rocksdb/db/db_write_test.cc

200 lines
6.7 KiB
C++
Raw Permalink Normal View History

// 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).
#include <atomic>
#include <memory>
#include <thread>
#include <vector>
#include "db/db_test_util.h"
#include "db/write_batch_internal.h"
#include "db/write_thread.h"
#include "port/port.h"
#include "port/stack_trace.h"
#include "test_util/fault_injection_test_env.h"
#include "test_util/sync_point.h"
#include "util/string_util.h"
namespace rocksdb {
// Test variations of WriteImpl.
class DBWriteTest : public DBTestBase, public testing::WithParamInterface<int> {
public:
DBWriteTest() : DBTestBase("/db_write_test") {}
Options GetOptions() { return DBTestBase::GetOptions(GetParam()); }
void Open() { DBTestBase::Reopen(GetOptions()); }
};
// It is invalid to do sync write while disabling WAL.
TEST_P(DBWriteTest, SyncAndDisableWAL) {
WriteOptions write_options;
write_options.sync = true;
write_options.disableWAL = true;
ASSERT_TRUE(dbfull()->Put(write_options, "foo", "bar").IsInvalidArgument());
WriteBatch batch;
ASSERT_OK(batch.Put("foo", "bar"));
ASSERT_TRUE(dbfull()->Write(write_options, &batch).IsInvalidArgument());
}
TEST_P(DBWriteTest, IOErrorOnWALWritePropagateToWriteThreadFollower) {
constexpr int kNumThreads = 5;
std::unique_ptr<FaultInjectionTestEnv> mock_env(
new FaultInjectionTestEnv(Env::Default()));
Options options = GetOptions();
options.env = mock_env.get();
Reopen(options);
std::atomic<int> ready_count{0};
std::atomic<int> leader_count{0};
std::vector<port::Thread> threads;
mock_env->SetFilesystemActive(false);
// Wait until all threads linked to write threads, to make sure
// all threads join the same batch group.
SyncPoint::GetInstance()->SetCallBack(
"WriteThread::JoinBatchGroup:Wait", [&](void* arg) {
ready_count++;
auto* w = reinterpret_cast<WriteThread::Writer*>(arg);
if (w->state == WriteThread::STATE_GROUP_LEADER) {
leader_count++;
while (ready_count < kNumThreads) {
// busy waiting
}
}
});
SyncPoint::GetInstance()->EnableProcessing();
for (int i = 0; i < kNumThreads; i++) {
threads.push_back(port::Thread(
[&](int index) {
// All threads should fail.
auto res = Put("key" + ToString(index), "value");
if (options.manual_wal_flush) {
ASSERT_TRUE(res.ok());
// we should see fs error when we do the flush
// TSAN reports a false alarm for lock-order-inversion but Open and
// FlushWAL are not run concurrently. Disabling this until TSAN is
// fixed.
// res = dbfull()->FlushWAL(false);
// ASSERT_FALSE(res.ok());
} else {
ASSERT_FALSE(res.ok());
}
},
i));
}
for (int i = 0; i < kNumThreads; i++) {
threads[i].join();
}
ASSERT_EQ(1, leader_count);
// Close before mock_env destruct.
Close();
}
TEST_P(DBWriteTest, ManualWalFlushInEffect) {
Options options = GetOptions();
Reopen(options);
// try the 1st WAL created during open
ASSERT_TRUE(Put("key" + ToString(0), "value").ok());
ASSERT_TRUE(options.manual_wal_flush != dbfull()->TEST_WALBufferIsEmpty());
ASSERT_TRUE(dbfull()->FlushWAL(false).ok());
ASSERT_TRUE(dbfull()->TEST_WALBufferIsEmpty());
// try the 2nd wal created during SwitchWAL
dbfull()->TEST_SwitchWAL();
ASSERT_TRUE(Put("key" + ToString(0), "value").ok());
ASSERT_TRUE(options.manual_wal_flush != dbfull()->TEST_WALBufferIsEmpty());
ASSERT_TRUE(dbfull()->FlushWAL(false).ok());
ASSERT_TRUE(dbfull()->TEST_WALBufferIsEmpty());
}
TEST_P(DBWriteTest, IOErrorOnWALWriteTriggersReadOnlyMode) {
std::unique_ptr<FaultInjectionTestEnv> mock_env(
new FaultInjectionTestEnv(Env::Default()));
Options options = GetOptions();
options.env = mock_env.get();
Reopen(options);
for (int i = 0; i < 2; i++) {
// Forcibly fail WAL write for the first Put only. Subsequent Puts should
// fail due to read-only mode
mock_env->SetFilesystemActive(i != 0);
auto res = Put("key" + ToString(i), "value");
// TSAN reports a false alarm for lock-order-inversion but Open and
// FlushWAL are not run concurrently. Disabling this until TSAN is
// fixed.
/*
if (options.manual_wal_flush && i == 0) {
// even with manual_wal_flush the 2nd Put should return error because of
// the read-only mode
ASSERT_TRUE(res.ok());
// we should see fs error when we do the flush
res = dbfull()->FlushWAL(false);
}
*/
if (!options.manual_wal_flush) {
ASSERT_FALSE(res.ok());
}
}
// Close before mock_env destruct.
