of stuff that the dockerfile installs, so as to give a faster startup.
2022-04-14 08:36:49 -06:00
"/etc/supervisor/supervisord.conf",
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
main_config_path=config_path,
Merge the Complement testing Docker images into a single, multi-purpose image. (#12881) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-08 03:57:05 -06:00
enable_redis=workers_in_use,
Unix Sockets for HTTP Replication (#15708) Unix socket support for `federation` and `client` Listeners has existed now for a little while(since [1.81.0](https://github.com/matrix-org/synapse/pull/15353)), but there was one last hold out before it could be complete: HTTP Replication communication. This should finish it up. The Listeners would have always worked, but would have had no way to be talked to/at. --------- Co-authored-by: Eric Eastwood <madlittlemods@gmail.com> Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org> Co-authored-by: Eric Eastwood <erice@element.io>
2023-07-11 12:08:06 -06:00
using_unix_sockets=using_unix_sockets,
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
)
Refactor the Dockerfile-workers configuration script to use Jinja2 templates in Synapse workers' Supervisord blocks. (#13054) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-27 04:43:20 -06:00
convert(
"/conf/synapse.supervisord.conf.j2",
"/etc/supervisor/conf.d/synapse.conf",
workers=worker_descriptors,
main_config_path=config_path,
Improve startup times in Complement test runs against workers, particularly in CPU-constrained environments. (#13127) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-30 05:58:12 -06:00
use_forking_launcher=environ.get("SYNAPSE_USE_EXPERIMENTAL_FORKING_LAUNCHER"),
Refactor the Dockerfile-workers configuration script to use Jinja2 templates in Synapse workers' Supervisord blocks. (#13054) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-27 04:43:20 -06:00
)
Create healthcheck script for synapse-workers container (#11429) The intent is to iterate through all the worker ports and only report healthy when all are healthy, starting with the main process.
2021-11-26 07:05:20 -07:00
# healthcheck config
convert(
"/conf/healthcheck.sh.j2",
"/healthcheck.sh",
healthcheck_urls=healthcheck_urls,
)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Ensure the logging directory exists
log_dir = data_dir + "/logs"
if not os.path.exists(log_dir):
os.mkdir(log_dir)
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
def generate_worker_log_config(
environ: Mapping[str, str], worker_name: str, data_dir: str
) -> str:
"""Generate a log.config file for the given worker.
Returns: the path to the generated file
"""
# Check whether we should write worker logs to disk, in addition to the console
Add the ability to set the log level using the `SYNAPSE_TEST_LOG_LEVEL` environment when using `complement.sh`. (#13152)
2022-07-05 03:46:20 -06:00
extra_log_template_args: Dict[str, Optional[str]] = {}
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
if environ.get("SYNAPSE_WORKERS_WRITE_LOGS_TO_DISK"):
Add the ability to set the log level using the `SYNAPSE_TEST_LOG_LEVEL` environment when using `complement.sh`. (#13152)
2022-07-05 03:46:20 -06:00
extra_log_template_args["LOG_FILE_PATH"] = f"{data_dir}/logs/{worker_name}.log"
extra_log_template_args["SYNAPSE_LOG_LEVEL"] = environ.get("SYNAPSE_LOG_LEVEL")
extra_log_template_args["SYNAPSE_LOG_SENSITIVE"] = environ.get(
"SYNAPSE_LOG_SENSITIVE"
)
Log when events are (unexpectedly) filtered out of responses in tests (#14213) See https://github.com/matrix-org/synapse/pull/14095#discussion_r990335492 This is useful because when see that a relevant event is an `outlier` or `soft-failed`, then that's a good unexpected indicator explaining why it's not showing up. `filter_events_for_client` is used in `/sync`, `/messages`, `/context` which are all common end-to-end assertion touch points (also notifications, relations).
2023-06-01 20:27:18 -06:00
extra_log_template_args["SYNAPSE_LOG_TESTING"] = environ.get("SYNAPSE_LOG_TESTING")
Add the ability to set the log level using the `SYNAPSE_TEST_LOG_LEVEL` environment when using `complement.sh`. (#13152)
2022-07-05 03:46:20 -06:00
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
# Render and write the file
Add the ability to set the log level using the `SYNAPSE_TEST_LOG_LEVEL` environment when using `complement.sh`. (#13152)
2022-07-05 03:46:20 -06:00
log_config_filepath = f"/conf/workers/{worker_name}.log.config"
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
convert(
"/conf/log.config",
log_config_filepath,
worker_name=worker_name,
**extra_log_template_args,
Improve startup times in Complement test runs against workers, particularly in CPU-constrained environments. (#13127) Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-06-30 05:58:12 -06:00
include_worker_name_in_log_line=environ.get(
"SYNAPSE_USE_EXPERIMENTAL_FORKING_LAUNCHER"
),
Dockerfile-workers: give the master its own log config (#12466) When we run a worker-mode synapse under docker, everything gets logged to stdout. Currently, output from the workers is tacked with a worker name, for example: ``` 2022-04-13 15:27:56,810 - worker:frontend_proxy1 - synapse.util.caches.lrucache - 154 - INFO - LruCache._expire_old_entries-0 - Dropped 0 items from caches ``` - note `worker:frontend_proxy1`. No such tag is applied to log lines from the master, which makes somewhat confusing reading. To fix this, we generate a dedicated log config file for the master in the same way that we do for the workers, and use that.
