You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
rocksdb/db/db_flush_test.cc

3085 lines
112 KiB

// 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).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include <atomic>
Memtable sampling for mempurge heuristic. (#8628) Summary: Changes the API of the MemPurge process: the `bool experimental_allow_mempurge` and `experimental_mempurge_policy` flags have been replaced by a `double experimental_mempurge_threshold` option. This change of API reflects another major change introduced in this PR: the MemPurgeDecider() function now works by sampling the memtables being flushed to estimate the overall amount of useful payload (payload minus the garbage), and then compare this useful payload estimate with the `double experimental_mempurge_threshold` value. Therefore, when the value of this flag is `0.0` (default value), mempurge is simply deactivated. On the other hand, a value of `DBL_MAX` would be equivalent to always going through a mempurge regardless of the garbage ratio estimate. At the moment, a `double experimental_mempurge_threshold` value else than 0.0 or `DBL_MAX` is opnly supported`with the `SkipList` memtable representation. Regarding the sampling, this PR includes the introduction of a `MemTable::UniqueRandomSample` function that collects (approximately) random entries from the memtable by using the new `SkipList::Iterator::RandomSeek()` under the hood, or by iterating through each memtable entry, depending on the target sample size and the total number of entries. The unit tests have been readapted to support this new API. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8628 Reviewed By: pdillinger Differential Revision: D30149315 Pulled By: bjlemaire fbshipit-source-id: 1feef5390c95db6f4480ab4434716533d3947f27
3 years ago
#include <limits>
#include "db/db_impl/db_impl.h"
#include "db/db_test_util.h"
#include "env/mock_env.h"
#include "file/filename.h"
#include "port/port.h"
#include "port/stack_trace.h"
#include "rocksdb/utilities/transaction_db.h"
#include "test_util/sync_point.h"
#include "test_util/testutil.h"
#include "util/cast_util.h"
#include "util/mutexlock.h"
#include "utilities/fault_injection_env.h"
#include "utilities/fault_injection_fs.h"
namespace ROCKSDB_NAMESPACE {
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// This is a static filter used for filtering
// kvs during the compaction process.
static std::string NEW_VALUE = "NewValue";
class DBFlushTest : public DBTestBase {
public:
DBFlushTest() : DBTestBase("db_flush_test", /*env_do_fsync=*/true) {}
};
class DBFlushDirectIOTest : public DBFlushTest,
public ::testing::WithParamInterface<bool> {
public:
DBFlushDirectIOTest() : DBFlushTest() {}
};
class DBAtomicFlushTest : public DBFlushTest,
public ::testing::WithParamInterface<bool> {
public:
DBAtomicFlushTest() : DBFlushTest() {}
};
// We had issue when two background threads trying to flush at the same time,
// only one of them get committed. The test verifies the issue is fixed.
TEST_F(DBFlushTest, FlushWhileWritingManifest) {
Options options;
options.disable_auto_compactions = true;
options.max_background_flushes = 2;
options.env = env_;
Reopen(options);
FlushOptions no_wait;
no_wait.wait = false;
no_wait.allow_write_stall=true;
SyncPoint::GetInstance()->LoadDependency(
{{"VersionSet::LogAndApply:WriteManifest",
"DBFlushTest::FlushWhileWritingManifest:1"},
{"MemTableList::TryInstallMemtableFlushResults:InProgress",
"VersionSet::LogAndApply:WriteManifestDone"}});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Put("foo", "v"));
ASSERT_OK(dbfull()->Flush(no_wait));
TEST_SYNC_POINT("DBFlushTest::FlushWhileWritingManifest:1");
ASSERT_OK(Put("bar", "v"));
ASSERT_OK(dbfull()->Flush(no_wait));
// If the issue is hit we will wait here forever.
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
#ifndef ROCKSDB_LITE
ASSERT_EQ(2, TotalTableFiles());
#endif // ROCKSDB_LITE
}
// Disable this test temporarily on Travis as it fails intermittently.
// Github issue: #4151
TEST_F(DBFlushTest, SyncFail) {
std::unique_ptr<FaultInjectionTestEnv> fault_injection_env(
new FaultInjectionTestEnv(env_));
Options options;
options.disable_auto_compactions = true;
options.env = fault_injection_env.get();
SyncPoint::GetInstance()->LoadDependency(
{{"DBFlushTest::SyncFail:1", "DBImpl::SyncClosedLogs:Start"},
{"DBImpl::SyncClosedLogs:Failed", "DBFlushTest::SyncFail:2"}});
SyncPoint::GetInstance()->EnableProcessing();
CreateAndReopenWithCF({"pikachu"}, options);
ASSERT_OK(Put("key", "value"));
FlushOptions flush_options;
flush_options.wait = false;
ASSERT_OK(dbfull()->Flush(flush_options));
Skip deleted WALs during recovery Summary: This patch record min log number to keep to the manifest while flushing SST files to ignore them and any WAL older than them during recovery. This is to avoid scenarios when we have a gap between the WAL files are fed to the recovery procedure. The gap could happen by for example out-of-order WAL deletion. Such gap could cause problems in 2PC recovery where the prepared and commit entry are placed into two separate WAL and gap in the WALs could result into not processing the WAL with the commit entry and hence breaking the 2PC recovery logic. Before the commit, for 2PC case, we determined which log number to keep in FindObsoleteFiles(). We looked at the earliest logs with outstanding prepare entries, or prepare entries whose respective commit or abort are in memtable. With the commit, the same calculation is done while we apply the SST flush. Just before installing the flush file, we precompute the earliest log file to keep after the flush finishes using the same logic (but skipping the memtables just flushed), record this information to the manifest entry for this new flushed SST file. This pre-computed value is also remembered in memory, and will later be used to determine whether a log file can be deleted. This value is unlikely to change until next flush because the commit entry will stay in memtable. (In WritePrepared, we could have removed the older log files as soon as all prepared entries are committed. It's not yet done anyway. Even if we do it, the only thing we loss with this new approach is earlier log deletion between two flushes, which does not guarantee to happen anyway because the obsolete file clean-up function is only executed after flush or compaction) This min log number to keep is stored in the manifest using the safely-ignore customized field of AddFile entry, in order to guarantee that the DB generated using newer release can be opened by previous releases no older than 4.2. Closes https://github.com/facebook/rocksdb/pull/3765 Differential Revision: D7747618 Pulled By: siying fbshipit-source-id: d00c92105b4f83852e9754a1b70d6b64cb590729
6 years ago
// Flush installs a new super-version. Get the ref count after that.
fault_injection_env->SetFilesystemActive(false);
TEST_SYNC_POINT("DBFlushTest::SyncFail:1");
TEST_SYNC_POINT("DBFlushTest::SyncFail:2");
fault_injection_env->SetFilesystemActive(true);
Skip deleted WALs during recovery Summary: This patch record min log number to keep to the manifest while flushing SST files to ignore them and any WAL older than them during recovery. This is to avoid scenarios when we have a gap between the WAL files are fed to the recovery procedure. The gap could happen by for example out-of-order WAL deletion. Such gap could cause problems in 2PC recovery where the prepared and commit entry are placed into two separate WAL and gap in the WALs could result into not processing the WAL with the commit entry and hence breaking the 2PC recovery logic. Before the commit, for 2PC case, we determined which log number to keep in FindObsoleteFiles(). We looked at the earliest logs with outstanding prepare entries, or prepare entries whose respective commit or abort are in memtable. With the commit, the same calculation is done while we apply the SST flush. Just before installing the flush file, we precompute the earliest log file to keep after the flush finishes using the same logic (but skipping the memtables just flushed), record this information to the manifest entry for this new flushed SST file. This pre-computed value is also remembered in memory, and will later be used to determine whether a log file can be deleted. This value is unlikely to change until next flush because the commit entry will stay in memtable. (In WritePrepared, we could have removed the older log files as soon as all prepared entries are committed. It's not yet done anyway. Even if we do it, the only thing we loss with this new approach is earlier log deletion between two flushes, which does not guarantee to happen anyway because the obsolete file clean-up function is only executed after flush or compaction) This min log number to keep is stored in the manifest using the safely-ignore customized field of AddFile entry, in order to guarantee that the DB generated using newer release can be opened by previous releases no older than 4.2. Closes https://github.com/facebook/rocksdb/pull/3765 Differential Revision: D7747618 Pulled By: siying fbshipit-source-id: d00c92105b4f83852e9754a1b70d6b64cb590729
6 years ago
// Now the background job will do the flush; wait for it.
// Returns the IO error happend during flush.
ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable());
#ifndef ROCKSDB_LITE
ASSERT_EQ("", FilesPerLevel()); // flush failed.
#endif // ROCKSDB_LITE
Destroy(options);
}
TEST_F(DBFlushTest, SyncSkip) {
Options options = CurrentOptions();
SyncPoint::GetInstance()->LoadDependency(
{{"DBFlushTest::SyncSkip:1", "DBImpl::SyncClosedLogs:Skip"},
{"DBImpl::SyncClosedLogs:Skip", "DBFlushTest::SyncSkip:2"}});
SyncPoint::GetInstance()->EnableProcessing();
Reopen(options);
ASSERT_OK(Put("key", "value"));
FlushOptions flush_options;
flush_options.wait = false;
ASSERT_OK(dbfull()->Flush(flush_options));
TEST_SYNC_POINT("DBFlushTest::SyncSkip:1");
TEST_SYNC_POINT("DBFlushTest::SyncSkip:2");
// Now the background job will do the flush; wait for it.
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
Destroy(options);
}
TEST_F(DBFlushTest, FlushInLowPriThreadPool) {
// Verify setting an empty high-pri (flush) thread pool causes flushes to be
// scheduled in the low-pri (compaction) thread pool.
Options options = CurrentOptions();
options.level0_file_num_compaction_trigger = 4;
options.memtable_factory.reset(test::NewSpecialSkipListFactory(1));
Reopen(options);
env_->SetBackgroundThreads(0, Env::HIGH);
std::thread::id tid;
int num_flushes = 0, num_compactions = 0;
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::BGWorkFlush", [&](void* /*arg*/) {
if (tid == std::thread::id()) {
tid = std::this_thread::get_id();
} else {
ASSERT_EQ(tid, std::this_thread::get_id());
}
++num_flushes;
});
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::BGWorkCompaction", [&](void* /*arg*/) {
ASSERT_EQ(tid, std::this_thread::get_id());
++num_compactions;
});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Put("key", "val"));
for (int i = 0; i < 4; ++i) {
ASSERT_OK(Put("key", "val"));
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
}
ASSERT_OK(dbfull()->TEST_WaitForCompact());
ASSERT_EQ(4, num_flushes);
ASSERT_EQ(1, num_compactions);
}
// Test when flush job is submitted to low priority thread pool and when DB is
// closed in the meanwhile, CloseHelper doesn't hang.
TEST_F(DBFlushTest, CloseDBWhenFlushInLowPri) {
Options options = CurrentOptions();
options.max_background_flushes = 1;
options.max_total_wal_size = 8192;
DestroyAndReopen(options);
CreateColumnFamilies({"cf1", "cf2"}, options);
env_->SetBackgroundThreads(0, Env::HIGH);
env_->SetBackgroundThreads(1, Env::LOW);
test::SleepingBackgroundTask sleeping_task_low;
int num_flushes = 0;
SyncPoint::GetInstance()->SetCallBack("DBImpl::BGWorkFlush",
[&](void* /*arg*/) { ++num_flushes; });
int num_low_flush_unscheduled = 0;
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::UnscheduleLowFlushCallback", [&](void* /*arg*/) {
num_low_flush_unscheduled++;
// There should be one flush job in low pool that needs to be
// unscheduled
ASSERT_EQ(num_low_flush_unscheduled, 1);
});
int num_high_flush_unscheduled = 0;
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::UnscheduleHighFlushCallback", [&](void* /*arg*/) {
num_high_flush_unscheduled++;
// There should be no flush job in high pool
ASSERT_EQ(num_high_flush_unscheduled, 0);
});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Put(0, "key1", DummyString(8192)));
// Block thread so that flush cannot be run and can be removed from the queue
// when called Unschedule.
env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask, &sleeping_task_low,
Env::Priority::LOW);
sleeping_task_low.WaitUntilSleeping();
// Trigger flush and flush job will be scheduled to LOW priority thread.
ASSERT_OK(Put(0, "key2", DummyString(8192)));
// Close DB and flush job in low priority queue will be removed without
// running.
Close();
sleeping_task_low.WakeUp();
sleeping_task_low.WaitUntilDone();
ASSERT_EQ(0, num_flushes);
TryReopenWithColumnFamilies({"default", "cf1", "cf2"}, options);
ASSERT_OK(Put(0, "key3", DummyString(8192)));
ASSERT_OK(Flush(0));
ASSERT_EQ(1, num_flushes);
}
TEST_F(DBFlushTest, ManualFlushWithMinWriteBufferNumberToMerge) {
Options options = CurrentOptions();
options.write_buffer_size = 100;
options.max_write_buffer_number = 4;
options.min_write_buffer_number_to_merge = 3;
Reopen(options);
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::BGWorkFlush",
"DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:1"},
{"DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:2",
"FlushJob::WriteLevel0Table"}});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Put("key1", "value1"));
port::Thread t([&]() {
// The call wait for flush to finish, i.e. with flush_options.wait = true.
ASSERT_OK(Flush());
});
// Wait for flush start.
TEST_SYNC_POINT("DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:1");
// Insert a second memtable before the manual flush finish.
// At the end of the manual flush job, it will check if further flush
// is needed, but it will not trigger flush of the second memtable because
// min_write_buffer_number_to_merge is not reached.
ASSERT_OK(Put("key2", "value2"));
ASSERT_OK(dbfull()->TEST_SwitchMemtable());
TEST_SYNC_POINT("DBFlushTest::ManualFlushWithMinWriteBufferNumberToMerge:2");
// Manual flush should return, without waiting for flush indefinitely.
t.join();
}
TEST_F(DBFlushTest, ScheduleOnlyOneBgThread) {
Options options = CurrentOptions();
Reopen(options);
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
int called = 0;
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::MaybeScheduleFlushOrCompaction:AfterSchedule:0", [&](void* arg) {
ASSERT_NE(nullptr, arg);
auto unscheduled_flushes = *reinterpret_cast<int*>(arg);
ASSERT_EQ(0, unscheduled_flushes);
++called;
});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Put("a", "foo"));
FlushOptions flush_opts;
ASSERT_OK(dbfull()->Flush(flush_opts));
ASSERT_EQ(1, called);
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
}
Added memtable garbage statistics (#8411) Summary: **Summary**: 2 new statistics counters are added to RocksDB: `MEMTABLE_PAYLOAD_BYTES_AT_FLUSH` and `MEMTABLE_GARBAGE_BYTES_AT_FLUSH`. The former tracks how many raw bytes of useful data are present on the memtable at flush time, whereas the latter is tracks how many of these raw bytes are considered garbage, meaning that they ended up not being imported on the SSTables resulting from the flush operations. **Unit test**: run `make db_flush_test -j$(nproc); ./db_flush_test` to run the unit test. This executable includes 3 tests, that test support and correct stat calculations for workloads with inserts, deletes, and DeleteRanges. The parameters are set such that the workloads are performed on a single memtable, and a single SSTable is created as a result of the flush operation. The flush operation is manually called in the test file. The tests verify that the values of these 2 statistics counters introduced in this PR can be exactly predicted, showing that we have a full understanding of the underlying operations. **Performance testing**: `./db_bench -statistics -benchmarks=fillrandom -num=10000000` repeated 10 times. Timing done using "date" function in a bash script. _Results_: Original Rocksdb fork: mean 66.6 sec, std 1.18 sec. This feature branch: mean 67.4 sec, std 1.35 sec. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8411 Reviewed By: akankshamahajan15 Differential Revision: D29150629 Pulled By: bjlemaire fbshipit-source-id: 7b3c2e86d50c6aa34fa50fd134282eacb543a5b1
3 years ago
// The following 3 tests are designed for testing garbage statistics at flush
// time.
//
// ======= General Information ======= (from GitHub Wiki).
// There are three scenarios where memtable flush can be triggered:
//
// 1 - Memtable size exceeds ColumnFamilyOptions::write_buffer_size
// after a write.
// 2 - Total memtable size across all column families exceeds
// DBOptions::db_write_buffer_size,
// or DBOptions::write_buffer_manager signals a flush. In this scenario
// the largest memtable will be flushed.
// 3 - Total WAL file size exceeds DBOptions::max_total_wal_size.
// In this scenario the memtable with the oldest data will be flushed,
// in order to allow the WAL file with data from this memtable to be
// purged.
//
// As a result, a memtable can be flushed before it is full. This is one
// reason the generated SST file can be smaller than the corresponding
// memtable. Compression is another factor to make SST file smaller than
// corresponding memtable, since data in memtable is uncompressed.
