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
|
#
# This file is licensed under the Affero General Public License (AGPL) version 3.
#
# Copyright 2020 The Matrix.org Foundation C.I.C.
# Copyright (C) 2023 New Vector, Ltd
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as
# published by the Free Software Foundation, either version 3 of the
# License, or (at your option) any later version.
#
# See the GNU Affero General Public License for more details:
# <https://www.gnu.org/licenses/agpl-3.0.html>.
#
# Originally licensed under the Apache License, Version 2.0:
# <http://www.apache.org/licenses/LICENSE-2.0>.
#
# [This file includes modifications made by New Vector Limited]
#
#
from typing import List, Optional
from twisted.test.proto_helpers import MemoryReactor
from synapse.server import HomeServer
from synapse.storage.database import (
DatabasePool,
LoggingDatabaseConnection,
LoggingTransaction,
)
from synapse.storage.engines import IncorrectDatabaseSetup
from synapse.storage.types import Cursor
from synapse.storage.util.id_generators import MultiWriterIdGenerator, StreamIdGenerator
from synapse.util import Clock
from tests.unittest import HomeserverTestCase
from tests.utils import USE_POSTGRES_FOR_TESTS
class StreamIdGeneratorTestCase(HomeserverTestCase):
def prepare(self, reactor: MemoryReactor, clock: Clock, hs: HomeServer) -> None:
self.store = hs.get_datastores().main
self.db_pool: DatabasePool = self.store.db_pool
self.get_success(self.db_pool.runInteraction("_setup_db", self._setup_db))
def _setup_db(self, txn: LoggingTransaction) -> None:
txn.execute(
"""
CREATE TABLE foobar (
stream_id BIGINT NOT NULL,
data TEXT
);
"""
)
txn.execute("INSERT INTO foobar VALUES (123, 'hello world');")
def _create_id_generator(self) -> StreamIdGenerator:
def _create(conn: LoggingDatabaseConnection) -> StreamIdGenerator:
return StreamIdGenerator(
db_conn=conn,
notifier=self.hs.get_replication_notifier(),
table="foobar",
column="stream_id",
)
return self.get_success_or_raise(self.db_pool.runWithConnection(_create))
def test_initial_value(self) -> None:
"""Check that we read the current token from the DB."""
id_gen = self._create_id_generator()
self.assertEqual(id_gen.get_current_token(), 123)
def test_single_gen_next(self) -> None:
"""Check that we correctly increment the current token from the DB."""
id_gen = self._create_id_generator()
async def test_gen_next() -> None:
async with id_gen.get_next() as next_id:
# We haven't persisted `next_id` yet; current token is still 123
self.assertEqual(id_gen.get_current_token(), 123)
# But we did learn what the next value is
self.assertEqual(next_id, 124)
# Once the context manager closes we assume that the `next_id` has been
# written to the DB.
self.assertEqual(id_gen.get_current_token(), 124)
self.get_success(test_gen_next())
def test_multiple_gen_nexts(self) -> None:
"""Check that we handle overlapping calls to gen_next sensibly."""
id_gen = self._create_id_generator()
async def test_gen_next() -> None:
ctx1 = id_gen.get_next()
ctx2 = id_gen.get_next()
ctx3 = id_gen.get_next()
# Request three new stream IDs.
self.assertEqual(await ctx1.__aenter__(), 124)
self.assertEqual(await ctx2.__aenter__(), 125)
self.assertEqual(await ctx3.__aenter__(), 126)
# None are persisted: current token unchanged.
self.assertEqual(id_gen.get_current_token(), 123)
# Persist each in turn.
await ctx1.__aexit__(None, None, None)
self.assertEqual(id_gen.get_current_token(), 124)
await ctx2.__aexit__(None, None, None)
self.assertEqual(id_gen.get_current_token(), 125)
await ctx3.__aexit__(None, None, None)
self.assertEqual(id_gen.get_current_token(), 126)
self.get_success(test_gen_next())
def test_multiple_gen_nexts_closed_in_different_order(self) -> None:
"""Check that we handle overlapping calls to gen_next, even when their IDs
created and persisted in different orders."""
