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
|
/**
* Copyright (C) 2017 MongoDB, Inc.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License, version 3,
* as published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* As a special exception, the copyright holders give permission to link the
* code of portions of this program with the OpenSSL library under certain
* conditions as described in each individual source file and distribute
* linked combinations including the program with the OpenSSL library. You
* must comply with the GNU Affero General Public License in all respects for
* all of the code used other than as permitted herein. If you modify file(s)
* with this exception, you may extend this exception to your version of the
* file(s), but you are not obligated to do so. If you do not wish to do so,
* delete this exception statement from your version. If you delete this
* exception statement from all source files in the program, then also delete
* it in the license file.
*/
#define MONGO_LOG_DEFAULT_COMPONENT ::mongo::logger::LogComponent::kDefault
#include "mongo/platform/basic.h"
#include "mongo/bson/bsonobj.h"
#include "mongo/bson/timestamp.h"
#include "mongo/db/dbdirectclient.h"
#include "mongo/db/logical_clock.h"
#include "mongo/db/logical_clock_test_fixture.h"
#include "mongo/db/logical_time.h"
#include "mongo/db/repl/replication_coordinator_mock.h"
#include "mongo/stdx/memory.h"
#include "mongo/unittest/unittest.h"
#include "mongo/util/clock_source_mock.h"
#include "mongo/util/log.h"
namespace mongo {
namespace {
std::string kDummyNamespaceString = "test.foo";
using LogicalClockTest = LogicalClockTestFixture;
LogicalTime buildLogicalTime(unsigned secs, unsigned inc) {
return LogicalTime(Timestamp(secs, inc));
}
// Check that the initial time does not change during logicalClock creation.
TEST_F(LogicalClockTest, roundtrip) {
Timestamp tX(1);
auto time = LogicalTime(tX);
getClock()->setClusterTimeFromTrustedSource(time);
auto storedTime(getClock()->getClusterTime());
ASSERT_TRUE(storedTime == time);
}
// Verify the reserve ticks functionality.
TEST_F(LogicalClockTest, reserveTicks) {
// Set clock to a non-zero time, so we can verify wall clock synchronization.
setMockClockSourceTime(Date_t::fromMillisSinceEpoch(10 * 1000));
auto t1 = getClock()->reserveTicks(1);
auto t2(getClock()->getClusterTime());
ASSERT_TRUE(t1 == t2);
// Make sure we synchronized with the wall clock.
ASSERT_TRUE(t2.asTimestamp().getSecs() == 10);
auto t3 = getClock()->reserveTicks(1);
t1.addTicks(1);
ASSERT_TRUE(t3 == t1);
t3 = getClock()->reserveTicks(100);
t1.addTicks(1);
ASSERT_TRUE(t3 == t1);
t3 = getClock()->reserveTicks(1);
t1.addTicks(100);
ASSERT_TRUE(t3 == t1);
// Ensure overflow to a new second.
auto initTimeSecs = getClock()->getClusterTime().asTimestamp().getSecs();
getClock()->reserveTicks((1U << 31) - 1);
auto newTimeSecs = getClock()->getClusterTime().asTimestamp().getSecs();
ASSERT_TRUE(newTimeSecs == initTimeSecs + 1);
}
// Verify the advanceClusterTime functionality.
TEST_F(LogicalClockTest, advanceClusterTime) {
auto t1 = getClock()->reserveTicks(1);
t1.addTicks(100);
ASSERT_OK(getClock()->advanceClusterTime(t1));
ASSERT_TRUE(t1 == getClock()->getClusterTime());
}
// Verify rate limiter rejects logical times whose seconds values are too far ahead.
TEST_F(LogicalClockTest, RateLimiterRejectsLogicalTimesTooFarAhead) {
setMockClockSourceTime(Date_t::fromMillisSinceEpoch(10 * 1000));
Timestamp tooFarAheadTimestamp(
durationCount<Seconds>(getMockClockSourceTime().toDurationSinceEpoch()) +
durationCount<Seconds>(LogicalClock::kMaxAcceptableLogicalClockDriftSecs) +
10, // Add 10 seconds to ensure limit is exceeded.
1);
LogicalTime t1(tooFarAheadTimestamp);
ASSERT_EQ(ErrorCodes::ClusterTimeFailsRateLimiter, getClock()->advanceClusterTime(t1));
}
// Verify cluster time can be initialized to a very old time.