TEST_F(DBFlushTest, StatisticsGarbageBasic) {
Options options = CurrentOptions();
// The following options are used to enforce several values that
// may already exist as default values to make this test resilient
// to default value updates in the future.
options.statistics = CreateDBStatistics();
// Record all statistics.
options.statistics->set_stats_level(StatsLevel::kAll);
// create the DB if it's not already present
options.create_if_missing = true;
// Useful for now as we are trying to compare uncompressed data savings on
// flush().
options.compression = kNoCompression;
// Prevent memtable in place updates. Should already be disabled
// (from Wiki:
// In place updates can be enabled by toggling on the bool
// inplace_update_support flag. However, this flag is by default set to
// false
// because this thread-safe in-place update support is not compatible
// with concurrent memtable writes. Note that the bool
// allow_concurrent_memtable_write is set to true by default )
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
// Enforce size of a single MemTable to 64MB (64MB = 67108864 bytes).
options.write_buffer_size = 64 << 20;
ASSERT_OK(TryReopen(options));
// Put multiple times the same key-values.
// The encoded length of a db entry in the memtable is
// defined in db/memtable.cc (MemTable::Add) as the variable:
// encoded_len= VarintLength(internal_key_size) --> =
// log_256(internal_key).
// Min # of bytes
// necessary to
// store
// internal_key_size.
// + internal_key_size --> = actual key string,
// (size key_size: w/o term null char)
// + 8 bytes for
// fixed uint64 "seq
// number
// +
// insertion type"
// + VarintLength(val_size) --> = min # of bytes to
// store val_size
// + val_size --> = actual value
// string
// For example, in our situation, "key1" : size 4, "value1" : size 6
// (the terminating null characters are not copied over to the memtable).
// And therefore encoded_len = 1 + (4+8) + 1 + 6 = 20 bytes per entry.
// However in terms of raw data contained in the memtable, and written
// over to the SSTable, we only count internal_key_size and val_size,
// because this is the only raw chunk of bytes that contains everything
// necessary to reconstruct a user entry: sequence number, insertion type,
// key, and value.
// To test the relevance of our Memtable garbage statistics,
// namely MEMTABLE_PAYLOAD_BYTES_AT_FLUSH and MEMTABLE_GARBAGE_BYTES_AT_FLUSH,
// we insert K-V pairs with 3 distinct keys (of length 4),
// and random values of arbitrary length RAND_VALUES_LENGTH,
// and we repeat this step NUM_REPEAT times total.
// At the end, we insert 3 final K-V pairs with the same 3 keys
// and known values (these will be the final values, of length 6).
// I chose NUM_REPEAT=2,000 such that no automatic flush is
// triggered (the number of bytes in the memtable is therefore
// well below any meaningful heuristic for a memtable of size 64MB).
// As a result, since each K-V pair is inserted as a payload
// of N meaningful bytes (sequence number, insertion type,
// key, and value = 8 + 4 + RAND_VALUE_LENGTH),
// MEMTABLE_GARBAGE_BYTES_AT_FLUSH should be equal to 2,000 * N bytes
// and MEMTABLE_PAYLAOD_BYTES_AT_FLUSH = MEMTABLE_GARBAGE_BYTES_AT_FLUSH +
// (3*(8 + 4 + 6)) bytes. For RAND_VALUE_LENGTH = 172 (arbitrary value), we
// expect:
// N = 8 + 4 + 172 = 184 bytes
// MEMTABLE_GARBAGE_BYTES_AT_FLUSH = 2,000 * 184 = 368,000 bytes.
// MEMTABLE_PAYLOAD_BYTES_AT_FLUSH = 368,000 + 3*18 = 368,054 bytes.
const size_t NUM_REPEAT = 2000;
const size_t RAND_VALUES_LENGTH = 172;
const std::string KEY1 = "key1";
const std::string KEY2 = "key2";
const std::string KEY3 = "key3";
const std::string VALUE1 = "value1";
const std::string VALUE2 = "value2";
const std::string VALUE3 = "value3";
uint64_t EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH = 0;
uint64_t EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH = 0;
Random rnd(301);
// Insertion of of K-V pairs, multiple times.
for (size_t i = 0; i < NUM_REPEAT; i++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
std::string p_v1 = rnd.RandomString(RAND_VALUES_LENGTH);
std::string p_v2 = rnd.RandomString(RAND_VALUES_LENGTH);
std::string p_v3 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEY1, p_v1));
ASSERT_OK(Put(KEY2, p_v2));
ASSERT_OK(Put(KEY3, p_v3));
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY1.size() + p_v1.size() + sizeof(uint64_t);
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY2.size() + p_v2.size() + sizeof(uint64_t);
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY3.size() + p_v3.size() + sizeof(uint64_t);
}
// The memtable data bytes includes the "garbage"
// bytes along with the useful payload.
EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH =
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH;
ASSERT_OK(Put(KEY1, VALUE1));
ASSERT_OK(Put(KEY2, VALUE2));
ASSERT_OK(Put(KEY3, VALUE3));
// Add useful payload to the memtable data bytes:
EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH +=
KEY1.size() + VALUE1.size() + KEY2.size() + VALUE2.size() + KEY3.size() +
VALUE3.size() + 3 * sizeof(uint64_t);
// We assert that the last K-V pairs have been successfully inserted,
// and that the valid values are VALUE1, VALUE2, VALUE3.
PinnableSlice value;
ASSERT_OK(Get(KEY1, &value));
ASSERT_EQ(value.ToString(), VALUE1);
ASSERT_OK(Get(KEY2, &value));
ASSERT_EQ(value.ToString(), VALUE2);
ASSERT_OK(Get(KEY3, &value));
ASSERT_EQ(value.ToString(), VALUE3);
// Force flush to SST. Increments the statistics counter.
ASSERT_OK(Flush());
// Collect statistics.
uint64_t mem_data_bytes =
TestGetTickerCount(options, MEMTABLE_PAYLOAD_BYTES_AT_FLUSH);
uint64_t mem_garbage_bytes =
TestGetTickerCount(options, MEMTABLE_GARBAGE_BYTES_AT_FLUSH);
EXPECT_EQ(mem_data_bytes, EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH);
EXPECT_EQ(mem_garbage_bytes, EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH);
Close();
}
TEST_F(DBFlushTest, StatisticsGarbageInsertAndDeletes) {
Options options = CurrentOptions();
options.statistics = CreateDBStatistics();
options.statistics->set_stats_level(StatsLevel::kAll);
options.create_if_missing = true;
options.compression = kNoCompression;
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
options.write_buffer_size = 67108864;
ASSERT_OK(TryReopen(options));
const size_t NUM_REPEAT = 2000;
const size_t RAND_VALUES_LENGTH = 37;
const std::string KEY1 = "key1";
const std::string KEY2 = "key2";
const std::string KEY3 = "key3";
const std::string KEY4 = "key4";
const std::string KEY5 = "key5";
const std::string KEY6 = "key6";
uint64_t EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH = 0;
uint64_t EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH = 0;
WriteBatch batch;
Random rnd(301);
// Insertion of of K-V pairs, multiple times.
for (size_t i = 0; i < NUM_REPEAT; i++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
std::string p_v1 = rnd.RandomString(RAND_VALUES_LENGTH);
std::string p_v2 = rnd.RandomString(RAND_VALUES_LENGTH);
std::string p_v3 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEY1, p_v1));
ASSERT_OK(Put(KEY2, p_v2));
ASSERT_OK(Put(KEY3, p_v3));
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY1.size() + p_v1.size() + sizeof(uint64_t);
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY2.size() + p_v2.size() + sizeof(uint64_t);
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY3.size() + p_v3.size() + sizeof(uint64_t);
ASSERT_OK(Delete(KEY1));
ASSERT_OK(Delete(KEY2));
ASSERT_OK(Delete(KEY3));
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY1.size() + KEY2.size() + KEY3.size() + 3 * sizeof(uint64_t);
}
// The memtable data bytes includes the "garbage"
// bytes along with the useful payload.
EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH =
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH;
// Note : one set of delete for KEY1, KEY2, KEY3 is written to
// SSTable to propagate the delete operations to K-V pairs
// that could have been inserted into the database during past Flush
// opeartions.
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH -=
KEY1.size() + KEY2.size() + KEY3.size() + 3 * sizeof(uint64_t);
// Additional useful paylaod.
ASSERT_OK(Delete(KEY4));
ASSERT_OK(Delete(KEY5));
ASSERT_OK(Delete(KEY6));
// // Add useful payload to the memtable data bytes:
EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH +=
KEY4.size() + KEY5.size() + KEY6.size() + 3 * sizeof(uint64_t);
// We assert that the K-V pairs have been successfully deleted.
PinnableSlice value;
ASSERT_NOK(Get(KEY1, &value));
ASSERT_NOK(Get(KEY2, &value));
ASSERT_NOK(Get(KEY3, &value));
// Force flush to SST. Increments the statistics counter.
ASSERT_OK(Flush());
// Collect statistics.
uint64_t mem_data_bytes =
TestGetTickerCount(options, MEMTABLE_PAYLOAD_BYTES_AT_FLUSH);
uint64_t mem_garbage_bytes =
TestGetTickerCount(options, MEMTABLE_GARBAGE_BYTES_AT_FLUSH);
EXPECT_EQ(mem_data_bytes, EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH);
EXPECT_EQ(mem_garbage_bytes, EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH);
Close();
}
TEST_F(DBFlushTest, StatisticsGarbageRangeDeletes) {
Options options = CurrentOptions();
options.statistics = CreateDBStatistics();
options.statistics->set_stats_level(StatsLevel::kAll);
options.create_if_missing = true;
options.compression = kNoCompression;
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
options.write_buffer_size = 67108864;
ASSERT_OK(TryReopen(options));
const size_t NUM_REPEAT = 1000;
const size_t RAND_VALUES_LENGTH = 42;
const std::string KEY1 = "key1";
const std::string KEY2 = "key2";
const std::string KEY3 = "key3";
const std::string KEY4 = "key4";
const std::string KEY5 = "key5";
const std::string KEY6 = "key6";
const std::string VALUE3 = "value3";
uint64_t EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH = 0;
uint64_t EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH = 0;
Random rnd(301);
// Insertion of of K-V pairs, multiple times.
// Also insert DeleteRange
for (size_t i = 0; i < NUM_REPEAT; i++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
std::string p_v1 = rnd.RandomString(RAND_VALUES_LENGTH);
std::string p_v2 = rnd.RandomString(RAND_VALUES_LENGTH);
std::string p_v3 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEY1, p_v1));
ASSERT_OK(Put(KEY2, p_v2));
ASSERT_OK(Put(KEY3, p_v3));
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY1.size() + p_v1.size() + sizeof(uint64_t);
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY2.size() + p_v2.size() + sizeof(uint64_t);
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
KEY3.size() + p_v3.size() + sizeof(uint64_t);
ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(), KEY1,
KEY2));
// Note: DeleteRange have an exclusive upper bound, e.g. here: [KEY2,KEY3)
// is deleted.
ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(), KEY2,
KEY3));
// Delete ranges are stored as a regular K-V pair, with key=STARTKEY,
// value=ENDKEY.
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH +=
(KEY1.size() + KEY2.size() + sizeof(uint64_t)) +
(KEY2.size() + KEY3.size() + sizeof(uint64_t));
}
// The memtable data bytes includes the "garbage"
// bytes along with the useful payload.
EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH =
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH;
// Note : one set of deleteRange for (KEY1, KEY2) and (KEY2, KEY3) is written
// to SSTable to propagate the deleteRange operations to K-V pairs that could
// have been inserted into the database during past Flush opeartions.
EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH -=
(KEY1.size() + KEY2.size() + sizeof(uint64_t)) +
(KEY2.size() + KEY3.size() + sizeof(uint64_t));
// Overwrite KEY3 with known value (VALUE3)
// Note that during the whole time KEY3 has never been deleted
// by the RangeDeletes.
ASSERT_OK(Put(KEY3, VALUE3));
EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH +=
KEY3.size() + VALUE3.size() + sizeof(uint64_t);
// Additional useful paylaod.
ASSERT_OK(
db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(), KEY4, KEY5));
ASSERT_OK(
db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(), KEY5, KEY6));
// Add useful payload to the memtable data bytes:
EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH +=
(KEY4.size() + KEY5.size() + sizeof(uint64_t)) +
(KEY5.size() + KEY6.size() + sizeof(uint64_t));
// We assert that the K-V pairs have been successfully deleted.
PinnableSlice value;
ASSERT_NOK(Get(KEY1, &value));
ASSERT_NOK(Get(KEY2, &value));
// And that KEY3's value is correct.
ASSERT_OK(Get(KEY3, &value));
ASSERT_EQ(value, VALUE3);
// Force flush to SST. Increments the statistics counter.
ASSERT_OK(Flush());
// Collect statistics.
uint64_t mem_data_bytes =
TestGetTickerCount(options, MEMTABLE_PAYLOAD_BYTES_AT_FLUSH);
uint64_t mem_garbage_bytes =
TestGetTickerCount(options, MEMTABLE_GARBAGE_BYTES_AT_FLUSH);
EXPECT_EQ(mem_data_bytes, EXPECTED_MEMTABLE_PAYLOAD_BYTES_AT_FLUSH);
EXPECT_EQ(mem_garbage_bytes, EXPECTED_MEMTABLE_GARBAGE_BYTES_AT_FLUSH);
Close();
}
#ifndef ROCKSDB_LITE
// This simple Listener can only handle one flush at a time.
class TestFlushListener : public EventListener {
public:
TestFlushListener(Env* env, DBFlushTest* test)
: slowdown_count(0), stop_count(0), db_closed(), env_(env), test_(test) {
db_closed = false;
}
~TestFlushListener() override {
prev_fc_info_.status.PermitUncheckedError(); // Ignore the status
}
void OnTableFileCreated(const TableFileCreationInfo& info) override {
// remember the info for later checking the FlushJobInfo.
prev_fc_info_ = info;
ASSERT_GT(info.db_name.size(), 0U);
ASSERT_GT(info.cf_name.size(), 0U);
ASSERT_GT(info.file_path.size(), 0U);
ASSERT_GT(info.job_id, 0);
ASSERT_GT(info.table_properties.data_size, 0U);
ASSERT_GT(info.table_properties.raw_key_size, 0U);
ASSERT_GT(info.table_properties.raw_value_size, 0U);
ASSERT_GT(info.table_properties.num_data_blocks, 0U);
ASSERT_GT(info.table_properties.num_entries, 0U);
ASSERT_EQ(info.file_checksum, kUnknownFileChecksum);
ASSERT_EQ(info.file_checksum_func_name, kUnknownFileChecksumFuncName);
}
void OnFlushCompleted(DB* db, const FlushJobInfo& info) override {
flushed_dbs_.push_back(db);
flushed_column_family_names_.push_back(info.cf_name);
if (info.triggered_writes_slowdown) {
slowdown_count++;
}
if (info.triggered_writes_stop) {
stop_count++;
}
// verify whether the previously created file matches the flushed file.
ASSERT_EQ(prev_fc_info_.db_name, db->GetName());
ASSERT_EQ(prev_fc_info_.cf_name, info.cf_name);
ASSERT_EQ(prev_fc_info_.job_id, info.job_id);
ASSERT_EQ(prev_fc_info_.file_path, info.file_path);
ASSERT_EQ(TableFileNameToNumber(info.file_path), info.file_number);
// Note: the following chunk relies on the notification pertaining to the
// database pointed to by DBTestBase::db_, and is thus bypassed when
// that assumption does not hold (see the test case MultiDBMultiListeners
// below).
ASSERT_TRUE(test_);
if (db == test_->db_) {
std::vector<std::vector<FileMetaData>> files_by_level;
test_->dbfull()->TEST_GetFilesMetaData(db->DefaultColumnFamily(),
&files_by_level);
ASSERT_FALSE(files_by_level.empty());
auto it = std::find_if(files_by_level[0].begin(), files_by_level[0].end(),
[&](const FileMetaData& meta) {
return meta.fd.GetNumber() == info.file_number;
});
ASSERT_NE(it, files_by_level[0].end());
ASSERT_EQ(info.oldest_blob_file_number, it->oldest_blob_file_number);
}
ASSERT_EQ(db->GetEnv()->GetThreadID(), info.thread_id);
ASSERT_GT(info.thread_id, 0U);
}
std::vector<std::string> flushed_column_family_names_;
std::vector<DB*> flushed_dbs_;
int slowdown_count;
int stop_count;
bool db_closing;
std::atomic_bool db_closed;
TableFileCreationInfo prev_fc_info_;
protected:
Env* env_;
DBFlushTest* test_;
};
#endif // !ROCKSDB_LITE
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
TEST_F(DBFlushTest, MemPurgeBasic) {
Options options = CurrentOptions();
// The following options are used to enforce several values that
// may already exist as default values to make this test resilient
// to default value updates in the future.
options.statistics = CreateDBStatistics();
// Record all statistics.
options.statistics->set_stats_level(StatsLevel::kAll);
// create the DB if it's not already present
options.create_if_missing = true;
// Useful for now as we are trying to compare uncompressed data savings on
// flush().
options.compression = kNoCompression;
// Prevent memtable in place updates. Should already be disabled
// (from Wiki:
// In place updates can be enabled by toggling on the bool
// inplace_update_support flag. However, this flag is by default set to
// false
// because this thread-safe in-place update support is not compatible
// with concurrent memtable writes. Note that the bool
// allow_concurrent_memtable_write is set to true by default )
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
// Enforce size of a single MemTable to 64MB (64MB = 67108864 bytes).