id_gen = self._create_id_generator()
async def test_gen_next() -> None:
ctx1 = id_gen.get_next()
ctx2 = id_gen.get_next()
ctx3 = id_gen.get_next()
# Request three new stream IDs.
self.assertEqual(await ctx1.__aenter__(), 124)
self.assertEqual(await ctx2.__aenter__(), 125)
self.assertEqual(await ctx3.__aenter__(), 126)
# None are persisted: current token unchanged.
self.assertEqual(id_gen.get_current_token(), 123)
# Persist them in a different order, starting with 126 from ctx3.
await ctx3.__aexit__(None, None, None)
# We haven't persisted 124 from ctx1 yet---current token is still 123.
self.assertEqual(id_gen.get_current_token(), 123)
# Now persist 124 from ctx1.
await ctx1.__aexit__(None, None, None)
# Current token is then 124, waiting for 125 to be persisted.
self.assertEqual(id_gen.get_current_token(), 124)
# Finally persist 125 from ctx2.
await ctx2.__aexit__(None, None, None)
# Current token is then 126 (skipping over 125).
self.assertEqual(id_gen.get_current_token(), 126)
self.get_success(test_gen_next())
def test_gen_next_while_still_waiting_for_persistence(self) -> None:
"""Check that we handle overlapping calls to gen_next."""
id_gen = self._create_id_generator()
async def test_gen_next() -> None:
ctx1 = id_gen.get_next()
ctx2 = id_gen.get_next()
ctx3 = id_gen.get_next()
# Request two new stream IDs.
self.assertEqual(await ctx1.__aenter__(), 124)
self.assertEqual(await ctx2.__aenter__(), 125)
# Persist ctx2 first.
await ctx2.__aexit__(None, None, None)
# Still waiting on ctx1's ID to be persisted.
self.assertEqual(id_gen.get_current_token(), 123)
# Now request a third stream ID. It should be 126 (the smallest ID that
# we've not yet handed out.)
self.assertEqual(await ctx3.__aenter__(), 126)
self.get_success(test_gen_next())
class MultiWriterIdGeneratorTestCase(HomeserverTestCase):
if not USE_POSTGRES_FOR_TESTS:
skip = "Requires Postgres"
def prepare(self, reactor: MemoryReactor, clock: Clock, hs: HomeServer) -> None:
self.store = hs.get_datastores().main
self.db_pool: DatabasePool = self.store.db_pool
self.get_success(self.db_pool.runInteraction("_setup_db", self._setup_db))
def _setup_db(self, txn: LoggingTransaction) -> None:
txn.execute("CREATE SEQUENCE foobar_seq")
txn.execute(
"""
CREATE TABLE foobar (
stream_id BIGINT NOT NULL,
instance_name TEXT NOT NULL,
data TEXT
);
"""
)
def _create_id_generator(
self, instance_name: str = "master", writers: Optional[List[str]] = None
) -> MultiWriterIdGenerator:
def _create(conn: LoggingDatabaseConnection) -> MultiWriterIdGenerator:
return MultiWriterIdGenerator(
conn,
self.db_pool,
notifier=self.hs.get_replication_notifier(),
stream_name="test_stream",
instance_name=instance_name,
tables=[("foobar", "instance_name", "stream_id")],
sequence_name="foobar_seq",
writers=writers or ["master"],
)
return self.get_success_or_raise(self.db_pool.runWithConnection(_create))
def _insert_rows(self, instance_name: str, number: int) -> None:
"""Insert N rows as the given instance, inserting with stream IDs pulled
from the postgres sequence.
