dolysis/rocksdb
3
0
Fork 0
mirror of https://github.com/facebook/rocksdb.git synced 2024-11-26 07:30:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update documentation

Browse source 
Summary:
Changed leveldb documentation with rocksdb in doc/index.html. Added some of the important options from options.h to doc.

Also removed benchmark files and impl.h, since this is all replaced by RocksDB wikis.

Test Plan: -

Reviewers: dhruba, haobo, kailiu, emayanke, sdong

Reviewed By: dhruba

CC: leveldb

Differential Revision: https://reviews.facebook.net/D13977
This commit is contained in:
Igor Canadi 2013-11-11 21:02:38 -08:00
parent 318a4919d2
commit 65e45f0c4a
7 changed files with 213 additions and 2059 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

718
doc/bench/db_bench_sqlite3.cc
View file

@ -1,718 +0,0 @@
// 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 <stdio.h>
#include <stdlib.h>
#include <sqlite3.h>
#include "util/histogram.h"
#include "util/random.h"
#include "util/testutil.h"
// Comma-separated list of operations to run in the specified order
// Actual benchmarks:
//
// fillseq -- write N values in sequential key order in async mode
// fillseqsync -- write N/100 values in sequential key order in sync mode
// fillseqbatch -- batch write N values in sequential key order in async mode
// fillrandom -- write N values in random key order in async mode
// fillrandsync -- write N/100 values in random key order in sync mode
// fillrandbatch -- batch write N values in sequential key order in async mode
// overwrite -- overwrite N values in random key order in async mode
// fillrand100K -- write N/1000 100K values in random order in async mode
// fillseq100K -- write N/1000 100K values in sequential order in async mode
// readseq -- read N times sequentially
// readrandom -- read N times in random order
// readrand100K -- read N/1000 100K values in sequential order in async mode
static const char* FLAGS_benchmarks =
"fillseq,"
"fillseqsync,"
"fillseqbatch,"
"fillrandom,"
"fillrandsync,"
"fillrandbatch,"
"overwrite,"
"overwritebatch,"
"readrandom,"
"readseq,"
"fillrand100K,"
"fillseq100K,"
"readseq,"
"readrand100K,"
;
// Number of key/values to place in database
static int FLAGS_num = 1000000;
// Number of read operations to do. If negative, do FLAGS_num reads.
static int FLAGS_reads = -1;
// Size of each value
static int FLAGS_value_size = 100;
// Print histogram of operation timings
static bool FLAGS_histogram = false;
// Arrange to generate values that shrink to this fraction of
// their original size after compression
static double FLAGS_compression_ratio = 0.5;
// Page size. Default 1 KB.
static int FLAGS_page_size = 1024;
// Number of pages.
// Default cache size = FLAGS_page_size * FLAGS_num_pages = 4 MB.
static int FLAGS_num_pages = 4096;
// If true, do not destroy the existing database. If you set this
// flag and also specify a benchmark that wants a fresh database, that
// benchmark will fail.
static bool FLAGS_use_existing_db = false;
// If true, we allow batch writes to occur
static bool FLAGS_transaction = true;
// If true, we enable Write-Ahead Logging
static bool FLAGS_WAL_enabled = true;
// Use the db with the following name.
static const char* FLAGS_db = NULL;
inline
static void ExecErrorCheck(int status, char *err_msg) {
if (status != SQLITE_OK) {
fprintf(stderr, "SQL error: %s\n", err_msg);