TEST_F(LogicalClockTest, InitFromTrustedSourceCanAcceptVeryOldLogicalTime) {
setMockClockSourceTime(Date_t::fromMillisSinceEpoch(
durationCount<Seconds>(LogicalClock::kMaxAcceptableLogicalClockDriftSecs) * 10 * 1000));
Timestamp veryOldTimestamp(
durationCount<Seconds>(getMockClockSourceTime().toDurationSinceEpoch()) -
(durationCount<Seconds>(LogicalClock::kMaxAcceptableLogicalClockDriftSecs) * 5));
auto veryOldTime = LogicalTime(veryOldTimestamp);
getClock()->setClusterTimeFromTrustedSource(veryOldTime);
ASSERT_TRUE(getClock()->getClusterTime() == veryOldTime);
}
// Verify writes to the oplog advance cluster time.
TEST_F(LogicalClockTest, WritesToOplogAdvanceClusterTime) {
Timestamp tX(1);
auto initialTime = LogicalTime(tX);
getClock()->setClusterTimeFromTrustedSource(initialTime);
ASSERT_TRUE(getClock()->getClusterTime() == initialTime);
getDBClient()->insert(kDummyNamespaceString, BSON("x" << 1));
ASSERT_TRUE(getClock()->getClusterTime() > initialTime);
ASSERT_EQ(getClock()->getClusterTime().asTimestamp(),
replicationCoordinator()->getMyLastAppliedOpTime().getTimestamp());
}
// Tests the scenario where an admin incorrectly sets the wall clock more than
// maxAcceptableLogicalClockDriftSecs in the past at startup, and cluster time is initialized to
// the incorrect past time, then the admin resets the clock to the current time. In this case,
// logical time can be advanced through metadata as long as the new time isn't
// maxAcceptableLogicalClockDriftSecs ahead of the correct current wall clock time, since the rate
// limiter compares new times to the wall clock, not the cluster time.
TEST_F(LogicalClockTest, WallClockSetTooFarInPast) {
auto oneDay = Seconds(24 * 60 * 60);
// Current wall clock and cluster time.
auto currentSecs = LogicalClock::kMaxAcceptableLogicalClockDriftSecs * 10;
LogicalTime currentTime(Timestamp(currentSecs, 0));
// Set wall clock more than maxAcceptableLogicalClockDriftSecs seconds in the past.
auto pastSecs = currentSecs - LogicalClock::kMaxAcceptableLogicalClockDriftSecs - oneDay;
setMockClockSourceTime(Date_t::fromDurationSinceEpoch(pastSecs));
// If cluster time is either uninitialized or even farther in the past, a write would set
// cluster time more than maxAcceptableLogicalClockDriftSecs in the past.
getDBClient()->insert(kDummyNamespaceString, BSON("x" << 1));
ASSERT_TRUE(
getClock()->getClusterTime() <
LogicalTime(Timestamp(currentSecs - LogicalClock::kMaxAcceptableLogicalClockDriftSecs, 0)));
// Set wall clock to the current time on the affected node.
setMockClockSourceTime(Date_t::fromDurationSinceEpoch(currentSecs));
// Verify that maxAcceptableLogicalClockDriftSecs parameter does not need to be increased to
// advance cluster time through metadata back to the current time.
ASSERT_OK(getClock()->advanceClusterTime(currentTime));
ASSERT_TRUE(getClock()->getClusterTime() == currentTime);
}
// Tests the scenario where an admin incorrectly sets the wall clock more than
// maxAcceptableLogicalClockDriftSecs in the future and a write is accepted, advancing cluster
// time, then the admin resets the clock to the current time. In this case, logical time cannot be
// advanced through metadata unless the drift parameter is increased.
TEST_F(LogicalClockTest, WallClockSetTooFarInFuture) {
auto oneDay = Seconds(24 * 60 * 60);
// Current wall clock and cluster time.
auto currentSecs = LogicalClock::kMaxAcceptableLogicalClockDriftSecs * 10;
LogicalTime currentTime(Timestamp(currentSecs, 0));
// Set wall clock more than maxAcceptableLogicalClockDriftSecs seconds in the future.
auto futureSecs = currentSecs + LogicalClock::kMaxAcceptableLogicalClockDriftSecs + oneDay;
setMockClockSourceTime(Date_t::fromDurationSinceEpoch(futureSecs));
// A write gets through and advances cluster time more than maxAcceptableLogicalClockDriftSecs
// in the future.
getDBClient()->insert(kDummyNamespaceString, BSON("x" << 1));
ASSERT_TRUE(
getClock()->getClusterTime() >
LogicalTime(Timestamp(currentSecs + LogicalClock::kMaxAcceptableLogicalClockDriftSecs, 0)));
// Set wall clock to the current time on the affected node.
setMockClockSourceTime(Date_t::fromDurationSinceEpoch(currentSecs));
// Verify that maxAcceptableLogicalClockDriftSecs parameter has to be increased to advance
// cluster time through metadata.
auto nextTime = getClock()->getClusterTime();
nextTime.addTicks(1); // The next lowest logical time.
