-
Notifications
You must be signed in to change notification settings - Fork 1
/
configure_workers_and_start.py
executable file
·2141 lines (1916 loc) · 91.6 KB
/
configure_workers_and_start.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#!/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.
# * 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'
# * SYNAPSE_AS_REGISTRATION_DIR: If specified, a directory in which .yaml and .yml
# files will be treated as Application Service registration files.
# * 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.
# * 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.
# * 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.
#
# 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.
import codecs
import os
import platform
import re
import shutil
import socket
import subprocess
import sys
import urllib.request
from collections import defaultdict
from pathlib import Path
from typing import Any, Dict, List, Mapping, MutableMapping, NoReturn, Optional, Set
import yaml
from jinja2 import Environment, FileSystemLoader
DEBUG = True
if DEBUG is True:
import json
MAIN_PROCESS_HTTP_LISTENER_PORT = 8008
MAIN_PROCESS_HTTP_FED_LISTENER_PORT = 8448
MAIN_PROCESS_NEW_CLIENT_PORT = 8080
MAIN_PROCESS_NEW_FEDERATION_PORT = 8081
MAIN_PROCESS_NEW_HEALTH_PORT = 8082
MAIN_PROCESS_NEW_REPLICATION_PORT = 9093
MAIN_PROCESS_HTTP_METRICS_LISTENER_PORT = 8060
MAIN_PROCESS_NEW_CLIENT_UNIX_SOCKET_PATH = "/run/main_public.sock"
MAIN_PROCESS_NEW_FEDERATION_UNIX_SOCKET_PATH = "/run/main_public.sock"
MAIN_PROCESS_NEW_HEALTH_UNIX_SOCKET_PATH = "/run/main_health.sock"
MAIN_PROCESS_NEW_REPLICATION_UNIX_SOCKET_PATH = "/run/main_replication.sock"
MAIN_PROCESS_NEW_METRICS_UNIX_SOCKET_PATH = "/run/main_metrics.sock"
enable_compressor = False
enable_coturn = False
enable_prometheus = False
enable_redis_exporter = False
enable_postgres_exporter = False
# Set this default to reflect the same as loaded below, so no juggling of types is needed
prometheus_storage_retention_time = "1y"
# 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.
WORKERS_CONFIG: Dict[str, Dict[str, Any]] = {
"pusher": {
"app": "synapse.app.generic_worker",
"listener_resources": [],
"endpoint_patterns": [],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"user_dir": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client"],
"endpoint_patterns": [
"^/_matrix/client/(api/v1|r0|v3|unstable)/user_directory/search$"
],
"shared_extra_conf": {"update_user_directory_from_worker": "placeholder_name"},
"worker_extra_conf": "",
},
"media_repository": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client", "federation", "media"],
"endpoint_patterns": [
"^/_matrix/media/",
"^/_synapse/admin/v1/purge_media_cache$",
"^/_synapse/admin/v1/room/.*/media.*$",
"^/_synapse/admin/v1/user/.*/media.*$",
"^/_synapse/admin/v1/users/.*/media.*$",
"^/_synapse/admin/v1/media/.*$",
"^/_synapse/admin/v1/room/.*/media/quarantine$",
"^/_matrix/client/v1/media/.*$",
"^/_matrix/federation/v1/media/.*$",
],
# The first configured media worker will run the media background jobs
"shared_extra_conf": {
"enable_media_repo": False,
"media_instance_running_background_jobs": "placeholder_name",
},
"worker_extra_conf": "enable_media_repo: true",
},
"appservice": {
"app": "synapse.app.generic_worker",
"listener_resources": [],
"endpoint_patterns": [],
"shared_extra_conf": {"notify_appservices_from_worker": "placeholder_name"},
"worker_extra_conf": "",
},
"federation_sender": {
"app": "synapse.app.generic_worker",
"listener_resources": [],
"endpoint_patterns": [],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"synchrotron": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client"],
"endpoint_patterns": [
"^/_matrix/client/(r0|v3|unstable)/sync$",
"^/_matrix/client/(api/v1|r0|v3)/events$",
"^/_matrix/client/(api/v1|r0|v3)/initialSync$",
"^/_matrix/client/(api/v1|r0|v3)/rooms/[^/]+/initialSync$",
],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"client_reader": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client"],
"endpoint_patterns": [
"^/_matrix/client/(api/v1|r0|v3|unstable)/publicRooms$",
"^/_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$",
"^/_matrix/client/(v1|unstable)/rooms/.*/relations/",
"^/_matrix/client/v1/rooms/.*/threads$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/login$",
"^/_matrix/client/(r0|v3|unstable)/account/3pid$",
"^/_matrix/client/(r0|v3|unstable)/account/whoami$",
"^/_matrix/client/versions$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/voip/turnServer$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/register$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/register/available$",