sqlite3_free(err_msg);
exit(1);
}
}
inline
static void StepErrorCheck(int status) {
if (status != SQLITE_DONE) {
fprintf(stderr, "SQL step error: status = %d\n", status);
exit(1);
}
}
inline
static void ErrorCheck(int status) {
if (status != SQLITE_OK) {
fprintf(stderr, "sqlite3 error: status = %d\n", status);
exit(1);
}
}
inline
static void WalCheckpoint(sqlite3* db_) {
// Flush all writes to disk
if (FLAGS_WAL_enabled) {
sqlite3_wal_checkpoint_v2(db_, NULL, SQLITE_CHECKPOINT_FULL, NULL, NULL);
}
}
namespace rocksdb {
// Helper for quickly generating random data.
namespace {
class RandomGenerator {
private:
std::string data_;
int pos_;
public:
RandomGenerator() {
// We use a limited amount of data over and over again and ensure
// that it is larger than the compression window (32KB), and also
// large enough to serve all typical value sizes we want to write.
Random rnd(301);
std::string piece;
while (data_.size() < 1048576) {
// Add a short fragment that is as compressible as specified
// by FLAGS_compression_ratio.
test::CompressibleString(&rnd, FLAGS_compression_ratio, 100, &piece);
data_.append(piece);
}
pos_ = 0;
}
Slice Generate(int len) {
if (pos_ + len > data_.size()) {
pos_ = 0;
assert(len < data_.size());
}
pos_ += len;
return Slice(data_.data() + pos_ - len, len);
}
};
static Slice TrimSpace(Slice s) {
int start = 0;
while (start < s.size() && isspace(s[start])) {
start++;
}
int limit = s.size();
while (limit > start && isspace(s[limit-1])) {
limit--;
}
return Slice(s.data() + start, limit - start);
}
} // namespace
class Benchmark {
private:
sqlite3* db_;
int db_num_;
int num_;
int reads_;
double start_;
double last_op_finish_;
int64_t bytes_;
std::string message_;
Histogram hist_;
RandomGenerator gen_;
Random rand_;
// State kept for progress messages
int done_;
int next_report_; // When to report next
void PrintHeader() {
const int kKeySize = 16;
PrintEnvironment();
fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
fprintf(stdout, "Values: %d bytes each\n", FLAGS_value_size);
fprintf(stdout, "Entries: %d\n", num_);
fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_)
/ 1048576.0));
PrintWarnings();
fprintf(stdout, "------------------------------------------------\n");
}
void PrintWarnings() {
#if defined(__GNUC__) && !defined(__OPTIMIZE__)
fprintf(stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n"
);
#endif
#ifndef NDEBUG
fprintf(stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
#endif
}
void PrintEnvironment() {
fprintf(stderr, "SQLite: version %s\n", SQLITE_VERSION);
#if defined(__linux)
time_t now = time(NULL);
fprintf(stderr, "Date: %s", ctime(&now)); // ctime() adds newline
FILE* cpuinfo = fopen("/proc/cpuinfo", "r");
if (cpuinfo != NULL) {
char line[1000];
int num_cpus = 0;
std::string cpu_type;
std::string cache_size;
while (fgets(line, sizeof(line), cpuinfo) != NULL) {
const char* sep = strchr(line, ':');
if (sep == NULL) {
continue;
}
Slice key = TrimSpace(Slice(line, sep - 1 - line));
Slice val = TrimSpace(Slice(sep + 1));
if (key == "model name") {
++num_cpus;
cpu_type = val.ToString();
} else if (key == "cache size") {
cache_size = val.ToString();
}
}
fclose(cpuinfo);
fprintf(stderr, "CPU: %d * %s\n", num_cpus, cpu_type.c_str());
fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
}
#endif
}
void Start() {
start_ = Env::Default()->NowMicros() * 1e-6;
bytes_ = 0;
message_.clear();
last_op_finish_ = start_;
