federation_sender.py 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325
  1. #!/usr/bin/env python
  2. # -*- coding: utf-8 -*-
  3. # Copyright 2016 OpenMarket Ltd
  4. #
  5. # Licensed under the Apache License, Version 2.0 (the "License");
  6. # you may not use this file except in compliance with the License.
  7. # You may obtain a copy of the License at
  8. #
  9. # http://www.apache.org/licenses/LICENSE-2.0
  10. #
  11. # Unless required by applicable law or agreed to in writing, software
  12. # distributed under the License is distributed on an "AS IS" BASIS,
  13. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  14. # See the License for the specific language governing permissions and
  15. # limitations under the License.
  16. import synapse
  17. from synapse.server import HomeServer
  18. from synapse.config._base import ConfigError
  19. from synapse.config.logger import setup_logging
  20. from synapse.config.homeserver import HomeServerConfig
  21. from synapse.crypto import context_factory
  22. from synapse.http.site import SynapseSite
  23. from synapse.federation import send_queue
  24. from synapse.federation.units import Edu
  25. from synapse.metrics.resource import MetricsResource, METRICS_PREFIX
  26. from synapse.replication.slave.storage.deviceinbox import SlavedDeviceInboxStore
  27. from synapse.replication.slave.storage.events import SlavedEventStore
  28. from synapse.replication.slave.storage.receipts import SlavedReceiptsStore
  29. from synapse.replication.slave.storage.registration import SlavedRegistrationStore
  30. from synapse.replication.slave.storage.transactions import TransactionStore
  31. from synapse.storage.engines import create_engine
  32. from synapse.storage.presence import UserPresenceState
  33. from synapse.util.async import sleep
  34. from synapse.util.httpresourcetree import create_resource_tree
  35. from synapse.util.logcontext import LoggingContext
  36. from synapse.util.manhole import manhole
  37. from synapse.util.rlimit import change_resource_limit
  38. from synapse.util.versionstring import get_version_string
  39. from synapse import events
  40. from twisted.internet import reactor, defer
  41. from twisted.web.resource import Resource
  42. from daemonize import Daemonize
  43. import sys
  44. import logging
  45. import gc
  46. import ujson as json
  47. logger = logging.getLogger("synapse.app.appservice")
  48. class FederationSenderSlaveStore(
  49. SlavedDeviceInboxStore, TransactionStore, SlavedReceiptsStore, SlavedEventStore,
  50. SlavedRegistrationStore,
  51. ):
  52. pass
  53. class FederationSenderServer(HomeServer):
  54. def get_db_conn(self, run_new_connection=True):
  55. # Any param beginning with cp_ is a parameter for adbapi, and should
  56. # not be passed to the database engine.
  57. db_params = {
  58. k: v for k, v in self.db_config.get("args", {}).items()
  59. if not k.startswith("cp_")
  60. }
  61. db_conn = self.database_engine.module.connect(**db_params)
  62. if run_new_connection:
  63. self.database_engine.on_new_connection(db_conn)
  64. return db_conn
  65. def setup(self):
  66. logger.info("Setting up.")
  67. self.datastore = FederationSenderSlaveStore(self.get_db_conn(), self)
  68. logger.info("Finished setting up.")
  69. def _listen_http(self, listener_config):
  70. port = listener_config["port"]
  71. bind_address = listener_config.get("bind_address", "")
  72. site_tag = listener_config.get("tag", port)
  73. resources = {}
  74. for res in listener_config["resources"]:
  75. for name in res["names"]:
  76. if name == "metrics":
  77. resources[METRICS_PREFIX] = MetricsResource(self)
  78. root_resource = create_resource_tree(resources, Resource())
  79. reactor.listenTCP(
  80. port,
  81. SynapseSite(
  82. "synapse.access.http.%s" % (site_tag,),
  83. site_tag,
  84. listener_config,
  85. root_resource,
  86. ),
  87. interface=bind_address
  88. )
  89. logger.info("Synapse federation_sender now listening on port %d", port)
  90. def start_listening(self, listeners):
  91. for listener in listeners:
  92. if listener["type"] == "http":
  93. self._listen_http(listener)
  94. elif listener["type"] == "manhole":
  95. reactor.listenTCP(
  96. listener["port"],
  97. manhole(
  98. username="matrix",
  99. password="rabbithole",
  100. globals={"hs": self},
  101. ),
  102. interface=listener.get("bind_address", '127.0.0.1')
  103. )
  104. else:
  105. logger.warn("Unrecognized listener type: %s", listener["type"])
  106. @defer.inlineCallbacks
  107. def replicate(self):
  108. http_client = self.get_simple_http_client()
  109. store = self.get_datastore()
  110. replication_url = self.config.worker_replication_url
  111. send_handler = FederationSenderHandler(self)
  112. send_handler.on_start()
  113. while True:
  114. try:
  115. args = store.stream_positions()
  116. args.update((yield send_handler.stream_positions()))
  117. args["timeout"] = 30000
  118. result = yield http_client.get_json(replication_url, args=args)
  119. yield store.process_replication(result)
  120. yield send_handler.process_replication(result)
  121. except:
  122. logger.exception("Error replicating from %r", replication_url)
  123. yield sleep(30)
  124. def start(config_options):
  125. try:
  126. config = HomeServerConfig.load_config(
  127. "Synapse federation sender", config_options
  128. )
  129. except ConfigError as e:
  130. sys.stderr.write("\n" + e.message + "\n")
  131. sys.exit(1)
  132. assert config.worker_app == "synapse.app.federation_sender"
  133. setup_logging(config.worker_log_config, config.worker_log_file)
  134. events.USE_FROZEN_DICTS = config.use_frozen_dicts
  135. database_engine = create_engine(config.database_config)
  136. if config.send_federation:
  137. sys.stderr.write(
  138. "\nThe send_federation must be disabled in the main synapse process"
  139. "\nbefore they can be run in a separate worker."