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
options.write_buffer_size = 1 << 20;
#ifndef ROCKSDB_LITE
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2 years ago
// Initially deactivate the MemPurge prototype.
options.experimental_mempurge_threshold = 0.0;
TestFlushListener* listener = new TestFlushListener(options.env, this);
options.listeners.emplace_back(listener);
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2 years ago
#else
// Activate directly the MemPurge prototype.
// (RocksDB lite does not support dynamic options)
options.experimental_mempurge_threshold = 1.0;
#endif // !ROCKSDB_LITE
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ASSERT_OK(TryReopen(options));
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2 years ago
// RocksDB lite does not support dynamic options
#ifndef ROCKSDB_LITE
// Dynamically activate the MemPurge prototype without restarting the DB.
ColumnFamilyHandle* cfh = db_->DefaultColumnFamily();
ASSERT_OK(db_->SetOptions(cfh, {{"experimental_mempurge_threshold", "1.0"}}));
#endif
std::atomic<uint32_t> mempurge_count{0};
std::atomic<uint32_t> sst_count{0};
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
"DBImpl::FlushJob:MemPurgeSuccessful",
[&](void* /*arg*/) { mempurge_count++; });
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
"DBImpl::FlushJob:SSTFileCreated", [&](void* /*arg*/) { sst_count++; });
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
std::string KEY1 = "IamKey1";
std::string KEY2 = "IamKey2";
std::string KEY3 = "IamKey3";
std::string KEY4 = "IamKey4";
std::string KEY5 = "IamKey5";
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
std::string KEY6 = "IamKey6";
std::string KEY7 = "IamKey7";
std::string KEY8 = "IamKey8";
std::string KEY9 = "IamKey9";
std::string RNDKEY1, RNDKEY2, RNDKEY3;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
const std::string NOT_FOUND = "NOT_FOUND";
// Heavy overwrite workload,
// more than would fit in maximum allowed memtables.
Random rnd(719);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
const size_t NUM_REPEAT = 100;
const size_t RAND_KEYS_LENGTH = 57;
const size_t RAND_VALUES_LENGTH = 10240;
std::string p_v1, p_v2, p_v3, p_v4, p_v5, p_v6, p_v7, p_v8, p_v9, p_rv1,
p_rv2, p_rv3;
// Insert a very first set of keys that will be
// mempurged at least once.
p_v1 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v2 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v3 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v4 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEY1, p_v1));
ASSERT_OK(Put(KEY2, p_v2));
ASSERT_OK(Put(KEY3, p_v3));
ASSERT_OK(Put(KEY4, p_v4));
ASSERT_EQ(Get(KEY1), p_v1);
ASSERT_EQ(Get(KEY2), p_v2);
ASSERT_EQ(Get(KEY3), p_v3);
ASSERT_EQ(Get(KEY4), p_v4);
// Insertion of of K-V pairs, multiple times (overwrites).
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
for (size_t i = 0; i < NUM_REPEAT; i++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
p_v5 = rnd.RandomString(RAND_VALUES_LENGTH);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
p_v6 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v7 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v8 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v9 = rnd.RandomString(RAND_VALUES_LENGTH);
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ASSERT_OK(Put(KEY5, p_v5));
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
ASSERT_OK(Put(KEY6, p_v6));
ASSERT_OK(Put(KEY7, p_v7));
ASSERT_OK(Put(KEY8, p_v8));
ASSERT_OK(Put(KEY9, p_v9));
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ASSERT_EQ(Get(KEY1), p_v1);
ASSERT_EQ(Get(KEY2), p_v2);
ASSERT_EQ(Get(KEY3), p_v3);
ASSERT_EQ(Get(KEY4), p_v4);
ASSERT_EQ(Get(KEY5), p_v5);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
ASSERT_EQ(Get(KEY6), p_v6);
ASSERT_EQ(Get(KEY7), p_v7);
ASSERT_EQ(Get(KEY8), p_v8);
ASSERT_EQ(Get(KEY9), p_v9);
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
}
// Check that there was at least one mempurge
const uint32_t EXPECTED_MIN_MEMPURGE_COUNT = 1;
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// Check that there was no SST files created during flush.
const uint32_t EXPECTED_SST_COUNT = 0;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
EXPECT_GE(mempurge_count.exchange(0), EXPECTED_MIN_MEMPURGE_COUNT);
EXPECT_EQ(sst_count.exchange(0), EXPECTED_SST_COUNT);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// Insertion of of K-V pairs, no overwrites.
for (size_t i = 0; i < NUM_REPEAT; i++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
RNDKEY1 = rnd.RandomString(RAND_KEYS_LENGTH);
RNDKEY2 = rnd.RandomString(RAND_KEYS_LENGTH);
RNDKEY3 = rnd.RandomString(RAND_KEYS_LENGTH);
p_rv1 = rnd.RandomString(RAND_VALUES_LENGTH);
p_rv2 = rnd.RandomString(RAND_VALUES_LENGTH);
p_rv3 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(RNDKEY1, p_rv1));
ASSERT_OK(Put(RNDKEY2, p_rv2));
ASSERT_OK(Put(RNDKEY3, p_rv3));
ASSERT_EQ(Get(KEY1), p_v1);
ASSERT_EQ(Get(KEY2), p_v2);
ASSERT_EQ(Get(KEY3), p_v3);
ASSERT_EQ(Get(KEY4), p_v4);
ASSERT_EQ(Get(KEY5), p_v5);
ASSERT_EQ(Get(KEY6), p_v6);
ASSERT_EQ(Get(KEY7), p_v7);
ASSERT_EQ(Get(KEY8), p_v8);
ASSERT_EQ(Get(KEY9), p_v9);
ASSERT_EQ(Get(RNDKEY1), p_rv1);
ASSERT_EQ(Get(RNDKEY2), p_rv2);
ASSERT_EQ(Get(RNDKEY3), p_rv3);
}
// Assert that at least one flush to storage has been performed
EXPECT_GT(sst_count.exchange(0), EXPECTED_SST_COUNT);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// (which will consequently increase the number of mempurges recorded too).
EXPECT_GE(mempurge_count.exchange(0), EXPECTED_MIN_MEMPURGE_COUNT);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// Assert that there is no data corruption, even with
// a flush to storage.
ASSERT_EQ(Get(KEY1), p_v1);
ASSERT_EQ(Get(KEY2), p_v2);
ASSERT_EQ(Get(KEY3), p_v3);
ASSERT_EQ(Get(KEY4), p_v4);
ASSERT_EQ(Get(KEY5), p_v5);
ASSERT_EQ(Get(KEY6), p_v6);
ASSERT_EQ(Get(KEY7), p_v7);
ASSERT_EQ(Get(KEY8), p_v8);
ASSERT_EQ(Get(KEY9), p_v9);
ASSERT_EQ(Get(RNDKEY1), p_rv1);
ASSERT_EQ(Get(RNDKEY2), p_rv2);
ASSERT_EQ(Get(RNDKEY3), p_rv3);
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
Close();
}
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2 years ago
// RocksDB lite does not support dynamic options
#ifndef ROCKSDB_LITE
TEST_F(DBFlushTest, MemPurgeBasicToggle) {
Options options = CurrentOptions();
// The following options are used to enforce several values that
// may already exist as default values to make this test resilient
// to default value updates in the future.
options.statistics = CreateDBStatistics();
// Record all statistics.
options.statistics->set_stats_level(StatsLevel::kAll);
// create the DB if it's not already present
options.create_if_missing = true;
// Useful for now as we are trying to compare uncompressed data savings on
// flush().
options.compression = kNoCompression;
// Prevent memtable in place updates. Should already be disabled
// (from Wiki:
// In place updates can be enabled by toggling on the bool
// inplace_update_support flag. However, this flag is by default set to
// false
// because this thread-safe in-place update support is not compatible
// with concurrent memtable writes. Note that the bool
// allow_concurrent_memtable_write is set to true by default )
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
// Enforce size of a single MemTable to 64MB (64MB = 67108864 bytes).
options.write_buffer_size = 1 << 20;
// Initially deactivate the MemPurge prototype.
// (negative values are equivalent to 0.0).
options.experimental_mempurge_threshold = -25.3;
TestFlushListener* listener = new TestFlushListener(options.env, this);
options.listeners.emplace_back(listener);
ASSERT_OK(TryReopen(options));
// Dynamically activate the MemPurge prototype without restarting the DB.
ColumnFamilyHandle* cfh = db_->DefaultColumnFamily();
// Values greater than 1.0 are equivalent to 1.0
ASSERT_OK(
db_->SetOptions(cfh, {{"experimental_mempurge_threshold", "3.7898"}}));
std::atomic<uint32_t> mempurge_count{0};
std::atomic<uint32_t> sst_count{0};
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:MemPurgeSuccessful",
[&](void* /*arg*/) { mempurge_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:SSTFileCreated", [&](void* /*arg*/) { sst_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
const size_t KVSIZE = 3;
std::vector<std::string> KEYS(KVSIZE);
for (size_t k = 0; k < KVSIZE; k++) {
KEYS[k] = "IamKey" + std::to_string(k);
}
std::vector<std::string> RNDVALS(KVSIZE);
const std::string NOT_FOUND = "NOT_FOUND";
// Heavy overwrite workload,
// more than would fit in maximum allowed memtables.
Random rnd(719);
const size_t NUM_REPEAT = 100;
const size_t RAND_VALUES_LENGTH = 10240;
// Insertion of of K-V pairs, multiple times (overwrites).
for (size_t i = 0; i < NUM_REPEAT; i++) {
for (size_t j = 0; j < KEYS.size(); j++) {
RNDVALS[j] = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEYS[j], RNDVALS[j]));
ASSERT_EQ(Get(KEYS[j]), RNDVALS[j]);
}
for (size_t j = 0; j < KEYS.size(); j++) {
ASSERT_EQ(Get(KEYS[j]), RNDVALS[j]);
}
}
// Check that there was at least one mempurge
const uint32_t EXPECTED_MIN_MEMPURGE_COUNT = 1;
// Check that there was no SST files created during flush.
const uint32_t EXPECTED_SST_COUNT = 0;
EXPECT_GE(mempurge_count.exchange(0), EXPECTED_MIN_MEMPURGE_COUNT);
EXPECT_EQ(sst_count.exchange(0), EXPECTED_SST_COUNT);
// Dynamically deactivate MemPurge.
ASSERT_OK(
db_->SetOptions(cfh, {{"experimental_mempurge_threshold", "-1023.0"}}));
// Insertion of of K-V pairs, multiple times (overwrites).
for (size_t i = 0; i < NUM_REPEAT; i++) {
for (size_t j = 0; j < KEYS.size(); j++) {
RNDVALS[j] = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEYS[j], RNDVALS[j]));
ASSERT_EQ(Get(KEYS[j]), RNDVALS[j]);
}
for (size_t j = 0; j < KEYS.size(); j++) {
ASSERT_EQ(Get(KEYS[j]), RNDVALS[j]);
}
}
// Check that there was at least one mempurge
const uint32_t ZERO = 0;
// Assert that at least one flush to storage has been performed
EXPECT_GT(sst_count.exchange(0), EXPECTED_SST_COUNT);
// The mempurge count is expected to be set to 0 when the options are updated.
// We expect no mempurge at all.
EXPECT_EQ(mempurge_count.exchange(0), ZERO);
Close();
}
// Closes the "#ifndef ROCKSDB_LITE"
// End of MemPurgeBasicToggle, which is not
// supported with RocksDB LITE because it
// relies on dynamically changing the option
// flag experimental_mempurge_threshold.
#endif
// At the moment, MemPurge feature is deactivated
// when atomic_flush is enabled. This is because the level
// of garbage between Column Families is not guaranteed to
// be consistent, therefore a CF could hypothetically
// trigger a MemPurge while another CF would trigger
// a regular Flush.
TEST_F(DBFlushTest, MemPurgeWithAtomicFlush) {
Options options = CurrentOptions();
// The following options are used to enforce several values that
// may already exist as default values to make this test resilient
// to default value updates in the future.
options.statistics = CreateDBStatistics();
// Record all statistics.
options.statistics->set_stats_level(StatsLevel::kAll);
// create the DB if it's not already present
options.create_if_missing = true;
// Useful for now as we are trying to compare uncompressed data savings on
// flush().
options.compression = kNoCompression;
// Prevent memtable in place updates. Should already be disabled
// (from Wiki:
// In place updates can be enabled by toggling on the bool
// inplace_update_support flag. However, this flag is by default set to
// false
// because this thread-safe in-place update support is not compatible
// with concurrent memtable writes. Note that the bool
// allow_concurrent_memtable_write is set to true by default )
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
// Enforce size of a single MemTable to 64KB (64KB = 65,536 bytes).
options.write_buffer_size = 1 << 20;
// Activate the MemPurge prototype.
options.experimental_mempurge_threshold = 153.245;
// Activate atomic_flush.
options.atomic_flush = true;
const std::vector<std::string> new_cf_names = {"pikachu", "eevie"};
CreateColumnFamilies(new_cf_names, options);
Close();
// 3 CFs: default will be filled with overwrites (would normally trigger
// mempurge)
// new_cf_names[1] will be filled with random values (would trigger
// flush) new_cf_names[2] not filled with anything.
ReopenWithColumnFamilies(
{kDefaultColumnFamilyName, new_cf_names[0], new_cf_names[1]}, options);
size_t num_cfs = handles_.size();
ASSERT_EQ(3, num_cfs);
ASSERT_OK(Put(1, "foo", "bar"));
ASSERT_OK(Put(2, "bar", "baz"));
std::atomic<uint32_t> mempurge_count{0};
std::atomic<uint32_t> sst_count{0};
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:MemPurgeSuccessful",
[&](void* /*arg*/) { mempurge_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:SSTFileCreated", [&](void* /*arg*/) { sst_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
const size_t KVSIZE = 3;
std::vector<std::string> KEYS(KVSIZE);
for (size_t k = 0; k < KVSIZE; k++) {
KEYS[k] = "IamKey" + std::to_string(k);
}
std::string RNDKEY;
std::vector<std::string> RNDVALS(KVSIZE);
const std::string NOT_FOUND = "NOT_FOUND";
// Heavy overwrite workload,
// more than would fit in maximum allowed memtables.
Random rnd(106);
const size_t NUM_REPEAT = 100;
const size_t RAND_KEY_LENGTH = 128;
const size_t RAND_VALUES_LENGTH = 10240;
// Insertion of of K-V pairs, multiple times (overwrites).
for (size_t i = 0; i < NUM_REPEAT; i++) {
for (size_t j = 0; j < KEYS.size(); j++) {
RNDKEY = rnd.RandomString(RAND_KEY_LENGTH);
RNDVALS[j] = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEYS[j], RNDVALS[j]));
ASSERT_OK(Put(1, RNDKEY, RNDVALS[j]));
ASSERT_EQ(Get(KEYS[j]), RNDVALS[j]);
ASSERT_EQ(Get(1, RNDKEY), RNDVALS[j]);
}
}
// Check that there was no mempurge because atomic_flush option is true.
const uint32_t EXPECTED_MIN_MEMPURGE_COUNT = 0;
// Check that there was at least one SST files created during flush.
const uint32_t EXPECTED_SST_COUNT = 1;
EXPECT_EQ(mempurge_count.exchange(0), EXPECTED_MIN_MEMPURGE_COUNT);
EXPECT_GE(sst_count.exchange(0), EXPECTED_SST_COUNT);
Close();
}
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
TEST_F(DBFlushTest, MemPurgeDeleteAndDeleteRange) {
Options options = CurrentOptions();
options.statistics = CreateDBStatistics();
options.statistics->set_stats_level(StatsLevel::kAll);
options.create_if_missing = true;
options.compression = kNoCompression;
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
#ifndef ROCKSDB_LITE
TestFlushListener* listener = new TestFlushListener(options.env, this);
options.listeners.emplace_back(listener);
#endif // !ROCKSDB_LITE
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Enforce size of a single MemTable to 64MB (64MB = 67108864 bytes).
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
options.write_buffer_size = 1 << 20;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Activate the MemPurge prototype.
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2 years ago
options.experimental_mempurge_threshold = 15.0;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ASSERT_OK(TryReopen(options));
std::atomic<uint32_t> mempurge_count{0};
std::atomic<uint32_t> sst_count{0};
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
"DBImpl::FlushJob:MemPurgeSuccessful",
[&](void* /*arg*/) { mempurge_count++; });
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
"DBImpl::FlushJob:SSTFileCreated", [&](void* /*arg*/) { sst_count++; });
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
std::string KEY1 = "ThisIsKey1";
std::string KEY2 = "ThisIsKey2";
std::string KEY3 = "ThisIsKey3";
std::string KEY4 = "ThisIsKey4";
std::string KEY5 = "ThisIsKey5";
const std::string NOT_FOUND = "NOT_FOUND";
Random rnd(117);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
const size_t NUM_REPEAT = 100;
const size_t RAND_VALUES_LENGTH = 10240;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
std::string key, value, p_v1, p_v2, p_v3, p_v3b, p_v4, p_v5;
int count = 0;
const int EXPECTED_COUNT_FORLOOP = 3;
const int EXPECTED_COUNT_END = 4;
ReadOptions ropt;
ropt.pin_data = true;
ropt.total_order_seek = true;
Iterator* iter = nullptr;
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Insertion of of K-V pairs, multiple times.