Close();
}
TEST_P(DBWriteTest, IOErrorOnSwitchMemtable) {
Random rnd(301);
std::unique_ptr<FaultInjectionTestEnv> mock_env(
new FaultInjectionTestEnv(Env::Default()));
Options options = GetOptions();
options.env = mock_env.get();
options.writable_file_max_buffer_size = 4 * 1024 * 1024;
options.write_buffer_size = 3 * 512 * 1024;
options.wal_bytes_per_sync = 256 * 1024;
options.manual_wal_flush = true;
Reopen(options);
mock_env->SetFilesystemActive(false, Status::IOError("Not active"));
Status s;
for (int i = 0; i < 4 * 512; ++i) {
s = Put(Key(i), RandomString(&rnd, 1024));
if (!s.ok()) {
break;
}
}
ASSERT_EQ(s.severity(), Status::Severity::kFatalError);
mock_env->SetFilesystemActive(true);
// Close before mock_env destruct.
Close();
}
Expose DB methods to lock and unlock the WAL (#5146) Summary: Expose DB methods to lock and unlock the WAL. These methods are intended to use by MyRocks in order to obtain WAL coordinates in consistent way. Usage scenario is following: MySQL has performance_schema.log_status which provides information that enables a backup tool to copy the required log files without locking for the duration of copy. To populate this table MySQL does following: 1. Lock the binary log. Transactions are not allowed to commit now 2. Save the binary log coordinates 3. Walk through the storage engines and lock writes on each engine. For InnoDB, redo log is locked. For MyRocks, WAL should be locked. 4. Ask storage engines for their coordinates. InnoDB reports its current LSN and checkpoint LSN. MyRocks should report active WAL files names and sizes. 5. Release storage engine's locks 6. Unlock binary log Backup tool will then use this information to copy InnoDB, RocksDB and MySQL binary logs up to specified positions to end up with consistent DB state after restore. Currently, RocksDB allows to obtain the list of WAL files. Only missing bit is the method to lock the writes to WAL files. LockWAL method must flush the WAL in order for the reported size to be accurate (GetSortedWALFiles is using file system stat call to return the file size), also, since backup tool is going to copy the WAL, it is better to be flushed. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5146 Differential Revision: D14815447 Pulled By: maysamyabandeh fbshipit-source-id: eec9535a6025229ed471119f19fe7b3d8ae888a3
2019-04-06 15:36:42 +02:00
// Test that db->LockWAL() flushes the WAL after locking.
TEST_P(DBWriteTest, LockWalInEffect) {
Options options = GetOptions();
Reopen(options);
// try the 1st WAL created during open
ASSERT_OK(Put("key" + ToString(0), "value"));
ASSERT_TRUE(options.manual_wal_flush != dbfull()->TEST_WALBufferIsEmpty());
ASSERT_OK(dbfull()->LockWAL());
ASSERT_TRUE(dbfull()->TEST_WALBufferIsEmpty(false));
ASSERT_OK(dbfull()->UnlockWAL());
// try the 2nd wal created during SwitchWAL
dbfull()->TEST_SwitchWAL();
ASSERT_OK(Put("key" + ToString(0), "value"));
ASSERT_TRUE(options.manual_wal_flush != dbfull()->TEST_WALBufferIsEmpty());
ASSERT_OK(dbfull()->LockWAL());
ASSERT_TRUE(dbfull()->TEST_WALBufferIsEmpty(false));
ASSERT_OK(dbfull()->UnlockWAL());
}
INSTANTIATE_TEST_CASE_P(DBWriteTestInstance, DBWriteTest,
testing::Values(DBTestBase::kDefault,
Optimize for serial commits in 2PC Summary: Throughput: 46k tps in our sysbench settings (filling the details later) The idea is to have the simplest change that gives us a reasonable boost in 2PC throughput. Major design changes: 1. The WAL file internal buffer is not flushed after each write. Instead it is flushed before critical operations (WAL copy via fs) or when FlushWAL is called by MySQL. Flushing the WAL buffer is also protected via mutex_. 2. Use two sequence numbers: last seq, and last seq for write. Last seq is the last visible sequence number for reads. Last seq for write is the next sequence number that should be used to write to WAL/memtable. This allows to have a memtable write be in parallel to WAL writes. 3. BatchGroup is not used for writes. This means that we can have parallel writers which changes a major assumption in the code base. To accommodate for that i) allow only 1 WriteImpl that intends to write to memtable via mem_mutex_--which is fine since in 2PC almost all of the memtable writes come via group commit phase which is serial anyway, ii) make all the parts in the code base that assumed to be the only writer (via EnterUnbatched) to also acquire mem_mutex_, iii) stat updates are protected via a stat_mutex_. Note: the first commit has the approach figured out but is not clean. Submitting the PR anyway to get the early feedback on the approach. If we are ok with the approach I will go ahead with this updates: 0) Rebase with Yi's pipelining changes 1) Currently batching is disabled by default to make sure that it will be consistent with all unit tests. Will make this optional via a config. 2) A couple of unit tests are disabled. They need to be updated with the serial commit of 2PC taken into account. 3) Replacing BatchGroup with mem_mutex_ got a bit ugly as it requires releasing mutex_ beforehand (the same way EnterUnbatched does). This needs to be cleaned up. Closes https://github.com/facebook/rocksdb/pull/2345 Differential Revision: D5210732 Pulled By: maysamyabandeh fbshipit-source-id: 78653bd95a35cd1e831e555e0e57bdfd695355a4
2017-06-24 23:06:43 +02:00
DBTestBase::kConcurrentWALWrites,
DBTestBase::kPipelinedWrite));
} // namespace rocksdb
int main(int argc, char** argv) {
rocksdb::port::InstallStackTraceHandler();
::testing::InitGoogleTest(&argc, argv);
return RUN_ALL_TESTS();
}