  140. "\nPlease add ``send_federation: false`` to the main config"
  141. "\n"
  142. )
  143. sys.exit(1)
  144. # Force the pushers to start since they will be disabled in the main config
  145. config.send_federation = True
  146. tls_server_context_factory = context_factory.ServerContextFactory(config)
  147. ps = FederationSenderServer(
  148. config.server_name,
  149. db_config=config.database_config,
  150. tls_server_context_factory=tls_server_context_factory,
  151. config=config,
  152. version_string="Synapse/" + get_version_string(synapse),
  153. database_engine=database_engine,
  154. )
  155. ps.setup()
  156. ps.start_listening(config.worker_listeners)
  157. def run():
  158. with LoggingContext("run"):
  159. logger.info("Running")
  160. change_resource_limit(config.soft_file_limit)
  161. if config.gc_thresholds:
  162. gc.set_threshold(*config.gc_thresholds)
  163. reactor.run()
  164. def start():
  165. ps.replicate()
  166. ps.get_datastore().start_profiling()
  167. ps.get_state_handler().start_caching()
  168. reactor.callWhenRunning(start)
  169. if config.worker_daemonize:
  170. daemon = Daemonize(
  171. app="synapse-federation-sender",
  172. pid=config.worker_pid_file,
  173. action=run,
  174. auto_close_fds=False,
  175. verbose=True,
  176. logger=logger,
  177. )
  178. daemon.start()
  179. else:
  180. run()
  181. class FederationSenderHandler(object):
  182. """Processes the replication stream and forwards the appropriate entries
  183. to the federation sender.
  184. """
  185. def __init__(self, hs):
  186. self.store = hs.get_datastore()
  187. self.federation_sender = hs.get_federation_sender()
  188. self._room_serials = {}
  189. self._room_typing = {}
  190. def on_start(self):
  191. # There may be some events that are persisted but haven't been sent,
  192. # so send them now.
  193. self.federation_sender.notify_new_events(
  194. self.store.get_room_max_stream_ordering()
  195. )
  196. @defer.inlineCallbacks
  197. def stream_positions(self):
  198. stream_id = yield self.store.get_federation_out_pos("federation")
  199. defer.returnValue({"federation": stream_id})
  200. @defer.inlineCallbacks
  201. def process_replication(self, result):
  202. # The federation stream contains things that we want to send out, e.g.
  203. # presence, typing, etc.
  204. fed_stream = result.get("federation")
  205. if fed_stream:
  206. latest_id = int(fed_stream["position"])
  207. # The federation stream containis a bunch of different types of
  208. # rows that need to be handled differently. We parse the rows, put
  209. # them into the appropriate collection and then send them off.
  210. presence_to_send = {}
  211. keyed_edus = {}
  212. edus = {}
  213. failures = {}
  214. device_destinations = set()
  215. # Parse the rows in the stream
  216. for row in fed_stream["rows"]:
  217. position, typ, content_js = row
  218. content = json.loads(content_js)
  219. if typ == send_queue.PRESENCE_TYPE:
  220. destination = content["destination"]
  221. state = UserPresenceState.from_dict(content["state"])
  222. presence_to_send.setdefault(destination, []).append(state)
  223. elif typ == send_queue.KEYED_EDU_TYPE:
  224. key = content["key"]
  225. edu = Edu(**content["edu"])
  226. keyed_edus.setdefault(
  227. edu.destination, {}
  228. )[(edu.destination, tuple(key))] = edu
  229. elif typ == send_queue.EDU_TYPE:
  230. edu = Edu(**content)
  231. edus.setdefault(edu.destination, []).append(edu)
  232. elif typ == send_queue.FAILURE_TYPE:
  233. destination = content["destination"]
  234. failure = content["failure"]
  235. failures.setdefault(destination, []).append(failure)
  236. elif typ == send_queue.DEVICE_MESSAGE_TYPE:
  237. device_destinations.add(content["destination"])
  238. else:
  239. raise Exception("Unrecognised federation type: %r", typ)
  240. # We've finished collecting, send everything off
  241. for destination, states in presence_to_send.items():
  242. self.federation_sender.send_presence(destination, states)
  243. for destination, edu_map in keyed_edus.items():
  244. for key, edu in edu_map.items():
  245. self.federation_sender.send_edu(
  246. edu.destination, edu.edu_type, edu.content, key=key,
  247. )
  248. for destination, edu_list in edus.items():
  249. for edu in edu_list:
  250. self.federation_sender.send_edu(
  251. edu.destination, edu.edu_type, edu.content, key=None,
  252. )
  253. for destination, failure_list in failures.items():
  254. for failure in failure_list:
  255. self.federation_sender.send_failure(destination, failure)
  256. for destination in device_destinations:
  257. self.federation_sender.send_device_messages(destination)
  258. # Record where we are in the stream.
  259. yield self.store.update_federation_out_pos(
  260. "federation", latest_id
  261. )
  262. # We also need to poke the federation sender when new events happen
  263. event_stream = result.get("events")
  264. if event_stream:
  265. latest_pos = event_stream["position"]
  266. self.federation_sender.notify_new_events(latest_pos)
  267. if __name__ == '__main__':
  268. with LoggingContext("main"):
  269. start(sys.argv[1:])