// Also insert DeleteRange
for (size_t i = 0; i < NUM_REPEAT; i++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
p_v1 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v2 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v3 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v3b = rnd.RandomString(RAND_VALUES_LENGTH);
p_v4 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v5 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEY1, p_v1));
ASSERT_OK(Put(KEY2, p_v2));
ASSERT_OK(Put(KEY3, p_v3));
ASSERT_OK(Put(KEY4, p_v4));
ASSERT_OK(Put(KEY5, p_v5));
ASSERT_OK(Delete(KEY2));
ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(), KEY2,
KEY4));
ASSERT_OK(Put(KEY3, p_v3b));
ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(), KEY1,
KEY3));
ASSERT_OK(Delete(KEY1));
ASSERT_EQ(Get(KEY1), NOT_FOUND);
ASSERT_EQ(Get(KEY2), NOT_FOUND);
ASSERT_EQ(Get(KEY3), p_v3b);
ASSERT_EQ(Get(KEY4), p_v4);
ASSERT_EQ(Get(KEY5), p_v5);
iter = db_->NewIterator(ropt);
iter->SeekToFirst();
count = 0;
for (; iter->Valid(); iter->Next()) {
ASSERT_OK(iter->status());
key = (iter->key()).ToString(false);
value = (iter->value()).ToString(false);
if (key.compare(KEY3) == 0)
ASSERT_EQ(value, p_v3b);
else if (key.compare(KEY4) == 0)
ASSERT_EQ(value, p_v4);
else if (key.compare(KEY5) == 0)
ASSERT_EQ(value, p_v5);
else
ASSERT_EQ(value, NOT_FOUND);
count++;
}
// Expected count here is 3: KEY3, KEY4, KEY5.
ASSERT_EQ(count, EXPECTED_COUNT_FORLOOP);
if (iter) {
delete iter;
}
}
// Check that there was at least one mempurge
const uint32_t EXPECTED_MIN_MEMPURGE_COUNT = 1;
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// Check that there was no SST files created during flush.
const uint32_t EXPECTED_SST_COUNT = 0;
EXPECT_GE(mempurge_count.exchange(0), EXPECTED_MIN_MEMPURGE_COUNT);
EXPECT_EQ(sst_count.exchange(0), EXPECTED_SST_COUNT);
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Additional test for the iterator+memPurge.
ASSERT_OK(Put(KEY2, p_v2));
iter = db_->NewIterator(ropt);
iter->SeekToFirst();
ASSERT_OK(Put(KEY4, p_v4));
count = 0;
for (; iter->Valid(); iter->Next()) {
ASSERT_OK(iter->status());
key = (iter->key()).ToString(false);
value = (iter->value()).ToString(false);
if (key.compare(KEY2) == 0)
ASSERT_EQ(value, p_v2);
else if (key.compare(KEY3) == 0)
ASSERT_EQ(value, p_v3b);
else if (key.compare(KEY4) == 0)
ASSERT_EQ(value, p_v4);
else if (key.compare(KEY5) == 0)
ASSERT_EQ(value, p_v5);
else
ASSERT_EQ(value, NOT_FOUND);
count++;
}
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Expected count here is 4: KEY2, KEY3, KEY4, KEY5.
ASSERT_EQ(count, EXPECTED_COUNT_END);
if (iter) delete iter;
Close();
}
// Create a Compaction Fitler that will be invoked
// at flush time and will update the value of a KV pair
// if the key string is "lower" than the filter_key_ string.
class ConditionalUpdateFilter : public CompactionFilter {
public:
explicit ConditionalUpdateFilter(const std::string* filtered_key)
: filtered_key_(filtered_key) {}
bool Filter(int /*level*/, const Slice& key, const Slice& /*value*/,
std::string* new_value, bool* value_changed) const override {
// If key<filtered_key_, update the value of the KV-pair.
if (key.compare(*filtered_key_) < 0) {
assert(new_value != nullptr);
*new_value = NEW_VALUE;
*value_changed = true;
}
return false /*do not remove this KV-pair*/;
}
const char* Name() const override { return "ConditionalUpdateFilter"; }
private:
const std::string* filtered_key_;
};
class ConditionalUpdateFilterFactory : public CompactionFilterFactory {
public:
explicit ConditionalUpdateFilterFactory(const Slice& filtered_key)
: filtered_key_(filtered_key.ToString()) {}
std::unique_ptr<CompactionFilter> CreateCompactionFilter(
const CompactionFilter::Context& /*context*/) override {
return std::unique_ptr<CompactionFilter>(
new ConditionalUpdateFilter(&filtered_key_));
}
const char* Name() const override { return "ConditionalUpdateFilterFactory"; }
bool ShouldFilterTableFileCreation(
TableFileCreationReason reason) const override {
// This compaction filter will be invoked
// at flush time (and therefore at MemPurge time).
return (reason == TableFileCreationReason::kFlush);
}
private:
std::string filtered_key_;
};
TEST_F(DBFlushTest, MemPurgeAndCompactionFilter) {
Options options = CurrentOptions();
std::string KEY1 = "ThisIsKey1";
std::string KEY2 = "ThisIsKey2";
std::string KEY3 = "ThisIsKey3";
std::string KEY4 = "ThisIsKey4";
std::string KEY5 = "ThisIsKey5";
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
std::string KEY6 = "ThisIsKey6";
std::string KEY7 = "ThisIsKey7";
std::string KEY8 = "ThisIsKey8";
std::string KEY9 = "ThisIsKey9";
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
const std::string NOT_FOUND = "NOT_FOUND";
options.statistics = CreateDBStatistics();
options.statistics->set_stats_level(StatsLevel::kAll);
options.create_if_missing = true;
options.compression = kNoCompression;
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
#ifndef ROCKSDB_LITE
TestFlushListener* listener = new TestFlushListener(options.env, this);
options.listeners.emplace_back(listener);
#endif // !ROCKSDB_LITE
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Create a ConditionalUpdate compaction filter
// that will update all the values of the KV pairs
// where the keys are "lower" than KEY4.
options.compaction_filter_factory =
std::make_shared<ConditionalUpdateFilterFactory>(KEY4);
// Enforce size of a single MemTable to 64MB (64MB = 67108864 bytes).
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
options.write_buffer_size = 1 << 20;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Activate the MemPurge prototype.
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2 years ago
options.experimental_mempurge_threshold = 26.55;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
ASSERT_OK(TryReopen(options));
std::atomic<uint32_t> mempurge_count{0};
std::atomic<uint32_t> sst_count{0};
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:MemPurgeSuccessful",
[&](void* /*arg*/) { mempurge_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:SSTFileCreated", [&](void* /*arg*/) { sst_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
Random rnd(53);
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
const size_t NUM_REPEAT = 1000;
const size_t RAND_VALUES_LENGTH = 10240;
std::string p_v1, p_v2, p_v3, p_v4, p_v5, p_v6, p_v7, p_v8, p_v9;
p_v1 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v2 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v3 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v4 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v5 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEY1, p_v1));
ASSERT_OK(Put(KEY2, p_v2));
ASSERT_OK(Put(KEY3, p_v3));
ASSERT_OK(Put(KEY4, p_v4));
ASSERT_OK(Put(KEY5, p_v5));
ASSERT_OK(Delete(KEY1));
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
// Insertion of of K-V pairs, multiple times.
for (size_t i = 0; i < NUM_REPEAT; i++) {
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// Create value strings of arbitrary
// length RAND_VALUES_LENGTH bytes.
p_v6 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v7 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v8 = rnd.RandomString(RAND_VALUES_LENGTH);
p_v9 = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEY6, p_v6));
ASSERT_OK(Put(KEY7, p_v7));
ASSERT_OK(Put(KEY8, p_v8));
ASSERT_OK(Put(KEY9, p_v9));
ASSERT_OK(Delete(KEY7));
}
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// Check that there was at least one mempurge
const uint32_t EXPECTED_MIN_MEMPURGE_COUNT = 1;
// Check that there was no SST files created during flush.
const uint32_t EXPECTED_SST_COUNT = 0;
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
EXPECT_GE(mempurge_count.exchange(0), EXPECTED_MIN_MEMPURGE_COUNT);
EXPECT_EQ(sst_count.exchange(0), EXPECTED_SST_COUNT);
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
3 years ago
// Verify that the ConditionalUpdateCompactionFilter
// updated the values of KEY2 and KEY3, and not KEY4 and KEY5.
ASSERT_EQ(Get(KEY1), NOT_FOUND);
ASSERT_EQ(Get(KEY2), NEW_VALUE);
ASSERT_EQ(Get(KEY3), NEW_VALUE);
ASSERT_EQ(Get(KEY4), p_v4);
ASSERT_EQ(Get(KEY5), p_v5);
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
3 years ago
}
TEST_F(DBFlushTest, DISABLED_MemPurgeWALSupport) {
Options options = CurrentOptions();
options.statistics = CreateDBStatistics();
options.statistics->set_stats_level(StatsLevel::kAll);
options.create_if_missing = true;
options.compression = kNoCompression;
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
Add simple heuristics for experimental mempurge. (#8583) Summary: Add `experimental_mempurge_policy` option flag and introduce two new `MemPurge` (Memtable Garbage Collection) policies: 'ALWAYS' and 'ALTERNATE'. Default value: ALTERNATE. `ALWAYS`: every flush will first go through a `MemPurge` process. If the output is too big to fit into a single memtable, then the mempurge is aborted and a regular flush process carries on. `ALWAYS` is designed for user that need to reduce the number of L0 SST file created to a strict minimum, and can afford a small dent in performance (possibly hits to CPU usage, read efficiency, and maximum burst write throughput). `ALTERNATE`: a flush is transformed into a `MemPurge` except if one of the memtables being flushed is the product of a previous `MemPurge`. `ALTERNATE` is a good tradeoff between reduction in number of L0 SST files created and performance. `ALTERNATE` perform particularly well for completely random garbage ratios, or garbage ratios anywhere in (0%,50%], and even higher when there is a wild variability in garbage ratios. This PR also includes support for `experimental_mempurge_policy` in `db_bench`. Testing was done locally by replacing all the `MemPurge` policies of the unit tests with `ALTERNATE`, as well as local testing with `db_crashtest.py` `whitebox` and `blackbox`. Overall, if an `ALWAYS` mempurge policy passes the tests, there is no reasons why an `ALTERNATE` policy would fail, and therefore the mempurge policy was set to `ALWAYS` for all mempurge unit tests. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8583 Reviewed By: pdillinger Differential Revision: D29888050 Pulled By: bjlemaire fbshipit-source-id: e2cf26646d66679f6f5fb29842624615610759c1
3 years ago
// Enforce size of a single MemTable to 128KB.
options.write_buffer_size = 128 << 10;
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2 years ago
// Activate the MemPurge prototype
// (values >1.0 are equivalent to 1.0).
options.experimental_mempurge_threshold = 2.5;
ASSERT_OK(TryReopen(options));
const size_t KVSIZE = 10;
do {
CreateAndReopenWithCF({"pikachu"}, options);
ASSERT_OK(Put(1, "foo", "v1"));
ASSERT_OK(Put(1, "baz", "v5"));
ReopenWithColumnFamilies({"default", "pikachu"}, options);
ASSERT_EQ("v1", Get(1, "foo"));
ASSERT_EQ("v1", Get(1, "foo"));
ASSERT_EQ("v5", Get(1, "baz"));
ASSERT_OK(Put(0, "bar", "v2"));
ASSERT_OK(Put(1, "bar", "v2"));
ASSERT_OK(Put(1, "foo", "v3"));
std::atomic<uint32_t> mempurge_count{0};
std::atomic<uint32_t> sst_count{0};
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:MemPurgeSuccessful",
[&](void* /*arg*/) { mempurge_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:SSTFileCreated", [&](void* /*arg*/) { sst_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
std::vector<std::string> keys;
for (size_t k = 0; k < KVSIZE; k++) {
keys.push_back("IamKey" + std::to_string(k));
}
std::string RNDKEY, RNDVALUE;
const std::string NOT_FOUND = "NOT_FOUND";
// Heavy overwrite workload,
// more than would fit in maximum allowed memtables.
Random rnd(719);
const size_t NUM_REPEAT = 100;
Add simple heuristics for experimental mempurge. (#8583) Summary: Add `experimental_mempurge_policy` option flag and introduce two new `MemPurge` (Memtable Garbage Collection) policies: 'ALWAYS' and 'ALTERNATE'. Default value: ALTERNATE. `ALWAYS`: every flush will first go through a `MemPurge` process. If the output is too big to fit into a single memtable, then the mempurge is aborted and a regular flush process carries on. `ALWAYS` is designed for user that need to reduce the number of L0 SST file created to a strict minimum, and can afford a small dent in performance (possibly hits to CPU usage, read efficiency, and maximum burst write throughput). `ALTERNATE`: a flush is transformed into a `MemPurge` except if one of the memtables being flushed is the product of a previous `MemPurge`. `ALTERNATE` is a good tradeoff between reduction in number of L0 SST files created and performance. `ALTERNATE` perform particularly well for completely random garbage ratios, or garbage ratios anywhere in (0%,50%], and even higher when there is a wild variability in garbage ratios. This PR also includes support for `experimental_mempurge_policy` in `db_bench`. Testing was done locally by replacing all the `MemPurge` policies of the unit tests with `ALTERNATE`, as well as local testing with `db_crashtest.py` `whitebox` and `blackbox`. Overall, if an `ALWAYS` mempurge policy passes the tests, there is no reasons why an `ALTERNATE` policy would fail, and therefore the mempurge policy was set to `ALWAYS` for all mempurge unit tests. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8583 Reviewed By: pdillinger Differential Revision: D29888050 Pulled By: bjlemaire fbshipit-source-id: e2cf26646d66679f6f5fb29842624615610759c1
3 years ago
const size_t RAND_KEY_LENGTH = 4096;
const size_t RAND_VALUES_LENGTH = 1024;
std::vector<std::string> values_default(KVSIZE), values_pikachu(KVSIZE);
// Insert a very first set of keys that will be
// mempurged at least once.
for (size_t k = 0; k < KVSIZE / 2; k++) {
values_default[k] = rnd.RandomString(RAND_VALUES_LENGTH);
values_pikachu[k] = rnd.RandomString(RAND_VALUES_LENGTH);
}
// Insert keys[0:KVSIZE/2] to
// both 'default' and 'pikachu' CFs.
for (size_t k = 0; k < KVSIZE / 2; k++) {
ASSERT_OK(Put(0, keys[k], values_default[k]));
ASSERT_OK(Put(1, keys[k], values_pikachu[k]));
}
// Check that the insertion was seamless.
for (size_t k = 0; k < KVSIZE / 2; k++) {
ASSERT_EQ(Get(0, keys[k]), values_default[k]);
ASSERT_EQ(Get(1, keys[k]), values_pikachu[k]);
}
// Insertion of of K-V pairs, multiple times (overwrites)
// into 'default' CF. Will trigger mempurge.
for (size_t j = 0; j < NUM_REPEAT; j++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
for (size_t k = KVSIZE / 2; k < KVSIZE; k++) {
values_default[k] = rnd.RandomString(RAND_VALUES_LENGTH);
}
// Insert K-V into default CF.
for (size_t k = KVSIZE / 2; k < KVSIZE; k++) {
ASSERT_OK(Put(0, keys[k], values_default[k]));
}
// Check key validity, for all keys, both in
// default and pikachu CFs.
for (size_t k = 0; k < KVSIZE; k++) {
ASSERT_EQ(Get(0, keys[k]), values_default[k]);
}
// Note that at this point, only keys[0:KVSIZE/2]
// have been inserted into Pikachu.
for (size_t k = 0; k < KVSIZE / 2; k++) {
ASSERT_EQ(Get(1, keys[k]), values_pikachu[k]);
}
}
// Insertion of of K-V pairs, multiple times (overwrites)
// into 'pikachu' CF. Will trigger mempurge.