"""
def _insert(txn: LoggingTransaction) -> None:
for _ in range(number):
txn.execute(
"INSERT INTO foobar VALUES (nextval('foobar_seq'), ?)",
(instance_name,),
)
txn.execute(
"""
INSERT INTO stream_positions VALUES ('test_stream', ?, lastval())
ON CONFLICT (stream_name, instance_name) DO UPDATE SET stream_id = lastval()
""",
(instance_name,),
)
self.get_success(self.db_pool.runInteraction("_insert_rows", _insert))
def _insert_row_with_id(self, instance_name: str, stream_id: int) -> None:
"""Insert one row as the given instance with given stream_id, updating
the postgres sequence position to match.
"""
def _insert(txn: LoggingTransaction) -> None:
txn.execute(
"INSERT INTO foobar VALUES (?, ?)",
(
stream_id,
instance_name,
),
)
txn.execute("SELECT setval('foobar_seq', ?)", (stream_id,))
txn.execute(
"""
INSERT INTO stream_positions VALUES ('test_stream', ?, ?)
ON CONFLICT (stream_name, instance_name) DO UPDATE SET stream_id = ?
""",
(instance_name, stream_id, stream_id),
)
self.get_success(self.db_pool.runInteraction("_insert_row_with_id", _insert))
def test_empty(self) -> None:
"""Test an ID generator against an empty database gives sensible
current positions.
"""
id_gen = self._create_id_generator()
# The table is empty so we expect the map for positions to have a dummy
# minimum value.
self.assertEqual(id_gen.get_positions(), {"master": 1})
def test_single_instance(self) -> None:
"""Test that reads and writes from a single process are handled
correctly.
"""
# Prefill table with 7 rows written by 'master'
self._insert_rows("master", 7)
id_gen = self._create_id_generator()
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
# Try allocating a new ID gen and check that we only see position
# advanced after we leave the context manager.
async def _get_next_async() -> None:
async with id_gen.get_next() as stream_id:
self.assertEqual(stream_id, 8)
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
self.get_success(_get_next_async())
self.assertEqual(id_gen.get_positions(), {"master": 8})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 8)
def test_out_of_order_finish(self) -> None:
"""Test that IDs persisted out of order are correctly handled"""
# Prefill table with 7 rows written by 'master'
self._insert_rows("master", 7)
id_gen = self._create_id_generator()
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
ctx1 = id_gen.get_next()
ctx2 = id_gen.get_next()
ctx3 = id_gen.get_next()
ctx4 = id_gen.get_next()
s1 = self.get_success(ctx1.__aenter__())
s2 = self.get_success(ctx2.__aenter__())
s3 = self.get_success(ctx3.__aenter__())
s4 = self.get_success(ctx4.__aenter__())
self.assertEqual(s1, 8)
self.assertEqual(s2, 9)
self.assertEqual(s3, 10)
self.assertEqual(s4, 11)
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
self.get_success(ctx2.__aexit__(None, None, None))
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
self.get_success(ctx1.__aexit__(None, None, None))
self.assertEqual(id_gen.get_positions(), {"master": 9})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 9)
self.get_success(ctx4.__aexit__(None, None, None))
self.assertEqual(id_gen.get_positions(), {"master": 9})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 9)
self.get_success(ctx3.__aexit__(None, None, None))
self.assertEqual(id_gen.get_positions(), {"master": 11})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 11)
def test_multi_instance(self) -> None:
"""Test that reads and writes from multiple processes are handled
correctly.
"""
self._insert_rows("first", 3)
self._insert_rows("second", 4)
first_id_gen = self._create_id_generator("first", writers=["first", "second"])
second_id_gen = self._create_id_generator("second", writers=["first", "second"])
self.assertEqual(first_id_gen.get_positions(), {"first": 3, "second": 7})
self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 7)
self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 7)
self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 7})
