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
|
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
#ifndef ROCKSDB_LITE
#include <mutex>
#include <string>
#include <thread>
#include <vector>
#include "db/db_impl/db_impl.h"
#include "port/port.h"
#include "rocksdb/db.h"
#include "rocksdb/env.h"
#include "test_util/sync_point.h"
#include "test_util/testharness.h"
#include "util/string_util.h"
namespace ROCKSDB_NAMESPACE {
class CompactFilesTest : public testing::Test {
public:
CompactFilesTest() {
env_ = Env::Default();
db_name_ = test::PerThreadDBPath("compact_files_test");
}
std::string db_name_;
Env* env_;
};
// A class which remembers the name of each flushed file.
class FlushedFileCollector : public EventListener {
public:
FlushedFileCollector() {}
~FlushedFileCollector() override {}
void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
std::lock_guard<std::mutex> lock(mutex_);
flushed_files_.push_back(info.file_path);
}
std::vector<std::string> GetFlushedFiles() {
std::lock_guard<std::mutex> lock(mutex_);
std::vector<std::string> result;
for (auto fname : flushed_files_) {
result.push_back(fname);
}
return result;
}
void ClearFlushedFiles() {
std::lock_guard<std::mutex> lock(mutex_);
flushed_files_.clear();
}
private:
std::vector<std::string> flushed_files_;
std::mutex mutex_;
};
TEST_F(CompactFilesTest, L0ConflictsFiles) {
Options options;
// to trigger compaction more easily
const int kWriteBufferSize = 10000;
const int kLevel0Trigger = 2;
options.create_if_missing = true;
options.compaction_style = kCompactionStyleLevel;
// Small slowdown and stop trigger for experimental purpose.
options.level0_slowdown_writes_trigger = 20;
options.level0_stop_writes_trigger = 20;
options.level0_stop_writes_trigger = 20;
options.write_buffer_size = kWriteBufferSize;
options.level0_file_num_compaction_trigger = kLevel0Trigger;
options.compression = kNoCompression;
DB* db = nullptr;
DestroyDB(db_name_, options);
Status s = DB::Open(options, db_name_, &db);
assert(s.ok());
assert(db);
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency({
{"CompactFilesImpl:0", "BackgroundCallCompaction:0"},
{"BackgroundCallCompaction:1", "CompactFilesImpl:1"},
});
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
// create couple files
// Background compaction starts and waits in BackgroundCallCompaction:0
for (int i = 0; i < kLevel0Trigger * 4; ++i) {
db->Put(WriteOptions(), ToString(i), "");
db->Put(WriteOptions(), ToString(100 - i), "");
db->Flush(FlushOptions());
}
ROCKSDB_NAMESPACE::ColumnFamilyMetaData meta;
db->GetColumnFamilyMetaData(&meta);
std::string file1;
for (auto& file : meta.levels[0].files) {
ASSERT_EQ(0, meta.levels[0].level);
if (file1 == "") {
file1 = file.db_path + "/" + file.name;
} else {
std::string file2 = file.db_path + "/" + file.name;
// Another thread starts a compact files and creates an L0 compaction
// The background compaction then notices that there is an L0 compaction
// already in progress and doesn't do an L0 compaction
// Once the background compaction finishes, the compact files finishes
ASSERT_OK(db->CompactFiles(ROCKSDB_NAMESPACE::CompactionOptions(),
{file1, file2}, 0));
break;
}
}
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
delete db;
}
TEST_F(CompactFilesTest, ObsoleteFiles) {
Options options;
// to trigger compaction more easily
const int kWriteBufferSize = 65536;
options.create_if_missing = true;
// Disable RocksDB background compaction.
options.compaction_style = kCompactionStyleNone;
options.level0_slowdown_writes_trigger = (1 << 30);
options.level0_stop_writes_trigger = (1 << 30);
options.write_buffer_size = kWriteBufferSize;
options.max_write_buffer_number = 2;
options.compression = kNoCompression;
// Add listener
FlushedFileCollector* collector = new FlushedFileCollector();
options.listeners.emplace_back(collector);
DB* db = nullptr;
DestroyDB(db_name_, options);
Status s = DB::Open(options, db_name_, &db);
assert(s.ok());
assert(db);
// create couple files
for (int i = 1000; i < 2000; ++i) {
db->Put(WriteOptions(), ToString(i),
std::string(kWriteBufferSize / 10, 'a' + (i % 26)));
}
auto l0_files = collector->GetFlushedFiles();
ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files, 1));
reinterpret_cast<DBImpl*>(db)->TEST_WaitForCompact();
// verify all compaction input files are deleted
for (auto fname : l0_files) {
ASSERT_EQ(Status::NotFound(), env_->FileExists(fname));
}
delete db;
}
TEST_F(CompactFilesTest, NotCutOutputOnLevel0) {
Options options;
options.create_if_missing = true;
// Disable RocksDB background compaction.