// Check that we keep the older logs for 'default' imm().
for (size_t j = 0; j < NUM_REPEAT; j++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
for (size_t k = KVSIZE / 2; k < KVSIZE; k++) {
values_pikachu[k] = rnd.RandomString(RAND_VALUES_LENGTH);
}
// Insert K-V into pikachu CF.
for (size_t k = KVSIZE / 2; k < KVSIZE; k++) {
ASSERT_OK(Put(1, keys[k], values_pikachu[k]));
}
// Check key validity, for all keys,
// both in default and pikachu.
for (size_t k = 0; k < KVSIZE; k++) {
ASSERT_EQ(Get(0, keys[k]), values_default[k]);
ASSERT_EQ(Get(1, keys[k]), values_pikachu[k]);
}
}
// Check that there was at least one mempurge
const uint32_t EXPECTED_MIN_MEMPURGE_COUNT = 1;
// Check that there was no SST files created during flush.
const uint32_t EXPECTED_SST_COUNT = 0;
EXPECT_GE(mempurge_count.exchange(0), EXPECTED_MIN_MEMPURGE_COUNT);
Memtable sampling for mempurge heuristic. (#8628) Summary: Changes the API of the MemPurge process: the `bool experimental_allow_mempurge` and `experimental_mempurge_policy` flags have been replaced by a `double experimental_mempurge_threshold` option. This change of API reflects another major change introduced in this PR: the MemPurgeDecider() function now works by sampling the memtables being flushed to estimate the overall amount of useful payload (payload minus the garbage), and then compare this useful payload estimate with the `double experimental_mempurge_threshold` value. Therefore, when the value of this flag is `0.0` (default value), mempurge is simply deactivated. On the other hand, a value of `DBL_MAX` would be equivalent to always going through a mempurge regardless of the garbage ratio estimate. At the moment, a `double experimental_mempurge_threshold` value else than 0.0 or `DBL_MAX` is opnly supported`with the `SkipList` memtable representation. Regarding the sampling, this PR includes the introduction of a `MemTable::UniqueRandomSample` function that collects (approximately) random entries from the memtable by using the new `SkipList::Iterator::RandomSeek()` under the hood, or by iterating through each memtable entry, depending on the target sample size and the total number of entries. The unit tests have been readapted to support this new API. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8628 Reviewed By: pdillinger Differential Revision: D30149315 Pulled By: bjlemaire fbshipit-source-id: 1feef5390c95db6f4480ab4434716533d3947f27
3 years ago
if (options.experimental_mempurge_threshold ==
std::numeric_limits<double>::max()) {
EXPECT_EQ(sst_count.exchange(0), EXPECTED_SST_COUNT);
Add simple heuristics for experimental mempurge. (#8583) Summary: Add `experimental_mempurge_policy` option flag and introduce two new `MemPurge` (Memtable Garbage Collection) policies: 'ALWAYS' and 'ALTERNATE'. Default value: ALTERNATE. `ALWAYS`: every flush will first go through a `MemPurge` process. If the output is too big to fit into a single memtable, then the mempurge is aborted and a regular flush process carries on. `ALWAYS` is designed for user that need to reduce the number of L0 SST file created to a strict minimum, and can afford a small dent in performance (possibly hits to CPU usage, read efficiency, and maximum burst write throughput). `ALTERNATE`: a flush is transformed into a `MemPurge` except if one of the memtables being flushed is the product of a previous `MemPurge`. `ALTERNATE` is a good tradeoff between reduction in number of L0 SST files created and performance. `ALTERNATE` perform particularly well for completely random garbage ratios, or garbage ratios anywhere in (0%,50%], and even higher when there is a wild variability in garbage ratios. This PR also includes support for `experimental_mempurge_policy` in `db_bench`. Testing was done locally by replacing all the `MemPurge` policies of the unit tests with `ALTERNATE`, as well as local testing with `db_crashtest.py` `whitebox` and `blackbox`. Overall, if an `ALWAYS` mempurge policy passes the tests, there is no reasons why an `ALTERNATE` policy would fail, and therefore the mempurge policy was set to `ALWAYS` for all mempurge unit tests. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8583 Reviewed By: pdillinger Differential Revision: D29888050 Pulled By: bjlemaire fbshipit-source-id: e2cf26646d66679f6f5fb29842624615610759c1
3 years ago
}
ReopenWithColumnFamilies({"default", "pikachu"}, options);
// Check that there was no data corruption anywhere,
// not in 'default' nor in 'Pikachu' CFs.
ASSERT_EQ("v3", Get(1, "foo"));
ASSERT_OK(Put(1, "foo", "v4"));
ASSERT_EQ("v4", Get(1, "foo"));
ASSERT_EQ("v2", Get(1, "bar"));
ASSERT_EQ("v5", Get(1, "baz"));
// Check keys in 'Default' and 'Pikachu'.
// keys[0:KVSIZE/2] were for sure contained
// in the imm() at Reopen/recovery time.
for (size_t k = 0; k < KVSIZE; k++) {
ASSERT_EQ(Get(0, keys[k]), values_default[k]);
ASSERT_EQ(Get(1, keys[k]), values_pikachu[k]);
}
// Insertion of random K-V pairs to trigger
// a flush in the Pikachu CF.
for (size_t j = 0; j < NUM_REPEAT; j++) {
RNDKEY = rnd.RandomString(RAND_KEY_LENGTH);
RNDVALUE = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(1, RNDKEY, RNDVALUE));
}
// ASsert than there was at least one flush to storage.
EXPECT_GT(sst_count.exchange(0), EXPECTED_SST_COUNT);
ReopenWithColumnFamilies({"default", "pikachu"}, options);
ASSERT_EQ("v4", Get(1, "foo"));
ASSERT_EQ("v2", Get(1, "bar"));
ASSERT_EQ("v5", Get(1, "baz"));
// Since values in default are held in mutable mem()
// and imm(), check if the flush in pikachu didn't
// affect these values.
for (size_t k = 0; k < KVSIZE; k++) {
ASSERT_EQ(Get(0, keys[k]), values_default[k]);
ASSERT_EQ(Get(1, keys[k]), values_pikachu[k]);
}
ASSERT_EQ(Get(1, RNDKEY), RNDVALUE);
} while (ChangeWalOptions());
}
Fix mempurge crash reported in #8958 (#9671) Summary: Change the `MemPurge` code to address a failure during a crash test reported in https://github.com/facebook/rocksdb/issues/8958. ### Details and results of the crash investigation: These failures happened in a specific scenario where the list of immutable tables was composed of 2 or more memtables, and the last memtable was the output of a previous `Mempurge` operation. Because the `PickMemtablesToFlush` function included a sorting of the memtables (previous PR related to the Mempurge project), and because the `VersionEdit` of the flush class is piggybacked onto a single one of these memtables, the `VersionEdit` was not properly selected and applied to the `VersionSet` of the DB. Since the `VersionSet` was not edited properly, the database was losing track of the SST file created during the flush process, which was subsequently deleted (and as you can expect, caused the tests to crash). The following command consistently failed, which was quite convenient to investigate the issue: `$ while rm -rf /dev/shm/single_stress && ./db_stress --clear_column_family_one_in=0 --column_families=1 --db=/dev/shm/single_stress --experimental_mempurge_threshold=5.493146827397074 --flush_one_in=10000 --reopen=0 --write_buffer_size=262144 --value_size_mult=33 --max_write_buffer_number=3 -ops_per_thread=10000; do : ; done` ### Solution proposed The memtables are no longer sorted based on their `memtableID` in the `PickMemtablesToFlush` function. Additionally, the `next_log_number` of the memtable created as an output of the `Mempurge` function now takes in the correct value (the log number of the first memtable being mempurged). Finally, the VersionEdit object of the flush class now takes the maximum `next_log_number` of the stack of memtables being flushed, which doesnt change anything when Mempurge is `off` but becomes necessary when Mempurge is `on`. ### Testing of the solution The following command no longer fails: ``$ while rm -rf /dev/shm/single_stress && ./db_stress --clear_column_family_one_in=0 --column_families=1 --db=/dev/shm/single_stress --experimental_mempurge_threshold=5.493146827397074 --flush_one_in=10000 --reopen=0 --write_buffer_size=262144 --value_size_mult=33 --max_write_buffer_number=3 -ops_per_thread=10000; do : ; done`` Additionally, I ran `db_crashtest` (`whitebox` and `blackbox`) for 2.5 hours with MemPurge on and did not observe any crash. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9671 Reviewed By: pdillinger Differential Revision: D34697424 Pulled By: bjlemaire fbshipit-source-id: d1ab675b361904351ac81a35c184030e52222874
2 years ago
TEST_F(DBFlushTest, MemPurgeCorrectLogNumberAndSSTFileCreation) {
// Before our bug fix, we noticed that when 2 memtables were
// being flushed (with one memtable being the output of a
// previous MemPurge and one memtable being a newly-sealed memtable),
// the SST file created was not properly added to the DB version
// (via the VersionEdit obj), leading to data loss (the SST file
// was later being purged as an obsolete file).
// Therefore, we reproduce this scenario to test our fix.
Options options = CurrentOptions();
options.create_if_missing = true;
options.compression = kNoCompression;
options.inplace_update_support = false;
options.allow_concurrent_memtable_write = true;
// Enforce size of a single MemTable to 1MB (64MB = 1048576 bytes).
options.write_buffer_size = 1 << 20;
// Activate the MemPurge prototype.
options.experimental_mempurge_threshold = 1.0;
// Force to have more than one memtable to trigger a flush.
// For some reason this option does not seem to be enforced,
// so the following test is designed to make sure that we
// are testing the correct test case.
options.min_write_buffer_number_to_merge = 3;
options.max_write_buffer_number = 5;
options.max_write_buffer_size_to_maintain = 2 * (options.write_buffer_size);
options.disable_auto_compactions = true;
ASSERT_OK(TryReopen(options));
std::atomic<uint32_t> mempurge_count{0};
std::atomic<uint32_t> sst_count{0};
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:MemPurgeSuccessful",
[&](void* /*arg*/) { mempurge_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushJob:SSTFileCreated", [&](void* /*arg*/) { sst_count++; });
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
// Dummy variable used for the following callback function.
uint64_t ZERO = 0;
// We will first execute mempurge operations exclusively.
// Therefore, when the first flush is triggered, we want to make
// sure there is at least 2 memtables being flushed: one output
// from a previous mempurge, and one newly sealed memtable.
// This is when we observed in the past that some SST files created
// were not properly added to the DB version (via the VersionEdit obj).
std::atomic<uint64_t> num_memtable_at_first_flush(0);
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"FlushJob::WriteLevel0Table:num_memtables", [&](void* arg) {
uint64_t* mems_size = reinterpret_cast<uint64_t*>(arg);
// atomic_compare_exchange_strong sometimes updates the value
// of ZERO (the "expected" object), so we make sure ZERO is indeed...
// zero.
ZERO = 0;
std::atomic_compare_exchange_strong(&num_memtable_at_first_flush, &ZERO,
*mems_size);
});
const std::vector<std::string> KEYS = {
"ThisIsKey1", "ThisIsKey2", "ThisIsKey3", "ThisIsKey4", "ThisIsKey5",
"ThisIsKey6", "ThisIsKey7", "ThisIsKey8", "ThisIsKey9"};
const std::string NOT_FOUND = "NOT_FOUND";
Random rnd(117);
const uint64_t NUM_REPEAT_OVERWRITES = 100;
const uint64_t NUM_RAND_INSERTS = 500;
const uint64_t RAND_VALUES_LENGTH = 10240;
std::string key, value;
std::vector<std::string> values(9, "");
// Keys used to check that no SST file disappeared.
for (uint64_t k = 0; k < 5; k++) {
values[k] = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEYS[k], values[k]));
}
// Insertion of of K-V pairs, multiple times.
// Trigger at least one mempurge and no SST file creation.
for (size_t i = 0; i < NUM_REPEAT_OVERWRITES; i++) {
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
for (uint64_t k = 5; k < values.size(); k++) {
values[k] = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(KEYS[k], values[k]));
}
// Check database consistency.
for (uint64_t k = 0; k < values.size(); k++) {
ASSERT_EQ(Get(KEYS[k]), values[k]);
}
}
// Check that there was at least one mempurge
uint32_t expected_min_mempurge_count = 1;
// Check that there was no SST files created during flush.
uint32_t expected_sst_count = 0;
EXPECT_GE(mempurge_count.load(), expected_min_mempurge_count);
EXPECT_EQ(sst_count.load(), expected_sst_count);
// Trigger an SST file creation and no mempurge.
for (size_t i = 0; i < NUM_RAND_INSERTS; i++) {
key = rnd.RandomString(RAND_VALUES_LENGTH);
// Create value strings of arbitrary length RAND_VALUES_LENGTH bytes.
value = rnd.RandomString(RAND_VALUES_LENGTH);
ASSERT_OK(Put(key, value));
// Check database consistency.
for (uint64_t k = 0; k < values.size(); k++) {
ASSERT_EQ(Get(KEYS[k]), values[k]);
}
ASSERT_EQ(Get(key), value);
}
// Check that there was at least one SST files created during flush.
expected_sst_count = 1;
EXPECT_GE(sst_count.load(), expected_sst_count);
// Oddly enough, num_memtable_at_first_flush is not enforced to be
// equal to min_write_buffer_number_to_merge. So by asserting that
// the first SST file creation comes from one output memtable
// from a previous mempurge, and one newly sealed memtable. This
// is the scenario where we observed that some SST files created
// were not properly added to the DB version before our bug fix.
ASSERT_GE(num_memtable_at_first_flush.load(), 2);
// Check that no data was lost after SST file creation.
for (uint64_t k = 0; k < values.size(); k++) {
ASSERT_EQ(Get(KEYS[k]), values[k]);
}
// Extra check of database consistency.
ASSERT_EQ(Get(key), value);
Close();
}
TEST_P(DBFlushDirectIOTest, DirectIO) {
Options options;
options.create_if_missing = true;
options.disable_auto_compactions = true;
options.max_background_flushes = 2;
options.use_direct_io_for_flush_and_compaction = GetParam();
options.env = MockEnv::Create(Env::Default());
SyncPoint::GetInstance()->SetCallBack(
"BuildTable:create_file", [&](void* arg) {
bool* use_direct_writes = static_cast<bool*>(arg);
ASSERT_EQ(*use_direct_writes,
options.use_direct_io_for_flush_and_compaction);
});
SyncPoint::GetInstance()->EnableProcessing();
Reopen(options);
ASSERT_OK(Put("foo", "v"));
FlushOptions flush_options;
flush_options.wait = true;
ASSERT_OK(dbfull()->Flush(flush_options));
Destroy(options);
delete options.env;
}
TEST_F(DBFlushTest, FlushError) {
Options options;
std::unique_ptr<FaultInjectionTestEnv> fault_injection_env(
new FaultInjectionTestEnv(env_));
options.write_buffer_size = 100;
options.max_write_buffer_number = 4;
options.min_write_buffer_number_to_merge = 3;
options.disable_auto_compactions = true;
options.env = fault_injection_env.get();
Reopen(options);
ASSERT_OK(Put("key1", "value1"));
ASSERT_OK(Put("key2", "value2"));
fault_injection_env->SetFilesystemActive(false);
Status s = dbfull()->TEST_SwitchMemtable();
fault_injection_env->SetFilesystemActive(true);
Destroy(options);
ASSERT_NE(s, Status::OK());
}
TEST_F(DBFlushTest, ManualFlushFailsInReadOnlyMode) {
// Regression test for bug where manual flush hangs forever when the DB
// is in read-only mode. Verify it now at least returns, despite failing.
Options options;
std::unique_ptr<FaultInjectionTestEnv> fault_injection_env(
new FaultInjectionTestEnv(env_));
options.env = fault_injection_env.get();
options.max_write_buffer_number = 2;
Reopen(options);
// Trigger a first flush but don't let it run
ASSERT_OK(db_->PauseBackgroundWork());
ASSERT_OK(Put("key1", "value1"));
FlushOptions flush_opts;
flush_opts.wait = false;
ASSERT_OK(db_->Flush(flush_opts));
// Write a key to the second memtable so we have something to flush later
// after the DB is in read-only mode.
ASSERT_OK(Put("key2", "value2"));
// Let the first flush continue, hit an error, and put the DB in read-only
// mode.
fault_injection_env->SetFilesystemActive(false);
ASSERT_OK(db_->ContinueBackgroundWork());
// We ingested the error to env, so the returned status is not OK.
ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable());
#ifndef ROCKSDB_LITE
uint64_t num_bg_errors;
ASSERT_TRUE(db_->GetIntProperty(DB::Properties::kBackgroundErrors,
&num_bg_errors));
ASSERT_GT(num_bg_errors, 0);
#endif // ROCKSDB_LITE
// In the bug scenario, triggering another flush would cause the second flush
// to hang forever. After the fix we expect it to return an error.
ASSERT_NOK(db_->Flush(FlushOptions()));
Close();
}
TEST_F(DBFlushTest, CFDropRaceWithWaitForFlushMemTables) {
Options options = CurrentOptions();
options.create_if_missing = true;
CreateAndReopenWithCF({"pikachu"}, options);
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::FlushMemTable:AfterScheduleFlush",
"DBFlushTest::CFDropRaceWithWaitForFlushMemTables:BeforeDrop"},
{"DBFlushTest::CFDropRaceWithWaitForFlushMemTables:AfterFree",
"DBImpl::BackgroundCallFlush:start"},
{"DBImpl::BackgroundCallFlush:start",
"DBImpl::FlushMemTable:BeforeWaitForBgFlush"}});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_EQ(2, handles_.size());
ASSERT_OK(Put(1, "key", "value"));
auto* cfd = static_cast<ColumnFamilyHandleImpl*>(handles_[1])->cfd();
port::Thread drop_cf_thr([&]() {
TEST_SYNC_POINT(
"DBFlushTest::CFDropRaceWithWaitForFlushMemTables:BeforeDrop");
ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
ASSERT_OK(dbfull()->DestroyColumnFamilyHandle(handles_[1]));
handles_.resize(1);
TEST_SYNC_POINT(
"DBFlushTest::CFDropRaceWithWaitForFlushMemTables:AfterFree");
});
FlushOptions flush_opts;
flush_opts.allow_write_stall = true;
ASSERT_NOK(dbfull()->TEST_FlushMemTable(cfd, flush_opts));
drop_cf_thr.join();
Close();
SyncPoint::GetInstance()->DisableProcessing();
}
#ifndef ROCKSDB_LITE
TEST_F(DBFlushTest, FireOnFlushCompletedAfterCommittedResult) {
class TestListener : public EventListener {
public:
void OnFlushCompleted(DB* db, const FlushJobInfo& info) override {
// There's only one key in each flush.