self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
# Try allocating a new ID gen and check that we only see position
# advanced after we leave the context manager.
async def _get_next_async() -> None:
async with first_id_gen.get_next() as stream_id:
self.assertEqual(stream_id, 8)
self.assertEqual(
first_id_gen.get_positions(), {"first": 3, "second": 7}
)
self.assertEqual(first_id_gen.get_persisted_upto_position(), 7)
self.get_success(_get_next_async())
self.assertEqual(first_id_gen.get_positions(), {"first": 8, "second": 7})
# However the ID gen on the second instance won't have seen the update
self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 7})
# ... but calling `get_next` on the second instance should give a unique
# stream ID
async def _get_next_async2() -> None:
async with second_id_gen.get_next() as stream_id:
self.assertEqual(stream_id, 9)
self.assertEqual(
second_id_gen.get_positions(), {"first": 3, "second": 7}
)
self.get_success(_get_next_async2())
self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 9})
# If the second ID gen gets told about the first, it correctly updates
second_id_gen.advance("first", 8)
self.assertEqual(second_id_gen.get_positions(), {"first": 8, "second": 9})
def test_multi_instance_empty_row(self) -> None:
"""Test that reads and writes from multiple processes are handled
correctly, when one of the writers starts without any rows.
"""
# Insert some rows for two out of three of the ID gens.
self._insert_rows("first", 3)
self._insert_rows("second", 4)
first_id_gen = self._create_id_generator(
"first", writers=["first", "second", "third"]
)
second_id_gen = self._create_id_generator(
"second", writers=["first", "second", "third"]
)
third_id_gen = self._create_id_generator(
"third", writers=["first", "second", "third"]
)
self.assertEqual(
first_id_gen.get_positions(), {"first": 3, "second": 7, "third": 7}
)
self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 7)
self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 7)
self.assertEqual(first_id_gen.get_current_token_for_writer("third"), 7)
self.assertEqual(
second_id_gen.get_positions(), {"first": 3, "second": 7, "third": 7}
)
self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
self.assertEqual(second_id_gen.get_current_token_for_writer("third"), 7)
# Try allocating a new ID gen and check that we only see position
# advanced after we leave the context manager.
async def _get_next_async() -> None:
async with third_id_gen.get_next() as stream_id:
self.assertEqual(stream_id, 8)
self.assertEqual(
third_id_gen.get_positions(), {"first": 3, "second": 7, "third": 7}
)
self.assertEqual(third_id_gen.get_persisted_upto_position(), 7)
self.get_success(_get_next_async())
self.assertEqual(
third_id_gen.get_positions(), {"first": 3, "second": 7, "third": 8}
)
def test_get_next_txn(self) -> None:
"""Test that the `get_next_txn` function works correctly."""
# Prefill table with 7 rows written by 'master'
self._insert_rows("master", 7)
id_gen = self._create_id_generator()
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
# Try allocating a new ID gen and check that we only see position
# advanced after we leave the context manager.
def _get_next_txn(txn: LoggingTransaction) -> None:
stream_id = id_gen.get_next_txn(txn)
self.assertEqual(stream_id, 8)
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
self.get_success(self.db_pool.runInteraction("test", _get_next_txn))
self.assertEqual(id_gen.get_positions(), {"master": 8})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 8)
def test_get_persisted_upto_position(self) -> None:
"""Test that `get_persisted_upto_position` correctly tracks updates to
positions.
"""
# The following tests are a bit cheeky in that we notify about new
# positions via `advance` without *actually* advancing the postgres
# sequence.
self._insert_row_with_id("first", 3)
self._insert_row_with_id("second", 5)
id_gen = self._create_id_generator("worker", writers=["first", "second"])
self.assertEqual(id_gen.get_positions(), {"first": 3, "second": 5})
# Min is 3 and there is a gap between 5, so we expect it to be 3.
self.assertEqual(id_gen.get_persisted_upto_position(), 3)
# We advance "first" straight to 6. Min is now 5 but there is no gap so
# we expect it to be 6
id_gen.advance("first", 6)
self.assertEqual(id_gen.get_persisted_upto_position(), 6)