hist_.Clear();
done_ = 0;
next_report_ = 100;
}
void FinishedSingleOp() {
if (FLAGS_histogram) {
double now = Env::Default()->NowMicros() * 1e-6;
double micros = (now - last_op_finish_) * 1e6;
hist_.Add(micros);
if (micros > 20000) {
fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
fflush(stderr);
}
last_op_finish_ = now;
}
done_++;
if (done_ >= next_report_) {
if (next_report_ < 1000) next_report_ += 100;
else if (next_report_ < 5000) next_report_ += 500;
else if (next_report_ < 10000) next_report_ += 1000;
else if (next_report_ < 50000) next_report_ += 5000;
else if (next_report_ < 100000) next_report_ += 10000;
else if (next_report_ < 500000) next_report_ += 50000;
else next_report_ += 100000;
fprintf(stderr, "... finished %d ops%30s\r", done_, "");
fflush(stderr);
}
}
void Stop(const Slice& name) {
double finish = Env::Default()->NowMicros() * 1e-6;
// Pretend at least one op was done in case we are running a benchmark
// that does not call FinishedSingleOp().
if (done_ < 1) done_ = 1;
if (bytes_ > 0) {
char rate[100];
snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / (finish - start_));
if (!message_.empty()) {
message_ = std::string(rate) + " " + message_;
} else {
message_ = rate;
}
}
fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(),
(finish - start_) * 1e6 / done_,
(message_.empty() ? "" : " "),
message_.c_str());
if (FLAGS_histogram) {
fprintf(stdout, "Microseconds per op:\n%s\n", hist_.ToString().c_str());
}
fflush(stdout);
}
public:
enum Order {
SEQUENTIAL,
RANDOM
};
enum DBState {
FRESH,
EXISTING
};
Benchmark()
: db_(NULL),
db_num_(0),
num_(FLAGS_num),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
bytes_(0),
rand_(301) {
std::vector<std::string> files;
std::string test_dir;
Env::Default()->GetTestDirectory(&test_dir);
Env::Default()->GetChildren(test_dir, &files);
if (!FLAGS_use_existing_db) {
for (int i = 0; i < files.size(); i++) {
if (Slice(files[i]).starts_with("dbbench_sqlite3")) {
std::string file_name(test_dir);
file_name += "/";
file_name += files[i];
Env::Default()->DeleteFile(file_name.c_str());
}
}
}
}
~Benchmark() {
int status = sqlite3_close(db_);
ErrorCheck(status);
}
void Run() {
PrintHeader();
Open();
const char* benchmarks = FLAGS_benchmarks;
while (benchmarks != NULL) {
const char* sep = strchr(benchmarks, ',');
Slice name;
if (sep == NULL) {
name = benchmarks;
benchmarks = NULL;
} else {
name = Slice(benchmarks, sep - benchmarks);
benchmarks = sep + 1;
}
bytes_ = 0;
Start();
bool known = true;
bool write_sync = false;
if (name == Slice("fillseq")) {
Write(write_sync, SEQUENTIAL, FRESH, num_, FLAGS_value_size, 1);
WalCheckpoint(db_);
} else if (name == Slice("fillseqbatch")) {
Write(write_sync, SEQUENTIAL, FRESH, num_, FLAGS_value_size, 1000);
WalCheckpoint(db_);
} else if (name == Slice("fillrandom")) {
Write(write_sync, RANDOM, FRESH, num_, FLAGS_value_size, 1);
WalCheckpoint(db_);
} else if (name == Slice("fillrandbatch")) {
Write(write_sync, RANDOM, FRESH, num_, FLAGS_value_size, 1000);
WalCheckpoint(db_);
} else if (name == Slice("overwrite")) {
Write(write_sync, RANDOM, EXISTING, num_, FLAGS_value_size, 1);
WalCheckpoint(db_);
} else if (name == Slice("overwritebatch")) {
Write(write_sync, RANDOM, EXISTING, num_, FLAGS_value_size, 1000);
WalCheckpoint(db_);
} else if (name == Slice("fillrandsync")) {
write_sync = true;
Write(write_sync, RANDOM, FRESH, num_ / 100, FLAGS_value_size, 1);