ASSERT_EQ(info.smallest_seqno, info.largest_seqno);
ASSERT_NE(0, info.smallest_seqno);
if (info.smallest_seqno == seq1) {
// First flush completed
ASSERT_FALSE(completed1);
completed1 = true;
CheckFlushResultCommitted(db, seq1);
} else {
// Second flush completed
ASSERT_FALSE(completed2);
completed2 = true;
ASSERT_EQ(info.smallest_seqno, seq2);
CheckFlushResultCommitted(db, seq2);
}
}
void CheckFlushResultCommitted(DB* db, SequenceNumber seq) {
DBImpl* db_impl = static_cast_with_check<DBImpl>(db);
InstrumentedMutex* mutex = db_impl->mutex();
mutex->Lock();
auto* cfd = static_cast_with_check<ColumnFamilyHandleImpl>(
db->DefaultColumnFamily())
->cfd();
ASSERT_LT(seq, cfd->imm()->current()->GetEarliestSequenceNumber());
mutex->Unlock();
}
std::atomic<SequenceNumber> seq1{0};
std::atomic<SequenceNumber> seq2{0};
std::atomic<bool> completed1{false};
std::atomic<bool> completed2{false};
};
std::shared_ptr<TestListener> listener = std::make_shared<TestListener>();
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::FlushMemTableToOutputFile:AfterPickMemtables",
"DBFlushTest::FireOnFlushCompletedAfterCommittedResult:WaitFirst"},
{"DBImpl::FlushMemTableToOutputFile:Finish",
"DBFlushTest::FireOnFlushCompletedAfterCommittedResult:WaitSecond"}});
SyncPoint::GetInstance()->SetCallBack(
"FlushJob::WriteLevel0Table", [&listener](void* arg) {
// Wait for the second flush finished, out of mutex.
auto* mems = reinterpret_cast<autovector<MemTable*>*>(arg);
if (mems->front()->GetEarliestSequenceNumber() == listener->seq1 - 1) {
TEST_SYNC_POINT(
"DBFlushTest::FireOnFlushCompletedAfterCommittedResult:"
"WaitSecond");
}
});
Options options = CurrentOptions();
options.create_if_missing = true;
options.listeners.push_back(listener);
// Setting max_flush_jobs = max_background_jobs / 4 = 2.
options.max_background_jobs = 8;
// Allow 2 immutable memtables.
options.max_write_buffer_number = 3;
Reopen(options);
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Put("foo", "v"));
listener->seq1 = db_->GetLatestSequenceNumber();
// t1 will wait for the second flush complete before committing flush result.
auto t1 = port::Thread([&]() {
// flush_opts.wait = true
ASSERT_OK(db_->Flush(FlushOptions()));
});
// Wait for first flush started.
TEST_SYNC_POINT(
"DBFlushTest::FireOnFlushCompletedAfterCommittedResult:WaitFirst");
// The second flush will exit early without commit its result. The work
// is delegated to the first flush.
ASSERT_OK(Put("bar", "v"));
listener->seq2 = db_->GetLatestSequenceNumber();
FlushOptions flush_opts;
flush_opts.wait = false;
ASSERT_OK(db_->Flush(flush_opts));
t1.join();
Fix test race conditions with OnFlushCompleted() (#9617) Summary: We often see flaky tests due to `DB::Flush()` or `DBImpl::TEST_WaitForFlushMemTable()` not waiting until event listeners complete. For example, https://github.com/facebook/rocksdb/issues/9084, https://github.com/facebook/rocksdb/issues/9400, https://github.com/facebook/rocksdb/issues/9528, plus two new ones this week: "EventListenerTest.OnSingleDBFlushTest" and "DBFlushTest.FireOnFlushCompletedAfterCommittedResult". I ran a `make check` with the below race condition-coercing patch and fixed issues it found besides old BlobDB. ``` diff --git a/db/db_impl/db_impl_compaction_flush.cc b/db/db_impl/db_impl_compaction_flush.cc index 0e1864788..aaba68c4a 100644 --- a/db/db_impl/db_impl_compaction_flush.cc +++ b/db/db_impl/db_impl_compaction_flush.cc @@ -861,6 +861,8 @@ void DBImpl::NotifyOnFlushCompleted( mutable_cf_options.level0_stop_writes_trigger); // release lock while notifying events mutex_.Unlock(); + bg_cv_.SignalAll(); + sleep(1); { for (auto& info : *flush_jobs_info) { info->triggered_writes_slowdown = triggered_writes_slowdown; ``` The reason I did not fix old BlobDB issues is because it appears to have a fundamental (non-test) issue. In particular, it uses an EventListener to keep track of the files. OnFlushCompleted() could be delayed until even after a compaction involving that flushed file completes, causing the compaction to unexpectedly delete an untracked file. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9617 Test Plan: `make check` including the race condition coercing patch Reviewed By: hx235 Differential Revision: D34384022 Pulled By: ajkr fbshipit-source-id: 2652ded39b415277c5d6a628414345223930514e
2 years ago
// Ensure background work is fully finished including listener callbacks
// before accessing listener state.
ASSERT_OK(dbfull()->TEST_WaitForBackgroundWork());
ASSERT_TRUE(listener->completed1);
ASSERT_TRUE(listener->completed2);
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
}
#endif // !ROCKSDB_LITE
TEST_F(DBFlushTest, FlushWithBlob) {
constexpr uint64_t min_blob_size = 10;
Options options;
options.enable_blob_files = true;
options.min_blob_size = min_blob_size;
options.disable_auto_compactions = true;
options.env = env_;
Reopen(options);
constexpr char short_value[] = "short";
static_assert(sizeof(short_value) - 1 < min_blob_size,
"short_value too long");
constexpr char long_value[] = "long_value";
static_assert(sizeof(long_value) - 1 >= min_blob_size,
"long_value too short");
ASSERT_OK(Put("key1", short_value));
ASSERT_OK(Put("key2", long_value));
ASSERT_OK(Flush());
ASSERT_EQ(Get("key1"), short_value);
ASSERT_EQ(Get("key2"), long_value);
VersionSet* const versions = dbfull()->GetVersionSet();
assert(versions);
ColumnFamilyData* const cfd = versions->GetColumnFamilySet()->GetDefault();
assert(cfd);
Version* const current = cfd->current();
assert(current);
const VersionStorageInfo* const storage_info = current->storage_info();
assert(storage_info);
const auto& l0_files = storage_info->LevelFiles(0);
ASSERT_EQ(l0_files.size(), 1);
const FileMetaData* const table_file = l0_files[0];
assert(table_file);
const auto& blob_files = storage_info->GetBlobFiles();
ASSERT_EQ(blob_files.size(), 1);
Use a sorted vector instead of a map to store blob file metadata (#9526) Summary: The patch replaces `std::map` with a sorted `std::vector` for `VersionStorageInfo::blob_files_` and preallocates the space for the `vector` before saving the `BlobFileMetaData` into the new `VersionStorageInfo` in `VersionBuilder::Rep::SaveBlobFilesTo`. These changes reduce the time the DB mutex is held while saving new `Version`s, and using a sorted `vector` also makes lookups faster thanks to better memory locality. In addition, the patch introduces helper methods `VersionStorageInfo::GetBlobFileMetaData` and `VersionStorageInfo::GetBlobFileMetaDataLB` that can be used by clients to perform lookups in the `vector`, and does some general cleanup in the parts of code where blob file metadata are used. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9526 Test Plan: Ran `make check` and the crash test script for a while. Performance was tested using a load-optimized benchmark (`fillseq` with vector memtable, no WAL) and small file sizes so that a significant number of files are produced: ``` numactl --interleave=all ./db_bench --benchmarks=fillseq --allow_concurrent_memtable_write=false --level0_file_num_compaction_trigger=4 --level0_slowdown_writes_trigger=20 --level0_stop_writes_trigger=30 --max_background_jobs=8 --max_write_buffer_number=8 --db=/data/ltamasi-dbbench --wal_dir=/data/ltamasi-dbbench --num=800000000 --num_levels=8 --key_size=20 --value_size=400 --block_size=8192 --cache_size=51539607552 --cache_numshardbits=6 --compression_max_dict_bytes=0 --compression_ratio=0.5 --compression_type=lz4 --bytes_per_sync=8388608 --cache_index_and_filter_blocks=1 --cache_high_pri_pool_ratio=0.5 --benchmark_write_rate_limit=0 --write_buffer_size=16777216 --target_file_size_base=16777216 --max_bytes_for_level_base=67108864 --verify_checksum=1 --delete_obsolete_files_period_micros=62914560 --max_bytes_for_level_multiplier=8 --statistics=0 --stats_per_interval=1 --stats_interval_seconds=20 --histogram=1 --memtablerep=skip_list --bloom_bits=10 --open_files=-1 --subcompactions=1 --compaction_style=0 --min_level_to_compress=3 --level_compaction_dynamic_level_bytes=true --pin_l0_filter_and_index_blocks_in_cache=1 --soft_pending_compaction_bytes_limit=167503724544 --hard_pending_compaction_bytes_limit=335007449088 --min_level_to_compress=0 --use_existing_db=0 --sync=0 --threads=1 --memtablerep=vector --allow_concurrent_memtable_write=false --disable_wal=1 --enable_blob_files=1 --blob_file_size=16777216 --min_blob_size=0 --blob_compression_type=lz4 --enable_blob_garbage_collection=1 --seed=<some value> ``` Final statistics before the patch: ``` Cumulative writes: 0 writes, 700M keys, 0 commit groups, 0.0 writes per commit group, ingest: 284.62 GB, 121.27 MB/s Interval writes: 0 writes, 334K keys, 0 commit groups, 0.0 writes per commit group, ingest: 139.28 MB, 72.46 MB/s ``` With the patch: ``` Cumulative writes: 0 writes, 760M keys, 0 commit groups, 0.0 writes per commit group, ingest: 308.66 GB, 131.52 MB/s Interval writes: 0 writes, 445K keys, 0 commit groups, 0.0 writes per commit group, ingest: 185.35 MB, 93.15 MB/s ``` Total time to complete the benchmark is 2611 seconds with the patch, down from 2986 secs. Reviewed By: riversand963 Differential Revision: D34082728 Pulled By: ltamasi fbshipit-source-id: fc598abf676dce436734d06bb9d2d99a26a004fc
2 years ago
const auto& blob_file = blob_files.front();
assert(blob_file);
ASSERT_EQ(table_file->smallest.user_key(), "key1");
ASSERT_EQ(table_file->largest.user_key(), "key2");
ASSERT_EQ(table_file->fd.smallest_seqno, 1);
ASSERT_EQ(table_file->fd.largest_seqno, 2);
ASSERT_EQ(table_file->oldest_blob_file_number,
blob_file->GetBlobFileNumber());
ASSERT_EQ(blob_file->GetTotalBlobCount(), 1);
#ifndef ROCKSDB_LITE
const InternalStats* const internal_stats = cfd->internal_stats();
assert(internal_stats);
const auto& compaction_stats = internal_stats->TEST_GetCompactionStats();
ASSERT_FALSE(compaction_stats.empty());
ASSERT_EQ(compaction_stats[0].bytes_written, table_file->fd.GetFileSize());
ASSERT_EQ(compaction_stats[0].bytes_written_blob,
blob_file->GetTotalBlobBytes());
ASSERT_EQ(compaction_stats[0].num_output_files, 1);
ASSERT_EQ(compaction_stats[0].num_output_files_blob, 1);
const uint64_t* const cf_stats_value = internal_stats->TEST_GetCFStatsValue();
ASSERT_EQ(cf_stats_value[InternalStats::BYTES_FLUSHED],
compaction_stats[0].bytes_written +
compaction_stats[0].bytes_written_blob);
#endif // ROCKSDB_LITE
}
TEST_F(DBFlushTest, FlushWithChecksumHandoff1) {
if (mem_env_ || encrypted_env_) {
ROCKSDB_GTEST_SKIP("Test requires non-mem or non-encrypted environment");
return;
}
std::shared_ptr<FaultInjectionTestFS> fault_fs(
new FaultInjectionTestFS(FileSystem::Default()));
std::unique_ptr<Env> fault_fs_env(NewCompositeEnv(fault_fs));
Options options = CurrentOptions();
options.write_buffer_size = 100;
options.max_write_buffer_number = 4;
options.min_write_buffer_number_to_merge = 3;
options.disable_auto_compactions = true;
options.env = fault_fs_env.get();
options.checksum_handoff_file_types.Add(FileType::kTableFile);
Reopen(options);
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kCRC32c);
ASSERT_OK(Put("key1", "value1"));
ASSERT_OK(Put("key2", "value2"));
ASSERT_OK(dbfull()->TEST_SwitchMemtable());
// The hash does not match, write fails
// fault_fs->SetChecksumHandoffFuncType(ChecksumType::kxxHash);
// Since the file system returns IOStatus::Corruption, it is an
// unrecoverable error.
SyncPoint::GetInstance()->SetCallBack("FlushJob::Start", [&](void*) {
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kxxHash);
});
ASSERT_OK(Put("key3", "value3"));
ASSERT_OK(Put("key4", "value4"));
SyncPoint::GetInstance()->EnableProcessing();
Status s = Flush();
ASSERT_EQ(s.severity(),
ROCKSDB_NAMESPACE::Status::Severity::kUnrecoverableError);
SyncPoint::GetInstance()->DisableProcessing();
Destroy(options);
Reopen(options);
// The file system does not support checksum handoff. The check
// will be ignored.
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kNoChecksum);
ASSERT_OK(Put("key5", "value5"));
ASSERT_OK(Put("key6", "value6"));
ASSERT_OK(dbfull()->TEST_SwitchMemtable());
// Each write will be similated as corrupted.
// Since the file system returns IOStatus::Corruption, it is an
// unrecoverable error.
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kCRC32c);
SyncPoint::GetInstance()->SetCallBack("FlushJob::Start", [&](void*) {
fault_fs->IngestDataCorruptionBeforeWrite();
});
ASSERT_OK(Put("key7", "value7"));
ASSERT_OK(Put("key8", "value8"));
SyncPoint::GetInstance()->EnableProcessing();
s = Flush();
ASSERT_EQ(s.severity(),
ROCKSDB_NAMESPACE::Status::Severity::kUnrecoverableError);
SyncPoint::GetInstance()->DisableProcessing();
Destroy(options);
}
TEST_F(DBFlushTest, FlushWithChecksumHandoff2) {
if (mem_env_ || encrypted_env_) {
ROCKSDB_GTEST_SKIP("Test requires non-mem or non-encrypted environment");
return;
}
std::shared_ptr<FaultInjectionTestFS> fault_fs(
new FaultInjectionTestFS(FileSystem::Default()));
std::unique_ptr<Env> fault_fs_env(NewCompositeEnv(fault_fs));
Options options = CurrentOptions();
options.write_buffer_size = 100;
options.max_write_buffer_number = 4;
options.min_write_buffer_number_to_merge = 3;
options.disable_auto_compactions = true;
options.env = fault_fs_env.get();
Reopen(options);
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kCRC32c);
ASSERT_OK(Put("key1", "value1"));
ASSERT_OK(Put("key2", "value2"));
ASSERT_OK(Flush());
// options is not set, the checksum handoff will not be triggered
SyncPoint::GetInstance()->SetCallBack("FlushJob::Start", [&](void*) {
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kxxHash);
});
ASSERT_OK(Put("key3", "value3"));
ASSERT_OK(Put("key4", "value4"));
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Flush());
SyncPoint::GetInstance()->DisableProcessing();
Destroy(options);
Reopen(options);
// The file system does not support checksum handoff. The check
// will be ignored.
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kNoChecksum);
ASSERT_OK(Put("key5", "value5"));
ASSERT_OK(Put("key6", "value6"));
ASSERT_OK(Flush());
// options is not set, the checksum handoff will not be triggered
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kCRC32c);
SyncPoint::GetInstance()->SetCallBack("FlushJob::Start", [&](void*) {
fault_fs->IngestDataCorruptionBeforeWrite();
});
ASSERT_OK(Put("key7", "value7"));
ASSERT_OK(Put("key8", "value8"));
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Flush());
SyncPoint::GetInstance()->DisableProcessing();
Destroy(options);
}
TEST_F(DBFlushTest, FlushWithChecksumHandoffManifest1) {
if (mem_env_ || encrypted_env_) {
ROCKSDB_GTEST_SKIP("Test requires non-mem or non-encrypted environment");
return;
}
std::shared_ptr<FaultInjectionTestFS> fault_fs(
new FaultInjectionTestFS(FileSystem::Default()));
std::unique_ptr<Env> fault_fs_env(NewCompositeEnv(fault_fs));
Options options = CurrentOptions();
options.write_buffer_size = 100;
options.max_write_buffer_number = 4;
options.min_write_buffer_number_to_merge = 3;
options.disable_auto_compactions = true;
options.env = fault_fs_env.get();
options.checksum_handoff_file_types.Add(FileType::kDescriptorFile);
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kCRC32c);
Reopen(options);
ASSERT_OK(Put("key1", "value1"));
ASSERT_OK(Put("key2", "value2"));
ASSERT_OK(Flush());
// The hash does not match, write fails
// fault_fs->SetChecksumHandoffFuncType(ChecksumType::kxxHash);
// Since the file system returns IOStatus::Corruption, it is mapped to
// kFatalError error.