# No gap, so we expect 7.
id_gen.advance("second", 7)
self.assertEqual(id_gen.get_persisted_upto_position(), 7)
# We haven't seen 8 yet, so we expect 7 still.
id_gen.advance("second", 9)
self.assertEqual(id_gen.get_persisted_upto_position(), 7)
# Now that we've seen 7, 8 and 9 we can got straight to 9.
id_gen.advance("first", 8)
self.assertEqual(id_gen.get_persisted_upto_position(), 9)
# Jump forward with gaps. The minimum is 11, even though we haven't seen
# 10 we know that everything before 11 must be persisted.
id_gen.advance("first", 11)
id_gen.advance("second", 15)
self.assertEqual(id_gen.get_persisted_upto_position(), 11)
def test_get_persisted_upto_position_get_next(self) -> None:
"""Test that `get_persisted_upto_position` correctly tracks updates to
positions when `get_next` is called.
"""
self._insert_row_with_id("first", 3)
self._insert_row_with_id("second", 5)
id_gen = self._create_id_generator("first", writers=["first", "second"])
self.assertEqual(id_gen.get_positions(), {"first": 3, "second": 5})
self.assertEqual(id_gen.get_persisted_upto_position(), 5)
async def _get_next_async() -> None:
async with id_gen.get_next() as stream_id:
self.assertEqual(stream_id, 6)
self.assertEqual(id_gen.get_persisted_upto_position(), 5)
self.get_success(_get_next_async())
self.assertEqual(id_gen.get_persisted_upto_position(), 6)
# We assume that so long as `get_next` does correctly advance the
# `persisted_upto_position` in this case, then it will be correct in the
# other cases that are tested above (since they'll hit the same code).
def test_restart_during_out_of_order_persistence(self) -> None:
"""Test that restarting a process while another process is writing out
of order updates are handled correctly.
"""
# Prefill table with 7 rows written by 'master'
self._insert_rows("master", 7)
id_gen = self._create_id_generator()
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
# Persist two rows at once
ctx1 = id_gen.get_next()
ctx2 = id_gen.get_next()
s1 = self.get_success(ctx1.__aenter__())
s2 = self.get_success(ctx2.__aenter__())
self.assertEqual(s1, 8)
self.assertEqual(s2, 9)
self.assertEqual(id_gen.get_positions(), {"master": 7})
self.assertEqual(id_gen.get_current_token_for_writer("master"), 7)
# We finish persisting the second row before restart
self.get_success(ctx2.__aexit__(None, None, None))
# We simulate a restart of another worker by just creating a new ID gen.
id_gen_worker = self._create_id_generator("worker")
# Restarted worker should not see the second persisted row
self.assertEqual(id_gen_worker.get_positions(), {"master": 7})
self.assertEqual(id_gen_worker.get_current_token_for_writer("master"), 7)
# Now if we persist the first row then both instances should jump ahead
# correctly.
self.get_success(ctx1.__aexit__(None, None, None))
self.assertEqual(id_gen.get_positions(), {"master": 9})
id_gen_worker.advance("master", 9)
self.assertEqual(id_gen_worker.get_positions(), {"master": 9})
def test_writer_config_change(self) -> None:
"""Test that changing the writer config correctly works."""
self._insert_row_with_id("first", 3)
self._insert_row_with_id("second", 5)
# Initial config has two writers
id_gen = self._create_id_generator("worker", writers=["first", "second"])
self.assertEqual(id_gen.get_persisted_upto_position(), 3)
self.assertEqual(id_gen.get_current_token_for_writer("first"), 3)
self.assertEqual(id_gen.get_current_token_for_writer("second"), 5)
# New config removes one of the configs. Note that if the writer is
# removed from config we assume that it has been shut down and has
# finished persisting, hence why the persisted upto position is 5.
id_gen_2 = self._create_id_generator("second", writers=["second"])
self.assertEqual(id_gen_2.get_persisted_upto_position(), 5)
self.assertEqual(id_gen_2.get_current_token_for_writer("second"), 5)