"^/_matrix/client/(r0|v3|unstable)/auth/.*/fallback/web$",
# This one needs to be routed by the .* cuz that's the room name.
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/messages$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/event",
"^/_matrix/client/(api/v1|r0|v3|unstable)/joined_rooms",
"^/_matrix/client/(api/v1|r0|v3|unstable/.*)/rooms/.*/aliases",
"^/_matrix/client/v1/rooms/.*/timestamp_to_event$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/search",
"^/_matrix/client/(r0|v3|unstable)/password_policy$",
"^/_matrix/client/(api/v1|r0|v3|unstable)/directory/room.*$",
"^/_matrix/client/(r0|v3|unstable)/capabilities$",
],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"federation_reader": {
"app": "synapse.app.generic_worker",
"listener_resources": ["federation"],
"endpoint_patterns": [
"^/_matrix/federation/(v1|v2)/event/",
"^/_matrix/federation/(v1|v2)/state/",
"^/_matrix/federation/(v1|v2)/state_ids/",
"^/_matrix/federation/(v1|v2)/backfill/",
"^/_matrix/federation/(v1|v2)/get_missing_events/",
"^/_matrix/federation/(v1|v2)/publicRooms",
"^/_matrix/federation/(v1|v2)/query/",
"^/_matrix/federation/(v1|v2)/make_join/",
"^/_matrix/federation/(v1|v2)/make_leave/",
"^/_matrix/federation/(v1|v2)/send_join/",
"^/_matrix/federation/(v1|v2)/send_leave/",
"^/_matrix/federation/(v1|v2)/invite/",
"^/_matrix/federation/(v1|v2)/query_auth/",
"^/_matrix/federation/(v1|v2)/event_auth/",
"^/_matrix/federation/v1/timestamp_to_event/",
"^/_matrix/federation/(v1|v2)/exchange_third_party_invite/",
"^/_matrix/federation/(v1|v2)/user/devices/",
"^/_matrix/federation/(v1|v2)/get_groups_publicised$",
"^/_matrix/key/v2/query",
],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"federation_inbound": {
"app": "synapse.app.generic_worker",
"listener_resources": ["federation"],
"endpoint_patterns": ["^/_matrix/federation/(v1|v2)/send/"],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"event_persister": {
"app": "synapse.app.generic_worker",
"listener_resources": ["replication"],
"endpoint_patterns": [],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"background_worker": {