ASSERT_OK(Put("key3", "value3"));
SyncPoint::GetInstance()->SetCallBack(
"VersionSet::LogAndApply:WriteManifest", [&](void*) {
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kxxHash);
});
ASSERT_OK(Put("key3", "value3"));
ASSERT_OK(Put("key4", "value4"));
SyncPoint::GetInstance()->EnableProcessing();
Status s = Flush();
ASSERT_EQ(s.severity(), ROCKSDB_NAMESPACE::Status::Severity::kFatalError);
SyncPoint::GetInstance()->DisableProcessing();
Destroy(options);
}
TEST_F(DBFlushTest, FlushWithChecksumHandoffManifest2) {
if (mem_env_ || encrypted_env_) {
ROCKSDB_GTEST_SKIP("Test requires non-mem or non-encrypted environment");
return;
}
std::shared_ptr<FaultInjectionTestFS> fault_fs(
new FaultInjectionTestFS(FileSystem::Default()));
std::unique_ptr<Env> fault_fs_env(NewCompositeEnv(fault_fs));
Options options = CurrentOptions();
options.write_buffer_size = 100;
options.max_write_buffer_number = 4;
options.min_write_buffer_number_to_merge = 3;
options.disable_auto_compactions = true;
options.env = fault_fs_env.get();
options.checksum_handoff_file_types.Add(FileType::kDescriptorFile);
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kNoChecksum);
Reopen(options);
// The file system does not support checksum handoff. The check
// will be ignored.
ASSERT_OK(Put("key5", "value5"));
ASSERT_OK(Put("key6", "value6"));
ASSERT_OK(Flush());
// Each write will be similated as corrupted.
// Since the file system returns IOStatus::Corruption, it is mapped to
// kFatalError error.
fault_fs->SetChecksumHandoffFuncType(ChecksumType::kCRC32c);
SyncPoint::GetInstance()->SetCallBack(
"VersionSet::LogAndApply:WriteManifest",
[&](void*) { fault_fs->IngestDataCorruptionBeforeWrite(); });
ASSERT_OK(Put("key7", "value7"));
ASSERT_OK(Put("key8", "value8"));
SyncPoint::GetInstance()->EnableProcessing();
Status s = Flush();
ASSERT_EQ(s.severity(), ROCKSDB_NAMESPACE::Status::Severity::kFatalError);
SyncPoint::GetInstance()->DisableProcessing();
Destroy(options);
}
TEST_F(DBFlushTest, PickRightMemtables) {
Options options = CurrentOptions();
DestroyAndReopen(options);
options.create_if_missing = true;
const std::string test_cf_name = "test_cf";
options.max_write_buffer_number = 128;
CreateColumnFamilies({test_cf_name}, options);
Close();
ReopenWithColumnFamilies({kDefaultColumnFamilyName, test_cf_name}, options);
ASSERT_OK(db_->Put(WriteOptions(), "key", "value"));
ASSERT_OK(db_->Put(WriteOptions(), handles_[1], "key", "value"));
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::SyncClosedLogs:BeforeReLock", [&](void* /*arg*/) {
ASSERT_OK(db_->Put(WriteOptions(), handles_[1], "what", "v"));
auto* cfhi =
static_cast_with_check<ColumnFamilyHandleImpl>(handles_[1]);
assert(cfhi);
ASSERT_OK(dbfull()->TEST_SwitchMemtable(cfhi->cfd()));
});
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::FlushMemTableToOutputFile:AfterPickMemtables", [&](void* arg) {
auto* job = reinterpret_cast<FlushJob*>(arg);
assert(job);
const auto& mems = job->GetMemTables();
assert(mems.size() == 1);
assert(mems[0]);
ASSERT_EQ(1, mems[0]->GetID());
});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(db_->Flush(FlushOptions(), handles_[1]));
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
}
class DBFlushTestBlobError : public DBFlushTest,
public testing::WithParamInterface<std::string> {
public:
DBFlushTestBlobError() : sync_point_(GetParam()) {}
std::string sync_point_;
};
INSTANTIATE_TEST_CASE_P(DBFlushTestBlobError, DBFlushTestBlobError,
::testing::ValuesIn(std::vector<std::string>{
"BlobFileBuilder::WriteBlobToFile:AddRecord",
"BlobFileBuilder::WriteBlobToFile:AppendFooter"}));
TEST_P(DBFlushTestBlobError, FlushError) {
Options options;
options.enable_blob_files = true;
options.disable_auto_compactions = true;
options.env = env_;
Reopen(options);
ASSERT_OK(Put("key", "blob"));
SyncPoint::GetInstance()->SetCallBack(sync_point_, [this](void* arg) {
Status* const s = static_cast<Status*>(arg);
assert(s);
(*s) = Status::IOError(sync_point_);
});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_NOK(Flush());
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
VersionSet* const versions = dbfull()->GetVersionSet();
assert(versions);
ColumnFamilyData* const cfd = versions->GetColumnFamilySet()->GetDefault();
assert(cfd);
Version* const current = cfd->current();
assert(current);
const VersionStorageInfo* const storage_info = current->storage_info();
assert(storage_info);
const auto& l0_files = storage_info->LevelFiles(0);
ASSERT_TRUE(l0_files.empty());
const auto& blob_files = storage_info->GetBlobFiles();
ASSERT_TRUE(blob_files.empty());
// Make sure the files generated by the failed job have been deleted
std::vector<std::string> files;
ASSERT_OK(env_->GetChildren(dbname_, &files));
for (const auto& file : files) {
uint64_t number = 0;
FileType type = kTableFile;
if (!ParseFileName(file, &number, &type)) {
continue;
}
ASSERT_NE(type, kTableFile);
ASSERT_NE(type, kBlobFile);
}
#ifndef ROCKSDB_LITE
const InternalStats* const internal_stats = cfd->internal_stats();
assert(internal_stats);
const auto& compaction_stats = internal_stats->TEST_GetCompactionStats();
ASSERT_FALSE(compaction_stats.empty());
if (sync_point_ == "BlobFileBuilder::WriteBlobToFile:AddRecord") {
ASSERT_EQ(compaction_stats[0].bytes_written, 0);
ASSERT_EQ(compaction_stats[0].bytes_written_blob, 0);
ASSERT_EQ(compaction_stats[0].num_output_files, 0);
ASSERT_EQ(compaction_stats[0].num_output_files_blob, 0);
} else {
// SST file writing succeeded; blob file writing failed (during Finish)
ASSERT_GT(compaction_stats[0].bytes_written, 0);
ASSERT_EQ(compaction_stats[0].bytes_written_blob, 0);
ASSERT_EQ(compaction_stats[0].num_output_files, 1);
ASSERT_EQ(compaction_stats[0].num_output_files_blob, 0);
}
const uint64_t* const cf_stats_value = internal_stats->TEST_GetCFStatsValue();
ASSERT_EQ(cf_stats_value[InternalStats::BYTES_FLUSHED],
compaction_stats[0].bytes_written +
compaction_stats[0].bytes_written_blob);
#endif // ROCKSDB_LITE
}
#ifndef ROCKSDB_LITE
TEST_F(DBFlushTest, TombstoneVisibleInSnapshot) {
class SimpleTestFlushListener : public EventListener {
public:
explicit SimpleTestFlushListener(DBFlushTest* _test) : test_(_test) {}
~SimpleTestFlushListener() override {}
void OnFlushBegin(DB* db, const FlushJobInfo& info) override {
ASSERT_EQ(static_cast<uint32_t>(0), info.cf_id);
ASSERT_OK(db->Delete(WriteOptions(), "foo"));
snapshot_ = db->GetSnapshot();
ASSERT_OK(db->Put(WriteOptions(), "foo", "value"));
auto* dbimpl = static_cast_with_check<DBImpl>(db);
assert(dbimpl);
ColumnFamilyHandle* cfh = db->DefaultColumnFamily();
auto* cfhi = static_cast_with_check<ColumnFamilyHandleImpl>(cfh);
assert(cfhi);
ASSERT_OK(dbimpl->TEST_SwitchMemtable(cfhi->cfd()));
}
DBFlushTest* test_ = nullptr;
const Snapshot* snapshot_ = nullptr;
};
Options options = CurrentOptions();
options.create_if_missing = true;
auto* listener = new SimpleTestFlushListener(this);
options.listeners.emplace_back(listener);
DestroyAndReopen(options);
ASSERT_OK(db_->Put(WriteOptions(), "foo", "value0"));
ManagedSnapshot snapshot_guard(db_);
ColumnFamilyHandle* default_cf = db_->DefaultColumnFamily();
ASSERT_OK(db_->Flush(FlushOptions(), default_cf));
const Snapshot* snapshot = listener->snapshot_;
assert(snapshot);
ReadOptions read_opts;
read_opts.snapshot = snapshot;
// Using snapshot should not see "foo".
{
std::string value;
Status s = db_->Get(read_opts, "foo", &value);
ASSERT_TRUE(s.IsNotFound());
}
db_->ReleaseSnapshot(snapshot);
}
TEST_P(DBAtomicFlushTest, ManualFlushUnder2PC) {
Options options = CurrentOptions();
options.create_if_missing = true;
options.allow_2pc = true;
options.atomic_flush = GetParam();
// 64MB so that memtable flush won't be trigger by the small writes.
options.write_buffer_size = (static_cast<size_t>(64) << 20);
// Destroy the DB to recreate as a TransactionDB.
Close();
Destroy(options, true);
// Create a TransactionDB.
TransactionDB* txn_db = nullptr;
TransactionDBOptions txn_db_opts;
txn_db_opts.write_policy = TxnDBWritePolicy::WRITE_COMMITTED;
ASSERT_OK(TransactionDB::Open(options, txn_db_opts, dbname_, &txn_db));
ASSERT_NE(txn_db, nullptr);
db_ = txn_db;
// Create two more columns other than default CF.
std::vector<std::string> cfs = {"puppy", "kitty"};
CreateColumnFamilies(cfs, options);
ASSERT_EQ(handles_.size(), 2);
ASSERT_EQ(handles_[0]->GetName(), cfs[0]);
ASSERT_EQ(handles_[1]->GetName(), cfs[1]);
const size_t kNumCfToFlush = options.atomic_flush ? 2 : 1;
WriteOptions wopts;
TransactionOptions txn_opts;
// txn1 only prepare, but does not commit.
// The WAL containing the prepared but uncommitted data must be kept.
Transaction* txn1 = txn_db->BeginTransaction(wopts, txn_opts, nullptr);
// txn2 not only prepare, but also commit.
Transaction* txn2 = txn_db->BeginTransaction(wopts, txn_opts, nullptr);
ASSERT_NE(txn1, nullptr);
ASSERT_NE(txn2, nullptr);
for (size_t i = 0; i < kNumCfToFlush; i++) {
ASSERT_OK(txn1->Put(handles_[i], "k1", "v1"));
ASSERT_OK(txn2->Put(handles_[i], "k2", "v2"));
}
// A txn must be named before prepare.
ASSERT_OK(txn1->SetName("txn1"));
ASSERT_OK(txn2->SetName("txn2"));
// Prepare writes to WAL, but not to memtable. (WriteCommitted)
ASSERT_OK(txn1->Prepare());
ASSERT_OK(txn2->Prepare());
// Commit writes to memtable.
ASSERT_OK(txn2->Commit());
delete txn1;
delete txn2;
// There are still data in memtable not flushed.
// But since data is small enough to reside in the active memtable,
// there are no immutable memtable.
for (size_t i = 0; i < kNumCfToFlush; i++) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
ASSERT_EQ(0, cfh->cfd()->imm()->NumNotFlushed());
ASSERT_FALSE(cfh->cfd()->mem()->IsEmpty());
}
// Atomic flush memtables,
// the min log with prepared data should be written to MANIFEST.
std::vector<ColumnFamilyHandle*> cfs_to_flush(kNumCfToFlush);
for (size_t i = 0; i < kNumCfToFlush; i++) {
cfs_to_flush[i] = handles_[i];
}
ASSERT_OK(txn_db->Flush(FlushOptions(), cfs_to_flush));
// There are no remaining data in memtable after flush.
for (size_t i = 0; i < kNumCfToFlush; i++) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
ASSERT_EQ(0, cfh->cfd()->imm()->NumNotFlushed());
ASSERT_TRUE(cfh->cfd()->mem()->IsEmpty());
ASSERT_EQ(cfh->cfd()->GetFlushReason(), FlushReason::kManualFlush);
}
// The recovered min log number with prepared data should be non-zero.
// In 2pc mode, MinLogNumberToKeep returns the
// VersionSet::min_log_number_to_keep recovered from MANIFEST, if it's 0,
// it means atomic flush didn't write the min_log_number_to_keep to MANIFEST.
cfs.push_back(kDefaultColumnFamilyName);
ASSERT_OK(TryReopenWithColumnFamilies(cfs, options));
DBImpl* db_impl = reinterpret_cast<DBImpl*>(db_);
ASSERT_TRUE(db_impl->allow_2pc());
ASSERT_NE(db_impl->MinLogNumberToKeep(), 0);
}
#endif // ROCKSDB_LITE
TEST_P(DBAtomicFlushTest, ManualAtomicFlush) {
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = GetParam();
options.write_buffer_size = (static_cast<size_t>(64) << 20);
CreateAndReopenWithCF({"pikachu", "eevee"}, options);
size_t num_cfs = handles_.size();
ASSERT_EQ(3, num_cfs);
WriteOptions wopts;
wopts.disableWAL = true;
for (size_t i = 0; i != num_cfs; ++i) {
ASSERT_OK(Put(static_cast<int>(i) /*cf*/, "key", "value", wopts));
}
for (size_t i = 0; i != num_cfs; ++i) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
ASSERT_EQ(0, cfh->cfd()->imm()->NumNotFlushed());
ASSERT_FALSE(cfh->cfd()->mem()->IsEmpty());
}
std::vector<int> cf_ids;
for (size_t i = 0; i != num_cfs; ++i) {
cf_ids.emplace_back(static_cast<int>(i));
}
ASSERT_OK(Flush(cf_ids));
for (size_t i = 0; i != num_cfs; ++i) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
ASSERT_EQ(cfh->cfd()->GetFlushReason(), FlushReason::kManualFlush);
ASSERT_EQ(0, cfh->cfd()->imm()->NumNotFlushed());
ASSERT_TRUE(cfh->cfd()->mem()->IsEmpty());
}
}
TEST_P(DBAtomicFlushTest, PrecomputeMinLogNumberToKeepNon2PC) {
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = GetParam();
options.write_buffer_size = (static_cast<size_t>(64) << 20);
CreateAndReopenWithCF({"pikachu"}, options);
const size_t num_cfs = handles_.size();
ASSERT_EQ(num_cfs, 2);
WriteOptions wopts;
for (size_t i = 0; i != num_cfs; ++i) {
ASSERT_OK(Put(static_cast<int>(i) /*cf*/, "key", "value", wopts));
}
{
// Flush the default CF only.
std::vector<int> cf_ids{0};
ASSERT_OK(Flush(cf_ids));
autovector<ColumnFamilyData*> flushed_cfds;
autovector<autovector<VersionEdit*>> flush_edits;
auto flushed_cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[0]);
flushed_cfds.push_back(flushed_cfh->cfd());
flush_edits.push_back({});
auto unflushed_cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[1]);
ASSERT_EQ(PrecomputeMinLogNumberToKeepNon2PC(dbfull()->GetVersionSet(),
flushed_cfds, flush_edits),
unflushed_cfh->cfd()->GetLogNumber());
}
{
// Flush all CFs.
std::vector<int> cf_ids;
for (size_t i = 0; i != num_cfs; ++i) {
cf_ids.emplace_back(static_cast<int>(i));
}
ASSERT_OK(Flush(cf_ids));
uint64_t log_num_after_flush = dbfull()->TEST_GetCurrentLogNumber();
uint64_t min_log_number_to_keep = std::numeric_limits<uint64_t>::max();
autovector<ColumnFamilyData*> flushed_cfds;
autovector<autovector<VersionEdit*>> flush_edits;
for (size_t i = 0; i != num_cfs; ++i) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
flushed_cfds.push_back(cfh->cfd());
flush_edits.push_back({});
min_log_number_to_keep =
std::min(min_log_number_to_keep, cfh->cfd()->GetLogNumber());
}
ASSERT_EQ(min_log_number_to_keep, log_num_after_flush);
ASSERT_EQ(PrecomputeMinLogNumberToKeepNon2PC(dbfull()->GetVersionSet(),
flushed_cfds, flush_edits),
min_log_number_to_keep);
}
}
TEST_P(DBAtomicFlushTest, AtomicFlushTriggeredByMemTableFull) {
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = GetParam();
// 4KB so that we can easily trigger auto flush.