# This config points to a single, previously unused writer.
id_gen_3 = self._create_id_generator("third", writers=["third"])
self.assertEqual(id_gen_3.get_persisted_upto_position(), 5)
# For new writers we assume their initial position to be the current
# persisted up to position. This stops Synapse from doing a full table
# scan when a new writer comes along.
self.assertEqual(id_gen_3.get_current_token_for_writer("third"), 5)
id_gen_4 = self._create_id_generator("fourth", writers=["third"])
self.assertEqual(id_gen_4.get_current_token_for_writer("third"), 5)
# Check that we get a sane next stream ID with this new config.
async def _get_next_async() -> None:
async with id_gen_3.get_next() as stream_id:
self.assertEqual(stream_id, 6)
self.get_success(_get_next_async())
self.assertEqual(id_gen_3.get_persisted_upto_position(), 6)
# If we add back the old "first" then we shouldn't see the persisted up
# to position revert back to 3.
id_gen_5 = self._create_id_generator("five", writers=["first", "third"])
self.assertEqual(id_gen_5.get_persisted_upto_position(), 6)
self.assertEqual(id_gen_5.get_current_token_for_writer("first"), 6)
self.assertEqual(id_gen_5.get_current_token_for_writer("third"), 6)
def test_sequence_consistency(self) -> None:
"""Test that we error out if the table and sequence diverges."""
# Prefill with some rows
self._insert_row_with_id("master", 3)
# Now we add a row *without* updating the stream ID
def _insert(txn: Cursor) -> None:
txn.execute("INSERT INTO foobar VALUES (26, 'master')")
self.get_success(self.db_pool.runInteraction("_insert", _insert))
# Creating the ID gen should error
with self.assertRaises(IncorrectDatabaseSetup):
self._create_id_generator("first")
def test_minimal_local_token(self) -> None:
self._insert_rows("first", 3)
self._insert_rows("second", 4)
first_id_gen = self._create_id_generator("first", writers=["first", "second"])
second_id_gen = self._create_id_generator("second", writers=["first", "second"])
self.assertEqual(first_id_gen.get_positions(), {"first": 3, "second": 7})
self.assertEqual(first_id_gen.get_minimal_local_current_token(), 3)
self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 7})
self.assertEqual(second_id_gen.get_minimal_local_current_token(), 7)
def test_current_token_gap(self) -> None:
"""Test that getting the current token for a writer returns the maximal
token when there are no writes.
"""
self._insert_rows("first", 3)
self._insert_rows("second", 4)
first_id_gen = self._create_id_generator(
"first", writers=["first", "second", "third"]
)
second_id_gen = self._create_id_generator(
"second", writers=["first", "second", "third"]
)
self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
self.assertEqual(second_id_gen.get_current_token(), 7)
# Check that the first ID gen advancing causes the second ID gen to
# advance (as the second ID gen has nothing in flight).
async def _get_next_async() -> None:
async with first_id_gen.get_next_mult(2):
pass
self.get_success(_get_next_async())
second_id_gen.advance("first", 9)
self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 9)
self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 9)
self.assertEqual(second_id_gen.get_current_token(), 7)
# Check that the first ID gen advancing doesn't advance the second ID
# gen when the second ID gen has stuff in flight.
self.get_success(_get_next_async())
ctxmgr = second_id_gen.get_next()
self.get_success(ctxmgr.__aenter__())
second_id_gen.advance("first", 11)
self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 11)
self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 9)
self.assertEqual(second_id_gen.get_current_token(), 7)
self.get_success(ctxmgr.__aexit__(None, None, None))
self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 11)
self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 12)
self.assertEqual(second_id_gen.get_current_token(), 7)
class BackwardsMultiWriterIdGeneratorTestCase(HomeserverTestCase):
"""Tests MultiWriterIdGenerator that produce *negative* stream IDs."""
