summary refs log tree commit diff
path: root/tests/replication/tcp/test_handler.py
blob: a8eb7fc523c2a718c9a89fe98abd837d89cf3e5c (plain) (blame)
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
#
# This file is licensed under the Affero General Public License (AGPL) version 3.
#
# Copyright 2022 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 twisted.internet import defer

from synapse.replication.tcp.commands import PositionCommand

from tests.replication._base import BaseMultiWorkerStreamTestCase


class ChannelsTestCase(BaseMultiWorkerStreamTestCase):
    def test_subscribed_to_enough_redis_channels(self) -> None:
        # The default main process is subscribed to the USER_IP channel.
        self.assertCountEqual(
            self.hs.get_replication_command_handler()._channels_to_subscribe_to,
            ["USER_IP"],
        )

    def test_background_worker_subscribed_to_user_ip(self) -> None:
        # The default main process is subscribed to the USER_IP channel.
        worker1 = self.make_worker_hs(
            "synapse.app.generic_worker",
            extra_config={
                "worker_name": "worker1",
                "run_background_tasks_on": "worker1",
                "redis": {"enabled": True},
            },
        )
        self.assertIn(
            "USER_IP",
            worker1.get_replication_command_handler()._channels_to_subscribe_to,
        )

        # Advance so the Redis subscription gets processed
        self.pump(0.1)

        # The counts are 2 because both the main process and the worker are subscribed.
        self.assertEqual(len(self._redis_server._subscribers_by_channel[b"test"]), 2)
        self.assertEqual(
            len(self._redis_server._subscribers_by_channel[b"test/USER_IP"]), 2
        )

    def test_non_background_worker_not_subscribed_to_user_ip(self) -> None:
        # The default main process is subscribed to the USER_IP channel.
        worker2 = self.make_worker_hs(
            "synapse.app.generic_worker",
            extra_config={
                "worker_name": "worker2",
                "run_background_tasks_on": "worker1",
                "redis": {"enabled": True},
            },
        )
        self.assertNotIn(
            "USER_IP",
            worker2.get_replication_command_handler()._channels_to_subscribe_to,
        )

        # Advance so the Redis subscription gets processed
        self.pump(0.1)

        # The count is 2 because both the main process and the worker are subscribed.
        self.assertEqual(len(self._redis_server._subscribers_by_channel[b"test"]), 2)
        # For USER_IP, the count is 1 because only the main process is subscribed.
        self.assertEqual(
            len(self._redis_server._subscribers_by_channel[b"test/USER_IP"]), 1
        )

    def test_wait_for_stream_position(self) -> None:
        """Check that wait for stream position correctly waits for an update from the
        correct instance.
        """
        store = self.hs.get_datastores().main
        cmd_handler = self.hs.get_replication_command_handler()
        data_handler = self.hs.get_replication_data_handler()

        worker1 = self.make_worker_hs(
            "synapse.app.generic_worker",
            extra_config={
                "worker_name": "worker1",
                "run_background_tasks_on": "worker1",
                "redis": {"enabled": True},
            },
        )

        cache_id_gen = worker1.get_datastores().main._cache_id_gen
        assert cache_id_gen is not None

        self.replicate()

        # First, make sure the master knows that `worker1` exists.
        initial_token = cache_id_gen.get_current_token()
        cmd_handler.send_command(
            PositionCommand("caches", "worker1", initial_token, initial_token)
        )
        self.replicate()

        # Next send out a normal RDATA, and check that waiting for that stream
        # ID returns immediately.
        ctx = cache_id_gen.get_next()
        next_token = self.get_success(ctx.__aenter__())
        self.get_success(ctx.__aexit__(None, None, None))

        self.get_success(
            data_handler.wait_for_stream_position("worker1", "caches", next_token)
        )

        # `wait_for_stream_position` should only return once master receives a
        # notification that `next_token` has persisted.
        ctx_worker1 = cache_id_gen.get_next()
        next_token = self.get_success(ctx_worker1.__aenter__())

        d = defer.ensureDeferred(
            data_handler.wait_for_stream_position("worker1", "caches", next_token)
        )
        self.assertFalse(d.called)

        # ... updating the cache ID gen on the master still shouldn't cause the
        # deferred to wake up.
        assert store._cache_id_gen is not None
        ctx = store._cache_id_gen.get_next()
        self.get_success(ctx.__aenter__())
        self.get_success(ctx.__aexit__(None, None, None))

        d = defer.ensureDeferred(
            data_handler.wait_for_stream_position("worker1", "caches", next_token)
        )
        self.assertFalse(d.called)

        # ... but worker1 finishing (and so sending an update) should.
        self.get_success(ctx_worker1.__aexit__(None, None, None))

        self.assertTrue(d.called)

    def test_wait_for_stream_position_rdata(self) -> None:
        """Check that wait for stream position correctly waits for an update
        from the correct instance, when RDATA is sent.
        """
        store = self.hs.get_datastores().main
        cmd_handler = self.hs.get_replication_command_handler()
        data_handler = self.hs.get_replication_data_handler()

        worker1 = self.make_worker_hs(
            "synapse.app.generic_worker",
            extra_config={
                "worker_name": "worker1",
                "run_background_tasks_on": "worker1",
                "redis": {"enabled": True},
            },
        )

        cache_id_gen = worker1.get_datastores().main._cache_id_gen
        assert cache_id_gen is not None

        self.replicate()

        # First, make sure the master knows that `worker1` exists.
        initial_token = cache_id_gen.get_current_token()
        cmd_handler.send_command(
            PositionCommand("caches", "worker1", initial_token, initial_token)
        )
        self.replicate()

        # `wait_for_stream_position` should only return once master receives a
        # notification that `next_token2` has persisted.
        ctx_worker1 = cache_id_gen.get_next_mult(2)
        next_token1, next_token2 = self.get_success(ctx_worker1.__aenter__())

        d = defer.ensureDeferred(
            data_handler.wait_for_stream_position("worker1", "caches", next_token2)
        )
        self.assertFalse(d.called)

        # Insert an entry into the cache stream with token `next_token1`, but
        # not `next_token2`.
        self.get_success(
            store.db_pool.simple_insert(
                table="cache_invalidation_stream_by_instance",
                values={
                    "stream_id": next_token1,
                    "instance_name": "worker1",
                    "cache_func": "foo",
                    "keys": [],
                    "invalidation_ts": 0,
                },
            )
        )

        # Finish the context manager, triggering the data to be sent to master.
        self.get_success(ctx_worker1.__aexit__(None, None, None))

        # Master should get told about `next_token2`, so the deferred should
        # resolve.
        self.assertTrue(d.called)