options.write_buffer_size = 4096;
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::BackgroundCallFlush:FlushFinish:0",
"DBAtomicFlushTest::AtomicFlushTriggeredByMemTableFull:BeforeCheck"}});
SyncPoint::GetInstance()->EnableProcessing();
CreateAndReopenWithCF({"pikachu", "eevee"}, options);
size_t num_cfs = handles_.size();
ASSERT_EQ(3, num_cfs);
WriteOptions wopts;
wopts.disableWAL = true;
for (size_t i = 0; i != num_cfs; ++i) {
ASSERT_OK(Put(static_cast<int>(i) /*cf*/, "key", "value", wopts));
}
// Keep writing to one of them column families to trigger auto flush.
for (int i = 0; i != 4000; ++i) {
ASSERT_OK(Put(static_cast<int>(num_cfs) - 1 /*cf*/,
"key" + std::to_string(i), "value" + std::to_string(i),
wopts));
}
TEST_SYNC_POINT(
"DBAtomicFlushTest::AtomicFlushTriggeredByMemTableFull:BeforeCheck");
if (options.atomic_flush) {
for (size_t i = 0; i + 1 != num_cfs; ++i) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
ASSERT_EQ(0, cfh->cfd()->imm()->NumNotFlushed());
ASSERT_TRUE(cfh->cfd()->mem()->IsEmpty());
}
} else {
for (size_t i = 0; i + 1 != num_cfs; ++i) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
ASSERT_EQ(0, cfh->cfd()->imm()->NumNotFlushed());
ASSERT_FALSE(cfh->cfd()->mem()->IsEmpty());
}
}
SyncPoint::GetInstance()->DisableProcessing();
}
TEST_P(DBAtomicFlushTest, AtomicFlushRollbackSomeJobs) {
bool atomic_flush = GetParam();
if (!atomic_flush) {
return;
}
std::unique_ptr<FaultInjectionTestEnv> fault_injection_env(
new FaultInjectionTestEnv(env_));
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = atomic_flush;
options.env = fault_injection_env.get();
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::AtomicFlushMemTablesToOutputFiles:SomeFlushJobsComplete:1",
"DBAtomicFlushTest::AtomicFlushRollbackSomeJobs:1"},
{"DBAtomicFlushTest::AtomicFlushRollbackSomeJobs:2",
"DBImpl::AtomicFlushMemTablesToOutputFiles:SomeFlushJobsComplete:2"}});
SyncPoint::GetInstance()->EnableProcessing();
CreateAndReopenWithCF({"pikachu", "eevee"}, options);
size_t num_cfs = handles_.size();
ASSERT_EQ(3, num_cfs);
WriteOptions wopts;
wopts.disableWAL = true;
for (size_t i = 0; i != num_cfs; ++i) {
int cf_id = static_cast<int>(i);
ASSERT_OK(Put(cf_id, "key", "value", wopts));
}
FlushOptions flush_opts;
flush_opts.wait = false;
ASSERT_OK(dbfull()->Flush(flush_opts, handles_));
TEST_SYNC_POINT("DBAtomicFlushTest::AtomicFlushRollbackSomeJobs:1");
fault_injection_env->SetFilesystemActive(false);
TEST_SYNC_POINT("DBAtomicFlushTest::AtomicFlushRollbackSomeJobs:2");
for (auto* cfh : handles_) {
// Returns the IO error happend during flush.
ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable(cfh));
}
for (size_t i = 0; i != num_cfs; ++i) {
auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);
ASSERT_EQ(1, cfh->cfd()->imm()->NumNotFlushed());
ASSERT_TRUE(cfh->cfd()->mem()->IsEmpty());
}
fault_injection_env->SetFilesystemActive(true);
Destroy(options);
}
TEST_P(DBAtomicFlushTest, FlushMultipleCFs_DropSomeBeforeRequestFlush) {
bool atomic_flush = GetParam();
if (!atomic_flush) {
return;
}
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = atomic_flush;
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
SyncPoint::GetInstance()->EnableProcessing();
CreateAndReopenWithCF({"pikachu", "eevee"}, options);
size_t num_cfs = handles_.size();
ASSERT_EQ(3, num_cfs);
WriteOptions wopts;
wopts.disableWAL = true;
std::vector<int> cf_ids;
for (size_t i = 0; i != num_cfs; ++i) {
int cf_id = static_cast<int>(i);
ASSERT_OK(Put(cf_id, "key", "value", wopts));
cf_ids.push_back(cf_id);
}
ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
ASSERT_TRUE(Flush(cf_ids).IsColumnFamilyDropped());
Destroy(options);
}
TEST_P(DBAtomicFlushTest,
FlushMultipleCFs_DropSomeAfterScheduleFlushBeforeFlushJobRun) {
bool atomic_flush = GetParam();
if (!atomic_flush) {
return;
}
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = atomic_flush;
CreateAndReopenWithCF({"pikachu", "eevee"}, options);
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::AtomicFlushMemTables:AfterScheduleFlush",
"DBAtomicFlushTest::BeforeDropCF"},
{"DBAtomicFlushTest::AfterDropCF",
"DBImpl::BackgroundCallFlush:start"}});
SyncPoint::GetInstance()->EnableProcessing();
size_t num_cfs = handles_.size();
ASSERT_EQ(3, num_cfs);
WriteOptions wopts;
wopts.disableWAL = true;
for (size_t i = 0; i != num_cfs; ++i) {
int cf_id = static_cast<int>(i);
ASSERT_OK(Put(cf_id, "key", "value", wopts));
}
port::Thread user_thread([&]() {
TEST_SYNC_POINT("DBAtomicFlushTest::BeforeDropCF");
ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
TEST_SYNC_POINT("DBAtomicFlushTest::AfterDropCF");
});
FlushOptions flush_opts;
flush_opts.wait = true;
ASSERT_OK(dbfull()->Flush(flush_opts, handles_));
user_thread.join();
for (size_t i = 0; i != num_cfs; ++i) {
int cf_id = static_cast<int>(i);
ASSERT_EQ("value", Get(cf_id, "key"));
}
ReopenWithColumnFamilies({kDefaultColumnFamilyName, "eevee"}, options);
num_cfs = handles_.size();
ASSERT_EQ(2, num_cfs);
for (size_t i = 0; i != num_cfs; ++i) {
int cf_id = static_cast<int>(i);
ASSERT_EQ("value", Get(cf_id, "key"));
}
Destroy(options);
}
TEST_P(DBAtomicFlushTest, TriggerFlushAndClose) {
bool atomic_flush = GetParam();
if (!atomic_flush) {
return;
}
const int kNumKeysTriggerFlush = 4;
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = atomic_flush;
options.memtable_factory.reset(
test::NewSpecialSkipListFactory(kNumKeysTriggerFlush));
CreateAndReopenWithCF({"pikachu"}, options);
for (int i = 0; i != kNumKeysTriggerFlush; ++i) {
ASSERT_OK(Put(0, "key" + std::to_string(i), "value" + std::to_string(i)));
}
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(Put(0, "key", "value"));
Close();
ReopenWithColumnFamilies({kDefaultColumnFamilyName, "pikachu"}, options);
ASSERT_EQ("value", Get(0, "key"));
}
TEST_P(DBAtomicFlushTest, PickMemtablesRaceWithBackgroundFlush) {
bool atomic_flush = GetParam();
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = atomic_flush;
options.max_write_buffer_number = 4;
// Set min_write_buffer_number_to_merge to be greater than 1, so that
// a column family with one memtable in the imm will not cause IsFlushPending
// to return true when flush_requested_ is false.
options.min_write_buffer_number_to_merge = 2;
CreateAndReopenWithCF({"pikachu"}, options);
ASSERT_EQ(2, handles_.size());
ASSERT_OK(dbfull()->PauseBackgroundWork());
ASSERT_OK(Put(0, "key00", "value00"));
ASSERT_OK(Put(1, "key10", "value10"));
FlushOptions flush_opts;
flush_opts.wait = false;
ASSERT_OK(dbfull()->Flush(flush_opts, handles_));
ASSERT_OK(Put(0, "key01", "value01"));
// Since max_write_buffer_number is 4, the following flush won't cause write
// stall.
ASSERT_OK(dbfull()->Flush(flush_opts));
ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
ASSERT_OK(dbfull()->DestroyColumnFamilyHandle(handles_[1]));
handles_[1] = nullptr;
ASSERT_OK(dbfull()->ContinueBackgroundWork());
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[0]));
delete handles_[0];
handles_.clear();
}
TEST_P(DBAtomicFlushTest, CFDropRaceWithWaitForFlushMemTables) {
bool atomic_flush = GetParam();
if (!atomic_flush) {
return;
}
Options options = CurrentOptions();
options.create_if_missing = true;
options.atomic_flush = atomic_flush;
CreateAndReopenWithCF({"pikachu"}, options);
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::AtomicFlushMemTables:AfterScheduleFlush",
"DBAtomicFlushTest::CFDropRaceWithWaitForFlushMemTables:BeforeDrop"},
{"DBAtomicFlushTest::CFDropRaceWithWaitForFlushMemTables:AfterFree",
"DBImpl::BackgroundCallFlush:start"},
{"DBImpl::BackgroundCallFlush:start",
"DBImpl::AtomicFlushMemTables:BeforeWaitForBgFlush"}});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_EQ(2, handles_.size());
ASSERT_OK(Put(0, "key", "value"));
ASSERT_OK(Put(1, "key", "value"));
auto* cfd_default =
static_cast<ColumnFamilyHandleImpl*>(dbfull()->DefaultColumnFamily())
->cfd();
auto* cfd_pikachu = static_cast<ColumnFamilyHandleImpl*>(handles_[1])->cfd();
port::Thread drop_cf_thr([&]() {
TEST_SYNC_POINT(
"DBAtomicFlushTest::CFDropRaceWithWaitForFlushMemTables:BeforeDrop");
ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
delete handles_[1];
handles_.resize(1);
TEST_SYNC_POINT(
"DBAtomicFlushTest::CFDropRaceWithWaitForFlushMemTables:AfterFree");
});
FlushOptions flush_opts;
flush_opts.allow_write_stall = true;
ASSERT_OK(dbfull()->TEST_AtomicFlushMemTables({cfd_default, cfd_pikachu},
flush_opts));
drop_cf_thr.join();
Close();
SyncPoint::GetInstance()->DisableProcessing();
}
TEST_P(DBAtomicFlushTest, RollbackAfterFailToInstallResults) {
bool atomic_flush = GetParam();
if (!atomic_flush) {
return;
}
auto fault_injection_env = std::make_shared<FaultInjectionTestEnv>(env_);
Options options = CurrentOptions();
options.env = fault_injection_env.get();
options.create_if_missing = true;
options.atomic_flush = atomic_flush;
CreateAndReopenWithCF({"pikachu"}, options);
ASSERT_EQ(2, handles_.size());
for (size_t cf = 0; cf < handles_.size(); ++cf) {
ASSERT_OK(Put(static_cast<int>(cf), "a", "value"));
}
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
SyncPoint::GetInstance()->SetCallBack(
"VersionSet::ProcessManifestWrites:BeforeWriteLastVersionEdit:0",
[&](void* /*arg*/) { fault_injection_env->SetFilesystemActive(false); });
SyncPoint::GetInstance()->EnableProcessing();
FlushOptions flush_opts;
Status s = db_->Flush(flush_opts, handles_);
ASSERT_NOK(s);
fault_injection_env->SetFilesystemActive(true);
Close();
SyncPoint::GetInstance()->ClearAllCallBacks();
}
// In atomic flush, concurrent bg flush threads commit to the MANIFEST in
// serial, in the order of their picked memtables for each column family.
// Only when a bg flush thread finds out that its memtables are the earliest
// unflushed ones for all the included column families will this bg flush
// thread continue to commit to MANIFEST.
// This unit test uses sync point to coordinate the execution of two bg threads
// executing the same sequence of functions. The interleaving are as follows.
// time bg1 bg2
// | pick memtables to flush
// | flush memtables cf1_m1, cf2_m1
// | join MANIFEST write queue
// | pick memtabls to flush
// | flush memtables cf1_(m1+1)
// | join MANIFEST write queue
// | wait to write MANIFEST
// | write MANIFEST
// | IO error
// | detect IO error and stop waiting
// V
TEST_P(DBAtomicFlushTest, BgThreadNoWaitAfterManifestError) {
bool atomic_flush = GetParam();
if (!atomic_flush) {
return;
}
auto fault_injection_env = std::make_shared<FaultInjectionTestEnv>(env_);
Options options = GetDefaultOptions();
options.create_if_missing = true;
options.atomic_flush = true;
options.env = fault_injection_env.get();
// Set a larger value than default so that RocksDB can schedule concurrent
// background flush threads.
options.max_background_jobs = 8;
options.max_write_buffer_number = 8;
CreateAndReopenWithCF({"pikachu"}, options);
assert(2 == handles_.size());
WriteOptions write_opts;
write_opts.disableWAL = true;
ASSERT_OK(Put(0, "a", "v_0_a", write_opts));
ASSERT_OK(Put(1, "a", "v_1_a", write_opts));
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
SyncPoint::GetInstance()->LoadDependency({
{"BgFlushThr2:WaitToCommit", "BgFlushThr1:BeforeWriteManifest"},
});
std::thread::id bg_flush_thr1, bg_flush_thr2;
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::BackgroundCallFlush:start", [&](void*) {
if (bg_flush_thr1 == std::thread::id()) {
bg_flush_thr1 = std::this_thread::get_id();
} else if (bg_flush_thr2 == std::thread::id()) {
bg_flush_thr2 = std::this_thread::get_id();
}
});
int called = 0;
SyncPoint::GetInstance()->SetCallBack(
"DBImpl::AtomicFlushMemTablesToOutputFiles:WaitToCommit", [&](void* arg) {
if (std::this_thread::get_id() == bg_flush_thr2) {
const auto* ptr = reinterpret_cast<std::pair<Status, bool>*>(arg);
assert(ptr);
if (0 == called) {
// When bg flush thread 2 reaches here for the first time.
ASSERT_OK(ptr->first);
ASSERT_TRUE(ptr->second);
} else if (1 == called) {
// When bg flush thread 2 reaches here for the second time.
ASSERT_TRUE(ptr->first.IsIOError());
ASSERT_FALSE(ptr->second);
}
++called;
TEST_SYNC_POINT("BgFlushThr2:WaitToCommit");
}
});
SyncPoint::GetInstance()->SetCallBack(
"VersionSet::ProcessManifestWrites:BeforeWriteLastVersionEdit:0",
[&](void*) {
if (std::this_thread::get_id() == bg_flush_thr1) {
TEST_SYNC_POINT("BgFlushThr1:BeforeWriteManifest");
}
});
SyncPoint::GetInstance()->SetCallBack(
"VersionSet::LogAndApply:WriteManifest", [&](void*) {
if (std::this_thread::get_id() != bg_flush_thr1) {
return;
}
ASSERT_OK(db_->Put(write_opts, "b", "v_1_b"));
FlushOptions flush_opts;
flush_opts.wait = false;
std::vector<ColumnFamilyHandle*> cfhs(1, db_->DefaultColumnFamily());
ASSERT_OK(dbfull()->Flush(flush_opts, cfhs));
});
SyncPoint::GetInstance()->SetCallBack(
"VersionSet::ProcessManifestWrites:AfterSyncManifest", [&](void* arg) {
auto* ptr = reinterpret_cast<IOStatus*>(arg);
assert(ptr);
*ptr = IOStatus::IOError("Injected failure");
});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_TRUE(dbfull()->Flush(FlushOptions(), handles_).IsIOError());
Close();
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->ClearAllCallBacks();
}
TEST_P(DBAtomicFlushTest, NoWaitWhenWritesStopped) {
Options options = GetDefaultOptions();
options.create_if_missing = true;
options.atomic_flush = GetParam();
options.max_write_buffer_number = 2;
options.memtable_factory.reset(test::NewSpecialSkipListFactory(1));
Reopen(options);
SyncPoint::GetInstance()->DisableProcessing();
SyncPoint::GetInstance()->LoadDependency(
{{"DBImpl::DelayWrite:Start",
"DBAtomicFlushTest::NoWaitWhenWritesStopped:0"}});
SyncPoint::GetInstance()->EnableProcessing();
ASSERT_OK(dbfull()->PauseBackgroundWork());
for (int i = 0; i < options.max_write_buffer_number; ++i) {
ASSERT_OK(Put("k" + std::to_string(i), "v" + std::to_string(i)));
}
std::thread stalled_writer([&]() { ASSERT_OK(Put("k", "v")); });
TEST_SYNC_POINT("DBAtomicFlushTest::NoWaitWhenWritesStopped:0");
{
FlushOptions flush_opts;
flush_opts.wait = false;
flush_opts.allow_write_stall = true;
ASSERT_TRUE(db_->Flush(flush_opts).IsTryAgain());
}
ASSERT_OK(dbfull()->ContinueBackgroundWork());
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
stalled_writer.join();
SyncPoint::GetInstance()->DisableProcessing();
}
INSTANTIATE_TEST_CASE_P(DBFlushDirectIOTest, DBFlushDirectIOTest,
testing::Bool());
INSTANTIATE_TEST_CASE_P(DBAtomicFlushTest, DBAtomicFlushTest, testing::Bool());
} // namespace ROCKSDB_NAMESPACE
int main(int argc, char** argv) {
ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
::testing::InitGoogleTest(&argc, argv);
return RUN_ALL_TESTS();
}