"app": "synapse.app.generic_worker",
"listener_resources": [],
"endpoint_patterns": [],
# This worker cannot be sharded. Therefore, there should only ever be one
# background worker. This is enforced for the safety of your database.
"shared_extra_conf": {"run_background_tasks_on": "placeholder_name"},
"worker_extra_conf": "",
},
"event_creator": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client"],
"endpoint_patterns": [
"^/_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/",
"^/_matrix/client/(api/v1|r0|v3|unstable)/knock/",
"^/_matrix/client/(api/v1|r0|v3|unstable)/profile/",
"^/_matrix/client/(v1|unstable/org.matrix.msc2716)/rooms/.*/batch_send",
],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"frontend_proxy": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client"],
"endpoint_patterns": ["^/_matrix/client/(api/v1|r0|v3|unstable)/keys/upload"],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"account_data": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client", "replication"],
"endpoint_patterns": [
"^/_matrix/client/(r0|v3|unstable)/.*/tags",
"^/_matrix/client/(r0|v3|unstable)/.*/account_data",
],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"presence": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client", "replication"],
"endpoint_patterns": ["^/_matrix/client/(api/v1|r0|v3|unstable)/presence/"],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"receipts": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client", "replication"],
"endpoint_patterns": [
"^/_matrix/client/(r0|v3|unstable)/rooms/.*/receipt",
"^/_matrix/client/(r0|v3|unstable)/rooms/.*/read_markers",
],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"to_device": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client", "replication"],
"endpoint_patterns": ["^/_matrix/client/(r0|v3|unstable)/sendToDevice/"],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
"typing": {
"app": "synapse.app.generic_worker",
"listener_resources": ["client", "replication"],
"endpoint_patterns": [
"^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/typing"
],
"shared_extra_conf": {},
"worker_extra_conf": "",
},
}
# Templates for sections that may be inserted multiple times in config files
NGINX_LOCATION_CONFIG_BLOCK = """
location ~* {endpoint} {{
proxy_pass {upstream};
proxy_set_header X-Forwarded-For $remote_addr;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header Host $host;
{additional_location_body}
}}
"""
# Bump zone shared memory upto 512k from 64k
NGINX_UPSTREAM_CONFIG_BLOCK = """
upstream {upstream_name} {{
zone upstreams 512K;
{body}
}}
"""
PROMETHEUS_SCRAPE_CONFIG_BLOCK = """
- targets: ["{metrics_target}"]
labels:
instance: "Synapse"
job: "{name}"
index: {index}
"""
class Worker:
"""
This is a base class representing a requested worker.
Attributes:
base_name: The basic name serves as a group identifier
name: The given name the worker should use. base_name + incrementing number
index: The number this worker was given on the end of base_name to make the name
app: 'synapse.app.generic_worker' for all now
listener_resources: Set of types of listeners needed. 'client, federation,
replication, media' etc.
main_port: The main port number for most primary listeners('client', 'federation', etc)
health_port: The port number for the /health endpoint
manhole_port: The port number for ssh-ing into a running synapse. Will not be a Unix socket
metrics_port: The port number for the metrics endpoints
replication_port: The port number for replication, if needed
endpoint_patterns: Set of url endpoints this worker watches. e.g. {'/url1','/url2'}
shared_extra_config: Dict of one-offs that enable special roles for specific
workers. Ends up in shared.yaml
worker_extra_conf: Only used by media_repository to enable that functionality.
Ends up in worker.yaml
types_list: the List of roles this worker fulfills.
"""
base_name: str
name: str
index: int
app: str
listener_resources: Set[str]
main_port: int
health_port: int
manhole_port: int
metrics_port: int
replication_port: int
endpoint_patterns: Set[str]
shared_extra_config: Dict[str, Any]
worker_extra_conf: str
types_list: List[str]
def extract_jinja_worker_template(self) -> Dict[str, Any]:
config: Dict[str, Any] = {}
config.setdefault("app", self.app)
config.setdefault("name", self.name)
config.setdefault("base_name", self.base_name)
config.setdefault("index", self.index)
config.setdefault("listener_resources", self.listener_resources)
config.setdefault("worker_extra_conf", self.worker_extra_conf)
return config
def __init__(self, name: str, worker_type_str: str) -> None:
"""
Initialize the parameters of this new worker by parsing the piece of string
stipulating the type of worker needed.
Args:
name: The name requested of the worker, will probably be updated at least
once.
worker_type_str: The string representing what roles this worker will
fulfill.