if not USE_POSTGRES_FOR_TESTS:
skip = "Requires Postgres"
def prepare(self, reactor: MemoryReactor, clock: Clock, hs: HomeServer) -> None:
self.store = hs.get_datastores().main
self.db_pool: DatabasePool = self.store.db_pool
self.get_success(self.db_pool.runInteraction("_setup_db", self._setup_db))
def _setup_db(self, txn: LoggingTransaction) -> None:
txn.execute("CREATE SEQUENCE foobar_seq")
txn.execute(
"""
CREATE TABLE foobar (
stream_id BIGINT NOT NULL,
instance_name TEXT NOT NULL,
data TEXT
);
"""
)
def _create_id_generator(
self, instance_name: str = "master", writers: Optional[List[str]] = None
) -> MultiWriterIdGenerator:
def _create(conn: LoggingDatabaseConnection) -> MultiWriterIdGenerator:
return MultiWriterIdGenerator(
conn,
self.db_pool,
notifier=self.hs.get_replication_notifier(),
stream_name="test_stream",
instance_name=instance_name,
tables=[("foobar", "instance_name", "stream_id")],
sequence_name="foobar_seq",
writers=writers or ["master"],
positive=False,
)
return self.get_success(self.db_pool.runWithConnection(_create))
def _insert_row(self, instance_name: str, stream_id: int) -> None:
"""Insert one row as the given instance with given stream_id."""
def _insert(txn: LoggingTransaction) -> None:
txn.execute(
"INSERT INTO foobar VALUES (?, ?)",
(
stream_id,
instance_name,
),
)
txn.execute(
"""
INSERT INTO stream_positions VALUES ('test_stream', ?, ?)
ON CONFLICT (stream_name, instance_name) DO UPDATE SET stream_id = ?
""",
(instance_name, -stream_id, -stream_id),
)
self.get_success(self.db_pool.runInteraction("_insert_row", _insert))
def test_single_instance(self) -> None:
"""Test that reads and writes from a single process are handled
correctly.
"""
id_gen = self._create_id_generator()
async def _get_next_async() -> None:
async with id_gen.get_next() as stream_id:
self._insert_row("master", stream_id)
self.get_success(_get_next_async())
self.assertEqual(id_gen.get_positions(), {"master": -1})
self.assertEqual(id_gen.get_current_token_for_writer("master"), -1)
self.assertEqual(id_gen.get_persisted_upto_position(), -1)
async def _get_next_async2() -> None:
async with id_gen.get_next_mult(3) as stream_ids:
for stream_id in stream_ids:
self._insert_row("master", stream_id)
self.get_success(_get_next_async2())
self.assertEqual(id_gen.get_positions(), {"master": -4})
self.assertEqual(id_gen.get_current_token_for_writer("master"), -4)
self.assertEqual(id_gen.get_persisted_upto_position(), -4)
# Test loading from DB by creating a second ID gen
second_id_gen = self._create_id_generator()
self.assertEqual(second_id_gen.get_positions(), {"master": -4})
self.assertEqual(second_id_gen.get_current_token_for_writer("master"), -4)
self.assertEqual(second_id_gen.get_persisted_upto_position(), -4)
def test_multiple_instance(self) -> None:
"""Tests that having multiple instances that get advanced over
federation works corretly.