options.compaction_style = kCompactionStyleNone;
options.level0_slowdown_writes_trigger = 1000;
options.level0_stop_writes_trigger = 1000;
options.write_buffer_size = 65536;
options.max_write_buffer_number = 2;
options.compression = kNoCompression;
options.max_compaction_bytes = 5000;
// Add listener
FlushedFileCollector* collector = new FlushedFileCollector();
options.listeners.emplace_back(collector);
DB* db = nullptr;
DestroyDB(db_name_, options);
Status s = DB::Open(options, db_name_, &db);
assert(s.ok());
assert(db);
// create couple files
for (int i = 0; i < 500; ++i) {
db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
}
reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
auto l0_files_1 = collector->GetFlushedFiles();
collector->ClearFlushedFiles();
for (int i = 0; i < 500; ++i) {
db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
}
reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
auto l0_files_2 = collector->GetFlushedFiles();
ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files_1, 0));
ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files_2, 0));
// no assertion failure
delete db;
}
TEST_F(CompactFilesTest, CapturingPendingFiles) {
Options options;
options.create_if_missing = true;
// Disable RocksDB background compaction.
options.compaction_style = kCompactionStyleNone;
// Always do full scans for obsolete files (needed to reproduce the issue).
options.delete_obsolete_files_period_micros = 0;
// Add listener.
FlushedFileCollector* collector = new FlushedFileCollector();
options.listeners.emplace_back(collector);
DB* db = nullptr;
DestroyDB(db_name_, options);
Status s = DB::Open(options, db_name_, &db);
assert(s.ok());
assert(db);
// Create 5 files.
for (int i = 0; i < 5; ++i) {
db->Put(WriteOptions(), "key" + ToString(i), "value");
db->Flush(FlushOptions());
}
auto l0_files = collector->GetFlushedFiles();
EXPECT_EQ(5, l0_files.size());
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency({
{"CompactFilesImpl:2", "CompactFilesTest.CapturingPendingFiles:0"},
{"CompactFilesTest.CapturingPendingFiles:1", "CompactFilesImpl:3"},
});
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
// Start compacting files.
ROCKSDB_NAMESPACE::port::Thread compaction_thread(
[&] { EXPECT_OK(db->CompactFiles(CompactionOptions(), l0_files, 1)); });
// In the meantime flush another file.
TEST_SYNC_POINT("CompactFilesTest.CapturingPendingFiles:0");
db->Put(WriteOptions(), "key5", "value");
db->Flush(FlushOptions());
TEST_SYNC_POINT("CompactFilesTest.CapturingPendingFiles:1");
compaction_thread.join();
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
delete db;
// Make sure we can reopen the DB.
s = DB::Open(options, db_name_, &db);
ASSERT_TRUE(s.ok());
assert(db);
delete db;
}
TEST_F(CompactFilesTest, CompactionFilterWithGetSv) {
class FilterWithGet : public CompactionFilter {
public:
bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
std::string* /*new_value*/,
bool* /*value_changed*/) const override {
if (db_ == nullptr) {
return true;
}
std::string res;
db_->Get(ReadOptions(), "", &res);
return true;
}
void SetDB(DB* db) {
db_ = db;
}
const char* Name() const override { return "FilterWithGet"; }
private:
DB* db_;
};
std::shared_ptr<FilterWithGet> cf(new FilterWithGet());
Options options;
options.create_if_missing = true;
options.compaction_filter = cf.get();
DB* db = nullptr;
DestroyDB(db_name_, options);
Status s = DB::Open(options, db_name_, &db);
ASSERT_OK(s);
cf->SetDB(db);
// Write one L0 file
db->Put(WriteOptions(), "K1", "V1");
db->Flush(FlushOptions());
// Compact all L0 files using CompactFiles
ROCKSDB_NAMESPACE::ColumnFamilyMetaData meta;
db->GetColumnFamilyMetaData(&meta);
for (auto& file : meta.levels[0].files) {
std::string fname = file.db_path + "/" + file.name;
ASSERT_OK(
db->CompactFiles(ROCKSDB_NAMESPACE::CompactionOptions(), {fname}, 0));
}
delete db;
}
TEST_F(CompactFilesTest, SentinelCompressionType) {
if (!Zlib_Supported()) {
fprintf(stderr, "zlib compression not supported, skip this test\n");
return;
}
if (!Snappy_Supported()) {
fprintf(stderr, "snappy compression not supported, skip this test\n");
return;
}
// Check that passing `CompressionType::kDisableCompressionOption` to
// `CompactFiles` causes it to use the column family compression options.