"""
self.listener_resources = set()
self.endpoint_patterns = set()
self.shared_extra_config = {}
self.main_port = 0
self.health_port = 0
self.manhole_port = 0
self.metrics_port = 0
self.replication_port = 0
self.types_list = []
self.worker_extra_conf = ""
self.base_name = ""
self.name = ""
self.index = 0
self.app = ""
# Assign the name provided
self.base_name = name
# Split the worker types from string into list. This will have already been
# stripped of the potential name and multiplier. Check for duplicates in the
# split worker type list. No advantage in having duplicated worker types on
# the same worker. Two would consolidate into one. (e.g. "pusher + pusher"
# would resolve to a single "pusher" which may not be what was intended.)
self.types_list = split_and_strip_string(worker_type_str, "+")
if len(self.types_list) != len(set(self.types_list)):
error(f"Duplicate worker type found in '{worker_type_str}'! Please fix.")
for role in self.types_list:
worker_config = WORKERS_CONFIG.get(role)
if worker_config:
worker_config = worker_config.copy()
else:
error(
f"'{role}' is an unknown worker type! Was found in "
f"'{worker_type_str}'. Please fix!"
)
# Get the app(all should now be synapse.app.generic_worker).
# TODO: factor this out
self.app = str(worker_config.get("app"))
# Get the listener_resources
listener_resources = worker_config.get("listener_resources", [])
if listener_resources:
self.listener_resources.update(listener_resources)
endpoint_patterns = worker_config.get("endpoint_patterns", [])
self.endpoint_patterns.update(endpoint_patterns)
# Get shared_extra_conf, if any
shared_extra_config = worker_config.get("shared_extra_conf")
if shared_extra_config:
self.shared_extra_config.update(shared_extra_config)
# Get worker_extra_conf, if any(there is only one at this time,
# and that is the media_repository. This goes in the worker yaml, so it's
# pretty safe, can't use 2 media_repo workers on the same worker.
worker_extra_conf = worker_config.get("worker_extra_conf")
if worker_extra_conf:
# Copy it, in case it takes it as a reference.
self.worker_extra_conf = worker_extra_conf
class Workers:
"""
This is a grouping of Worker, containing all the workers requested.
Attributes:
total_count: int of how many workers there are total.
worker: the collection of workers to access data about.
worker_type_counter: The broad counter that is used to increment the worker name
worker_type_to_name_map: Used to check for cross naming violations.
worker_type_fine_grain_counter: A finer grained counter for total number of a
specific type of worker.
current_port_counter: A simple integer counter to track the next port number to
assign.
map_of_worker_to_upstream: dict of 'worker_name':'upstream_host'.
e.g. 'event_creator1':'event_creator.client'
"""
total_count: int
worker: Dict[str, Worker]
worker_type_counter: Dict[str, int]
worker_type_to_name_map: Dict[str, str]
worker_type_fine_grain_counter: Dict[str, int]
current_port_counter: int
map_of_worker_to_upstream: Dict[str, str]
SHARDING_NOT_ALLOWED_LIST: List[str] = [
"background_worker",
"account_data",
"presence",
"receipts",
"typing",
"to_device",
]
def __init__(self, port_starting_num: int) -> None:
self.map_of_worker_to_upstream = {}
self.total_count = 0
self.worker = {}
self.worker_type_counter = defaultdict(int)
self.worker_type_to_name_map = {}
self.worker_type_fine_grain_counter = defaultdict(int)
self.current_port_counter = port_starting_num
def add_worker(self, requested_worker_type: str) -> str:
"""
Make a worker, check its name is sane, and collect its configuration bits for
later.
Args:
requested_worker_type: the string combination of roles this worker needs to
fulfill(and possibly a requested name).