2022-04-13 13:50:08 -06:00
)
return log_config_filepath
disallow-untyped-defs in `docker` and `stubs` directories (#12528)
2022-04-25 06:32:35 -06:00
def main(args: List[str], environ: MutableMapping[str, str]) -> None:
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
config_dir = environ.get("SYNAPSE_CONFIG_DIR", "/data")
config_path = environ.get("SYNAPSE_CONFIG_PATH", config_dir + "/homeserver.yaml")
data_dir = environ.get("SYNAPSE_DATA_DIR", "/data")
# override SYNAPSE_NO_TLS, we don't support TLS in worker mode,
# this needs to be handled by a frontend proxy
environ["SYNAPSE_NO_TLS"] = "yes"
# Generate the base homeserver config if one does not yet exist
if not os.path.exists(config_path):
log("Generating base homeserver config")
generate_base_homeserver_config()
Dockerfile-workers: spell out when config isn't generated (#15186) * Complement: Spell out when config isn't generated * Changelog
2023-03-02 08:55:26 -07:00
else:
log("Base homeserver config exists—not regenerating")
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# This script may be run multiple times (mostly by Complement, see note at top of
# file). Don't re-configure workers in this instance.
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
mark_filepath = "/conf/workers_have_been_configured"
if not os.path.exists(mark_filepath):
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
# Collect and validate worker_type requests
# Read the desired worker configuration from the environment
worker_types_env = environ.get("SYNAPSE_WORKER_TYPES", "").strip()
# Only process worker_types if they exist
if not worker_types_env:
# No workers, just the main process
worker_types = []
requested_worker_types: Dict[str, Any] = {}
else:
# Split type names by comma, ignoring whitespace.
worker_types = split_and_strip_string(worker_types_env, ",")
requested_worker_types = parse_worker_types(worker_types)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Always regenerate all other config files
Dockerfile-workers: spell out when config isn't generated (#15186) * Complement: Spell out when config isn't generated * Changelog
2023-03-02 08:55:26 -07:00
log("Generating worker config files")
Additional functionality for declaring worker types in Complement (#14921) Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
2023-03-14 10:29:33 -06:00
generate_worker_files(environ, config_path, data_dir, requested_worker_types)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Mark workers as being configured
with open(mark_filepath, "w") as f:
f.write("")
Dockerfile-workers: spell out when config isn't generated (#15186) * Complement: Spell out when config isn't generated * Changelog
2023-03-02 08:55:26 -07:00
else:
log("Worker config exists—not regenerating")
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
Set LD_PRELOAD to load jemalloc in Dockerfile-workers. (#14182)
2022-10-14 06:29:49 -06:00
# Lifted right out of start.py
jemallocpath = "/usr/lib/%s-linux-gnu/libjemalloc.so.2" % (platform.machine(),)
if os.path.isfile(jemallocpath):
environ["LD_PRELOAD"] = jemallocpath
else:
log("Could not find %s, will not use" % (jemallocpath,))
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
# Start supervisord, which will start Synapse, all of the configured worker
# processes, redis, nginx etc. according to the config we created above.
Dockerfile-workers: reduce the amount we install (#12464) This is an attempt to reduce the rebuild time. In short, we reduce the amount of stuff that the dockerfile installs, so as to give a faster startup.
2022-04-14 08:36:49 -06:00
log("Starting supervisord")
Flush stdout/err in Dockerfile-workers before replacing the current process (#14195) Also update `subprocess.check_output` to the slightly newer `subprocess.run`. Signed-off-by: Jason Little <realtyem@gmail.com>
2022-10-18 05:56:20 -06:00
flush_buffers()
Set LD_PRELOAD to load jemalloc in Dockerfile-workers. (#14182)
2022-10-14 06:29:49 -06:00
os.execle(
Dockerfile-workers: reduce the amount we install (#12464) This is an attempt to reduce the rebuild time. In short, we reduce the amount of stuff that the dockerfile installs, so as to give a faster startup.
2022-04-14 08:36:49 -06:00
"/usr/local/bin/supervisord",
"supervisord",
"-c",
"/etc/supervisor/supervisord.conf",
Set LD_PRELOAD to load jemalloc in Dockerfile-workers. (#14182)
2022-10-14 06:29:49 -06:00
environ,
Dockerfile-workers: reduce the amount we install (#12464) This is an attempt to reduce the rebuild time. In short, we reduce the amount of stuff that the dockerfile installs, so as to give a faster startup.
2022-04-14 08:36:49 -06:00
)
Add a dockerfile for running a set of Synapse worker processes (#9162) This PR adds a Dockerfile and some supporting files to the `docker/` directory. The Dockerfile's intention is to spin up a container with: * A Synapse main process. * Any desired worker processes, defined by a `SYNAPSE_WORKERS` environment variable supplied at runtime. * A redis for worker communication. * A nginx for routing traffic. * A supervisord to start all worker processes and monitor them if any go down. Note that **this is not currently intended to be used in production**. If you'd like to use Synapse workers with Docker, instead make use of the official image, with one worker per container. The purpose of this dockerfile is currently to allow testing Synapse in worker mode with the [Complement](https://github.com/matrix-org/complement/) test suite. `configure_workers_and_start.py` is where most of the magic happens in this PR. It reads from environment variables (documented in the file) and creates all necessary config files for the processes. It is the entrypoint of the Dockerfile, and thus is run any time the docker container is spun up, recreating all config files in case you want to use a different set of workers. One can specify which workers they'd like to use by setting the `SYNAPSE_WORKERS` environment variable (as a comma-separated list of arbitrary worker names) or by setting it to `*` for all worker processes. We will be using the latter in CI. Huge thanks to @MatMaul for helping get this all working :tada: This PR is paired with its equivalent on the Complement side: https://github.com/matrix-org/complement/pull/62. Note, for the purpose of testing this PR before it's merged: You'll need to (re)build the base Synapse docker image for everything to work (`matrixdotorg/synapse:latest`). Then build the worker-based docker image on top (`matrixdotorg/synapse:workers`).
2021-04-14 06:54:49 -06:00
if __name__ == "__main__":
main(sys.argv, os.environ)