for (auto compaction_style :
{CompactionStyle::kCompactionStyleLevel,
CompactionStyle::kCompactionStyleUniversal,
CompactionStyle::kCompactionStyleNone}) {
DestroyDB(db_name_, Options());
Options options;
options.compaction_style = compaction_style;
// L0: Snappy, L1: ZSTD, L2: Snappy
options.compression_per_level = {CompressionType::kSnappyCompression,
CompressionType::kZlibCompression,
CompressionType::kSnappyCompression};
options.create_if_missing = true;
FlushedFileCollector* collector = new FlushedFileCollector();
options.listeners.emplace_back(collector);
DB* db = nullptr;
ASSERT_OK(DB::Open(options, db_name_, &db));
db->Put(WriteOptions(), "key", "val");
db->Flush(FlushOptions());
auto l0_files = collector->GetFlushedFiles();
ASSERT_EQ(1, l0_files.size());
// L0->L1 compaction, so output should be ZSTD-compressed
CompactionOptions compaction_opts;
compaction_opts.compression = CompressionType::kDisableCompressionOption;
ASSERT_OK(db->CompactFiles(compaction_opts, l0_files, 1));
ROCKSDB_NAMESPACE::TablePropertiesCollection all_tables_props;
ASSERT_OK(db->GetPropertiesOfAllTables(&all_tables_props));
for (const auto& name_and_table_props : all_tables_props) {
ASSERT_EQ(CompressionTypeToString(CompressionType::kZlibCompression),
name_and_table_props.second->compression_name);
}
delete db;
}
}
TEST_F(CompactFilesTest, GetCompactionJobInfo) {
Options options;
options.create_if_missing = true;
// Disable RocksDB background compaction.
options.compaction_style = kCompactionStyleNone;
options.level0_slowdown_writes_trigger = 1000;
options.level0_stop_writes_trigger = 1000;
options.write_buffer_size = 65536;
options.max_write_buffer_number = 2;
options.compression = kNoCompression;
options.max_compaction_bytes = 5000;
// Add listener
FlushedFileCollector* collector = new FlushedFileCollector();
options.listeners.emplace_back(collector);
DB* db = nullptr;
DestroyDB(db_name_, options);
Status s = DB::Open(options, db_name_, &db);
assert(s.ok());
assert(db);
// create couple files
for (int i = 0; i < 500; ++i) {
db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
}
reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
auto l0_files_1 = collector->GetFlushedFiles();
CompactionOptions co;
co.compression = CompressionType::kLZ4Compression;
CompactionJobInfo compaction_job_info{};
ASSERT_OK(
db->CompactFiles(co, l0_files_1, 0, -1, nullptr, &compaction_job_info));
ASSERT_EQ(compaction_job_info.base_input_level, 0);
ASSERT_EQ(compaction_job_info.cf_id, db->DefaultColumnFamily()->GetID());
ASSERT_EQ(compaction_job_info.cf_name, db->DefaultColumnFamily()->GetName());
ASSERT_EQ(compaction_job_info.compaction_reason,
CompactionReason::kManualCompaction);
ASSERT_EQ(compaction_job_info.compression, CompressionType::kLZ4Compression);
ASSERT_EQ(compaction_job_info.output_level, 0);
ASSERT_OK(compaction_job_info.status);
// no assertion failure
delete db;
}
} // namespace ROCKSDB_NAMESPACE
int main(int argc, char** argv) {
::testing::InitGoogleTest(&argc, argv);
return RUN_ALL_TESTS();
}
#else
#include <stdio.h>
int main(int /*argc*/, char** /*argv*/) {
fprintf(stderr,
"SKIPPED as DBImpl::CompactFiles is not supported in ROCKSDB_LITE\n");
return 0;
}
#endif // !ROCKSDB_LITE
|