Returns: string of the new worker's full name
"""
# Peel off any name designated before a '=' to use later.
worker_base_name = ""
if "=" in requested_worker_type:
# Split on "=", remove extra whitespace from ends then make list
worker_type_split = split_and_strip_string(requested_worker_type, "=")
if len(worker_type_split) > 2:
error(
"To many worker names requested for a single worker, or to many "
f"'='. Please fix: '{requested_worker_type}'"
)
# 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}"
)
# Reassign the worker_type string with no name on it.
requested_worker_type = worker_type_split[1]
# If there was a requested name, it's been removed now
worker_type_set: Set[str] = set(
split_and_strip_string(requested_worker_type, "+")
)
if worker_base_name:
# It'll be useful to have this in the log in case it's a complex of many
# workers merged together. Note for Complement: it would only be seen in the
# logs for blueprint construction(which are not collected).
log(
f"Worker name request found: '{worker_base_name}', for: "
f"'{requested_worker_type}'"
)
else:
# The worker name will be the worker_type as a sorted and joined string
worker_base_name = "+".join(sorted(worker_type_set))
# Check worker base name isn't in use by something else already. Will error and
# stop if it is.
name_to_check = self.worker_type_to_name_map.get(worker_base_name)
if (name_to_check is None) or (name_to_check == requested_worker_type):
new_worker = Worker(worker_base_name, requested_worker_type)
# Check if there is to many of a worker there can only be one of.
# Will error and stop if it is a problem, e.g. 'background_worker'.
for role in new_worker.types_list:
if role in self.worker_type_fine_grain_counter:
# It's in the counter, even once. Check for sharding.
if role in self.SHARDING_NOT_ALLOWED_LIST:
error(
f"There can be only a single worker with '{role}' type. "
"Please check and remove."
)
# Either it's not in counter or it is but is not a shard hazard,
# therefore it's safe to add. Don't need the return value here.
self.worker_type_fine_grain_counter[role] += 1
# Add to the name:type map, if it already exists this will no-op and that's
# what we want.
self.worker_type_to_name_map.setdefault(
worker_base_name, requested_worker_type
)
# Now add or increment it on the global counter
self.worker_type_counter[requested_worker_type] += 1
# Save the count as an index
new_worker.index = self.worker_type_counter[requested_worker_type]
# Name workers by their type or requested name concatenated with an
# incrementing number. e.g. event_creator+event_persister1,
# federation_reader1 or bob1
new_worker.name = new_worker.base_name + str(new_worker.index)
# Save it to the kettle for later cooking
self.worker.setdefault(new_worker.name, new_worker)
# Give back the new worker name that's been settled on as a copy of the
# string, more work to do.
return str(new_worker.name)
else:
error(
f"Can not use '{worker_base_name}' with requested worker_type: "
f"'{requested_worker_type}', it is in use by: "
f"'{name_to_check}'"
)
def update_local_shared_config(self, worker_name: str) -> None:
"""Insert a given worker name into the worker's configuration dict.
Args:
worker_name: The name of the worker to insert.
"""
dict_to_edit = self.worker[worker_name].shared_extra_config.copy()
for k, v in dict_to_edit.items():
# Only proceed if it's a string as some values are boolean
if isinstance(dict_to_edit[k], str):
# This will be ignored if the text isn't the placeholder.
dict_to_edit[k] = v.replace("placeholder_name", worker_name)
if isinstance(dict_to_edit[k], list):
# I think this is where we can add stream writers and other special list
pass
self.worker[worker_name].shared_extra_config = dict_to_edit
def get_next_port_number(self) -> int:
"""
Increment and return the port number counter. This will initially create a gap
at the beginning at the sequence as one number gets skipped over. Ignore it as
annoying but not world-breaking
Returns: int of the next port number to assign
"""
self.current_port_counter += 1
return self.current_port_counter
class NginxConfig:
"""
This represents collected data to plug into the various nginx configuration points.
Attributes:
locations: A dict of locations and the name of the upstream_host responsible for
it. This will be used in the final nginx file. e.g.
'/_matrix/federation/.*/send':'bob-federation_inbound.federation'
upstreams_to_ports: A map of upstream host to a Set of ports. This is used in
the final nginx file.
upstreams_roles: Set of worker_type roles. Used to calculate load-balancing.