ASSERT_EQ(ErrorCodes::ClusterTimeFailsRateLimiter, getClock()->advanceClusterTime(nextTime));
// Set wall clock to the current time + 1 day to simulate increasing the
// maxAcceptableLogicalClockDriftSecs parameter, which can only be set at startup, and verify
// time can be advanced through metadata again.
setMockClockSourceTime(Date_t::fromDurationSinceEpoch(currentSecs + oneDay));
ASSERT_OK(getClock()->advanceClusterTime(nextTime));
ASSERT_TRUE(getClock()->getClusterTime() == nextTime);
}
// Verify the behavior of advancing cluster time around the max allowed values.
TEST_F(LogicalClockTest, ReserveTicksBehaviorAroundMaxTime) {
unsigned maxVal = LogicalClock::kMaxSignedInt;
// Verify clock can be advanced near the max values.
// Can always advance to the max value for the inc field.
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal - 1, maxVal - 1));
getClock()->reserveTicks(1);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal - 1, maxVal));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal - 1, maxVal - 5));
getClock()->reserveTicks(5);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal - 1, maxVal));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(0, maxVal - 1));
getClock()->reserveTicks(1);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(0, maxVal));
// Can overflow inc into seconds to reach max seconds value.
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal - 1, maxVal));
getClock()->reserveTicks(1);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, 1));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal - 1, maxVal - 5));
getClock()->reserveTicks(10);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, 10));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal - 1, 1));
getClock()->reserveTicks(maxVal);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, maxVal));
// Can advance inc field when seconds field is at the max value.
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, 1));
getClock()->reserveTicks(1);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, 2));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, 1));
getClock()->reserveTicks(100);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, 101));
// Can advance to the max value for both the inc and seconds fields.
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, maxVal - 1));
getClock()->reserveTicks(1);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, maxVal));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, maxVal - 5));
getClock()->reserveTicks(5);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, maxVal));
// Verify scenarios where the clock cannot be advanced.
// Can't overflow inc into seconds when seconds field is at the max value.
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, maxVal));
ASSERT_THROWS(getClock()->reserveTicks(1), std::exception);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, maxVal));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, maxVal));
ASSERT_THROWS(getClock()->reserveTicks(5), std::exception);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, maxVal));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, maxVal - 1));
ASSERT_THROWS(getClock()->reserveTicks(2), std::exception);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, maxVal - 1));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(maxVal, maxVal - 11));
ASSERT_THROWS(getClock()->reserveTicks(12), std::exception);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, maxVal - 11));
}
// Verify behavior of advancing cluster time when the wall clock is near the max allowed value.
TEST_F(LogicalClockTest, ReserveTicksBehaviorWhenWallClockNearMaxTime) {
unsigned maxVal = LogicalClock::kMaxSignedInt;
// Can be set to the max possible time by catching up to the wall clock.
setMockClockSourceTime(Date_t::fromDurationSinceEpoch(Seconds(maxVal)));
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(1, 1));
getClock()->reserveTicks(1);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(maxVal, 1));
// Should fail when wall clock would advance cluster time beyond the max allowed time.
setMockClockSourceTime(Date_t::max());
resetClock()->setClusterTimeFromTrustedSource(buildLogicalTime(1, 1));
ASSERT_THROWS(getClock()->reserveTicks(1), std::exception);
ASSERT_TRUE(getClock()->getClusterTime() == buildLogicalTime(1, 1));
}
// Verify the clock rejects logical times greater than the max allowed time.
TEST_F(LogicalClockTest, RejectsLogicalTimesGreaterThanMaxTime) {
unsigned maxVal = LogicalClock::kMaxSignedInt;
// A logical time can be greater than the maximum value allowed because the signed integer
// maximum is used for legacy compatibility, but these fields are stored as unsigned integers.
auto beyondMaxTime = buildLogicalTime(maxVal + 1, maxVal + 1);
// The clock can't be initialized to a time greater than the max possible.
resetClock();
ASSERT_THROWS(getClock()->setClusterTimeFromTrustedSource(beyondMaxTime), std::exception);
ASSERT_TRUE(getClock()->getClusterTime() == LogicalTime());
// The time can't be advanced through metadata to a time greater than the max possible.
// Advance the wall clock close enough to the new value so the rate check is passed.
auto almostMaxSecs =
Seconds(maxVal) - LogicalClock::kMaxAcceptableLogicalClockDriftSecs + Seconds(10);
setMockClockSourceTime(Date_t::fromDurationSinceEpoch(almostMaxSecs));
ASSERT_THROWS(getClock()->advanceClusterTime(beyondMaxTime), std::exception);
ASSERT_TRUE(getClock()->getClusterTime() == LogicalTime());
}
} // unnamed namespace
} // namespace mongo
|