"""
id_gen_1 = self._create_id_generator("first", writers=["first", "second"])
id_gen_2 = self._create_id_generator("second", writers=["first", "second"])
async def _get_next_async() -> None:
async with id_gen_1.get_next() as stream_id:
self._insert_row("first", stream_id)
id_gen_2.advance("first", stream_id)
self.get_success(_get_next_async())
self.assertEqual(id_gen_1.get_positions(), {"first": -1, "second": -1})
self.assertEqual(id_gen_2.get_positions(), {"first": -1, "second": -1})
self.assertEqual(id_gen_1.get_persisted_upto_position(), -1)
self.assertEqual(id_gen_2.get_persisted_upto_position(), -1)
async def _get_next_async2() -> None:
async with id_gen_2.get_next() as stream_id:
self._insert_row("second", stream_id)
id_gen_1.advance("second", stream_id)
self.get_success(_get_next_async2())
self.assertEqual(id_gen_1.get_positions(), {"first": -1, "second": -2})
self.assertEqual(id_gen_2.get_positions(), {"first": -1, "second": -2})
self.assertEqual(id_gen_1.get_persisted_upto_position(), -2)
self.assertEqual(id_gen_2.get_persisted_upto_position(), -2)
class MultiTableMultiWriterIdGeneratorTestCase(HomeserverTestCase):
if not USE_POSTGRES_FOR_TESTS:
skip = "Requires Postgres"
def prepare(self, reactor: MemoryReactor, clock: Clock, hs: HomeServer) -> None:
self.store = hs.get_datastores().main
self.db_pool: DatabasePool = self.store.db_pool
self.get_success(self.db_pool.runInteraction("_setup_db", self._setup_db))
def _setup_db(self, txn: LoggingTransaction) -> None:
txn.execute("CREATE SEQUENCE foobar_seq")
txn.execute(
"""
CREATE TABLE foobar1 (
stream_id BIGINT NOT NULL,
instance_name TEXT NOT NULL,
data TEXT
);
"""
)
txn.execute(
"""
CREATE TABLE foobar2 (
stream_id BIGINT NOT NULL,
instance_name TEXT NOT NULL,
data TEXT
);
"""
)
def _create_id_generator(
self, instance_name: str = "master", writers: Optional[List[str]] = None
) -> MultiWriterIdGenerator:
def _create(conn: LoggingDatabaseConnection) -> MultiWriterIdGenerator:
return MultiWriterIdGenerator(
conn,
self.db_pool,
notifier=self.hs.get_replication_notifier(),
stream_name="test_stream",
instance_name=instance_name,
tables=[
("foobar1", "instance_name", "stream_id"),
("foobar2", "instance_name", "stream_id"),
],
sequence_name="foobar_seq",
writers=writers or ["master"],
)
return self.get_success_or_raise(self.db_pool.runWithConnection(_create))
def _insert_rows(
self,
table: str,
instance_name: str,
number: int,
update_stream_table: bool = True,
) -> None:
"""Insert N rows as the given instance, inserting with stream IDs pulled
from the postgres sequence.
"""
def _insert(txn: LoggingTransaction) -> None:
for _ in range(number):
txn.execute(
"INSERT INTO %s VALUES (nextval('foobar_seq'), ?)" % (table,),
(instance_name,),
)
if update_stream_table:
txn.execute(
"""
INSERT INTO stream_positions VALUES ('test_stream', ?, lastval())
ON CONFLICT (stream_name, instance_name) DO UPDATE SET stream_id = lastval()
""",
(instance_name,),
)
self.get_success(self.db_pool.runInteraction("_insert_rows", _insert))
def test_load_existing_stream(self) -> None:
"""Test creating ID gens with multiple tables that have rows from after
the position in `stream_positions` table.
"""
self._insert_rows("foobar1", "first", 3)
self._insert_rows("foobar2", "second", 3)
self._insert_rows("foobar2", "second", 1, update_stream_table=False)
first_id_gen = self._create_id_generator("first", writers=["first", "second"])
second_id_gen = self._create_id_generator("second", writers=["first", "second"])
self.assertEqual(first_id_gen.get_positions(), {"first": 3, "second": 6})
self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 7)
self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 7)
self.assertEqual(first_id_gen.get_persisted_upto_position(), 7)
self.assertEqual(second_id_gen.get_positions(), {"first": 3, "second": 7})
self.assertEqual(second_id_gen.get_current_token_for_writer("first"), 7)
self.assertEqual(second_id_gen.get_current_token_for_writer("second"), 7)
self.assertEqual(second_id_gen.get_persisted_upto_position(), 7)
|