"""
locations: Dict[str, str]
upstreams_to_ports: Dict[str, Set[int]]
upstreams_roles: Dict[str, Set[str]]
def __init__(self) -> None:
self.locations = {}
self.upstreams_to_ports = {}
self.upstreams_roles = {}
def add_upstreams(
self,
host: str,
worker_roles: Set[str],
ports: Set[int],
) -> None:
"""
Add a new upstream host to NginxConfig.upstreams_to_ports and
NginxConfig.upstreams_roles when given the worker roles as a set and a
set of ports to assign to it.
Args:
host: The name to use for the load-balancing upstream nginx block.
worker_roles: A Set of roles this upstream block can be responsible for.
ports: A Set containing a single port or multiple ports to assign.
Returns: None
"""
debug(
f"Making new upstream: {host}\n - roles: {worker_roles}\n - ports: {ports}"
)
# Initialize this (possible new) host entries
self.upstreams_to_ports.setdefault(host, set()).update(ports)
self.upstreams_roles.setdefault(host, set()).update(worker_roles)
def create_upstreams_and_locations(self, workers: Workers):
"""
Use to create upstream data and endpoint locations for the written out Nginx
config file. Check class attributes for locations and upstreams_to_ports
"""
# Primarily used to create the necessary data for an upstream block to be
# written into the nginx config file. For such, the upstream needs a name and
# port(s). Additionally, some endpoints are more efficiently used if hashed
# to consistently point to the same worker(or port, in this context). Finally,
# will create the nginx.locations map used for the final nginx config file that
# associates the correct upstream name to a given endpoint.
# The upstream name created will be in the format of "name.listener_type" and
# will not have "http://" added to it.
# Note that this does create upstreams even with a single port. Nginx says
# that doing so, and adding in correct 'zone' and 'keepalive' declarations,
# allows for better performance because of shared_mem usage for state
# information on upstreams and by recycling open connections instead of
# constantly creating and closing them. The connection pool is naturally
# quite large, but a given closed connection holds the file descriptors for 2
# minutes before adding it back to the pool.
# Uses data from
# locations_to_port_set to find out what ports can serve a given endpoint
# port_to_upstream_name to get the hostname used by this port, typically a
# worker's base name is used for convenient grouping. Multiple
# non-matching hostnames will be concatenated with a '-' when creating
# this upstream name.
# port_to_listener_type to get the information on what kind of listener
# type(like 'client' or 'federation') this port is associated with.
# Functionally, this allows a given upstream to separate client requests
# from federation requests(and etc) since each listener type has it's own
# port. Also has a visual use for a person inspecting the config file.
# upstreams_roles to add a worker's roles for use in the specializing
# load-balancing by client IP or other hashing method.
# locations_to_port_set: A dict of locations and a set of collected ports.
# e.g. {'/endpoint1': {1234, 1235, ...}}
locations_to_port_set: Dict[str, Set[int]] = {}
# port_to_upstream_name: A map of port to the worker's base name related to it.
# e.g. {1234: "worker_base_name"}
port_to_upstream_name: Dict[int, str] = {}
# Add nginx location blocks for this worker's endpoints (if any are defined)
# Inappropriate types of listeners were already filtered out.
for worker in workers.worker.values():
for pattern in worker.endpoint_patterns:
# Collect port numbers for this endpoint pattern
locations_to_port_set.setdefault(pattern, set()).add(worker.main_port)
# Set lookup maps to be used for combining upstreams in a moment.
# Need the worker's base name
port_to_upstream_name.setdefault(worker.main_port, worker.base_name)
# And the list of roles this(possibly combination) worker can fill
self.upstreams_roles.setdefault(worker.base_name, set()).update(
worker.types_list
)
for endpoint_pattern, port_set in locations_to_port_set.items():
# Reset these for each run
new_nginx_upstream_set: Set[str] = set()
new_nginx_upstream_listener_set: Set[str] = set()
new_nginx_upstream_roles: Set[str] = set()
# Check each port number. For a single-port upstream this only iterates once
for each_port in port_set:
# Get the worker.base_name for the upstream name
new_nginx_upstream_set.add(port_to_upstream_name[each_port])
# Get the workers roles for specialized load-balancing
new_nginx_upstream_roles.update(
self.upstreams_roles[port_to_upstream_name[each_port]]
)
# This will be the name of the upstream
new_nginx_upstream = f"{'-'.join(sorted(new_nginx_upstream_set))}"
# Check this upstream exists, if not then make it
if new_nginx_upstream not in self.upstreams_to_ports:
self.add_upstreams(
new_nginx_upstream, new_nginx_upstream_roles, port_set
)
# Finally, update nginx.locations with new upstream
self.locations[endpoint_pattern] = new_nginx_upstream
# Utility functions
def log(txt: str) -> None:
print(txt, flush=True)
def error(txt: str) -> NoReturn:
print(txt, file=sys.stderr, flush=True)
sys.exit(2)
def debug(txt: str) -> None:
if DEBUG:
print(txt, flush=True)
def flush_buffers() -> None:
sys.stdout.flush()
sys.stderr.flush()
def convert(src: str, dst: str, mode: str = "a", **template_vars: object) -> None:
"""Generate a file from a template
Args:
src: Path to the input file.
dst: Path to write to.
mode: "a" to append a file, "w" to wipe and write a file. See open docstring.
template_vars: The arguments to replace placeholder variables in the template
with.
"""
# Read the template file
# 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))
# Generate a string from the template.
rendered = template.render(**template_vars)
# 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, mode) as outfile:
# In case the existing file doesn't end with a newline
outfile.write("\n")
outfile.write(rendered)
def getenv_bool(name: str, default: bool = False) -> bool:
return os.getenv(name, str(default)).lower() in ("yes", "y", "true", "1", "t", "on")
def add_worker_roles_to_shared_config(
shared_config: dict,
worker_type_list: list,
worker_name: str,
worker_port: int,
use_unix_socket: bool = False,
) -> None:
"""Given a dictionary representing a config file shared across all workers,
append appropriate worker information to it for the current worker_type instance.
Args:
shared_config: The config dict that all worker instances share (after being
converted to YAML)
worker_type_list: The type of worker (one of those defined in WORKERS_CONFIG).
This list can be a single worker type or multiple.
worker_name: The name of the worker instance.
worker_port: The replication port that the worker instance is listening on.
use_unix_socket: If a socket path should be used instead of a host/port combo
"""
# The instance_map config field marks the workers that write to various replication
# streams
instance_map = shared_config.setdefault("instance_map", {})
# This is a list of the stream_writers that there can be only one of. Events can be
# sharded, and therefore doesn't belong here.
singular_stream_writers = [
"account_data",
"presence",
"receipts",
"to_device",
"typing",
]
# Worker-type specific sharding config. Now a single worker can fulfill multiple
# roles, check each.
if "pusher" in worker_type_list:
shared_config.setdefault("pusher_instances", []).append(worker_name)
if "federation_sender" in worker_type_list:
shared_config.setdefault("federation_sender_instances", []).append(worker_name)
if "event_persister" in worker_type_list:
# Event persisters write to the events stream, so we need to update
# the list of event stream writers
shared_config.setdefault("stream_writers", {}).setdefault("events", []).append(
worker_name
)
# This 'for' loop serves two purposes:
# 1. Update the list of stream writers. It's convenient that the name of the worker
# type is the same as the stream to write. Iterate over the whole list in case there
# is more than one.
# 2. Any worker type that has a 'replication' listener gets add to the
# 'instance_map'.
for worker in worker_type_list:
if worker in singular_stream_writers:
shared_config.setdefault("stream_writers", {}).setdefault(
worker, []
).append(worker_name)
if "replication" in WORKERS_CONFIG.get(worker, {}).get(
"listener_resources", []
):
# Map of worker instance names to path or host/ports combos. If a worker
# type in WORKERS_CONFIG needs to be added here in the future, just add a
# 'replication' entry to the list in listener_resources for that worker.
# NOTE: in theory, the worker_port could be a 0 but that should not be the case
# as if a 'replication' listener is defined then this will be populated by the
# counter system
assert worker_port > 0
if use_unix_socket:
instance_map[worker_name] = {
"path": f"/run/worker.{worker_port}",
}
else:
instance_map[worker_name] = {
"host": "localhost",
"port": worker_port,
}
def combine_shared_config_fragments(
shared_config: Dict[str, Any], entries_to_add: Dict[str, Any]
) -> Dict[str, Any]:
# This takes the new dict and copies the old one over top of it, so that it
# overwrites any duplicate values with the pre-existing ones.
new_shared_config = entries_to_add.copy()
new_shared_config.update(shared_config)
return new_shared_config
def split_and_strip_string(given_string: str, split_char: str) -> List:
# Removes whitespace from ends of result strings before adding to list.
return [x.strip() for x in given_string.split(split_char)]
def extract_port_number_from_original_listeners(
original_listeners_list: List[Any], listener_to_find: str
) -> int:
"""
Iterate and return the port number for a given listener from the original
configuration files.
Args:
original_listeners_list: The list after extraction from the existing config
listener_to_find: The name of the listener to track down. e.g. 'client'
Returns: int of the port number found, 0 if not found
"""
# The original_listeners_list is a list with each entry being a dict containing, at
# least, a 'type' and a 'port' and another list(resources) with potentially more
# than one dict again containing, at least, one or more 'names' as a list.
port_to_return = 0
for list_entry in original_listeners_list:
if "http" in list_entry["type"]:
for resource in list_entry["resources"]:
if listener_to_find in resource["names"]:
if "port" in list_entry:
debug(f"{listener_to_find} port found: {list_entry['port']}")
port_to_return = list_entry["port"]
if listener_to_find in list_entry["type"]:
debug(f"{listener_to_find} port found: {list_entry['port']}")
port_to_return = list_entry["port"]
return port_to_return
def extract_socket_path_from_original_listeners(
original_listeners_list: List[Any], listener_to_find: str
) -> Optional[str]:
"""
Iterate and return the socket path for a given listener from the original
configuration files.
Args:
original_listeners_list: The list after extraction from the existing config
listener_to_find: The name of the listener to track down. e.g. 'client'
Returns: str of the path found, None if not found
"""
# The original_listeners_list is a list with each entry being a dict containing, at
# least, a 'type' and a 'port' and another list(resources) with potentially more
# than one dict again containing, at least, one or more 'names' as a list.
path_to_return = None
for list_entry in original_listeners_list:
if "http" in list_entry["type"]:
for resource in list_entry["resources"]:
if listener_to_find in resource["names"]:
if "path" in list_entry:
debug(f"{listener_to_find} path found: {list_entry['path']}")
path_to_return = list_entry["path"]
return path_to_return
def construct_worker_listener_block(
port_or_path_number: int,
listener_type_as_list: List[str],
use_socket: bool = False,
use_compress: bool = False,
) -> Dict[str, str | List[Dict[str, List]]]:
"""
Construct a JSON block for a worker listener, using either a path or a port. A port
assumes that 0.0.0.0 is the host to bind to. This is a much simplified version of a
listener compared to a regular listener.
Args:
port_or_path_number: This number is used either way, as an obvious port or to
individualize the worker socket file.
listener_type_as_list: The types of listener, e.g. 'client', 'federation' or
'replication'
use_socket: If True, set up as a Unix socket
use_compress: If True, enable compression for the listener
Return: A dict structure that translates into a JSON block and ends up being YAML
"""
port_or_path_key = "path" if use_socket else "port"
port_or_path_value = (
f"/run/worker.{port_or_path_number}" if use_socket else port_or_path_number