WalCheckpoint(db_);
} else if (name == Slice("fillseqsync")) {
write_sync = true;
Write(write_sync, SEQUENTIAL, FRESH, num_ / 100, FLAGS_value_size, 1);
WalCheckpoint(db_);
} else if (name == Slice("fillrand100K")) {
Write(write_sync, RANDOM, FRESH, num_ / 1000, 100 * 1000, 1);
WalCheckpoint(db_);
} else if (name == Slice("fillseq100K")) {
Write(write_sync, SEQUENTIAL, FRESH, num_ / 1000, 100 * 1000, 1);
WalCheckpoint(db_);
} else if (name == Slice("readseq")) {
ReadSequential();
} else if (name == Slice("readrandom")) {
Read(RANDOM, 1);
} else if (name == Slice("readrand100K")) {
int n = reads_;
reads_ /= 1000;
Read(RANDOM, 1);
reads_ = n;
} else {
known = false;
if (name != Slice()) { // No error message for empty name
fprintf(stderr, "unknown benchmark '%s'\n", name.ToString().c_str());
}
}
if (known) {
Stop(name);
}
}
}
void Open() {
assert(db_ == NULL);
int status;
char file_name[100];
char* err_msg = NULL;
db_num_++;
// Open database
std::string tmp_dir;
Env::Default()->GetTestDirectory(&tmp_dir);
snprintf(file_name, sizeof(file_name),
"%s/dbbench_sqlite3-%d.db",
tmp_dir.c_str(),
db_num_);
status = sqlite3_open(file_name, &db_);
if (status) {
fprintf(stderr, "open error: %s\n", sqlite3_errmsg(db_));
exit(1);
}
// Change SQLite cache size
char cache_size[100];
snprintf(cache_size, sizeof(cache_size), "PRAGMA cache_size = %d",
FLAGS_num_pages);
status = sqlite3_exec(db_, cache_size, NULL, NULL, &err_msg);
ExecErrorCheck(status, err_msg);
// FLAGS_page_size is defaulted to 1024
if (FLAGS_page_size != 1024) {
char page_size[100];
snprintf(page_size, sizeof(page_size), "PRAGMA page_size = %d",
FLAGS_page_size);
status = sqlite3_exec(db_, page_size, NULL, NULL, &err_msg);
ExecErrorCheck(status, err_msg);
}
// Change journal mode to WAL if WAL enabled flag is on
if (FLAGS_WAL_enabled) {
std::string WAL_stmt = "PRAGMA journal_mode = WAL";
// LevelDB's default cache size is a combined 4 MB
std::string WAL_checkpoint = "PRAGMA wal_autocheckpoint = 4096";
status = sqlite3_exec(db_, WAL_stmt.c_str(), NULL, NULL, &err_msg);
ExecErrorCheck(status, err_msg);
status = sqlite3_exec(db_, WAL_checkpoint.c_str(), NULL, NULL, &err_msg);
ExecErrorCheck(status, err_msg);
}
// Change locking mode to exclusive and create tables/index for database
std::string locking_stmt = "PRAGMA locking_mode = EXCLUSIVE";
std::string create_stmt =
"CREATE TABLE test (key blob, value blob, PRIMARY KEY(key))";
std::string stmt_array[] = { locking_stmt, create_stmt };
int stmt_array_length = sizeof(stmt_array) / sizeof(std::string);
for (int i = 0; i < stmt_array_length; i++) {
status = sqlite3_exec(db_, stmt_array[i].c_str(), NULL, NULL, &err_msg);
ExecErrorCheck(status, err_msg);
}
}
void Write(bool write_sync, Order order, DBState state,
int num_entries, int value_size, int entries_per_batch) {
// Create new database if state == FRESH
if (state == FRESH) {
if (FLAGS_use_existing_db) {
message_ = "skipping (--use_existing_db is true)";
return;
}
sqlite3_close(db_);
db_ = NULL;
Open();
Start();
}
if (num_entries != num_) {
char msg[100];
snprintf(msg, sizeof(msg), "(%d ops)", num_entries);
message_ = msg;
}
char* err_msg = NULL;
int status;
sqlite3_stmt *replace_stmt, *begin_trans_stmt, *end_trans_stmt;
std::string replace_str = "REPLACE INTO test (key, value) VALUES (?, ?)";
std::string begin_trans_str = "BEGIN TRANSACTION;";
std::string end_trans_str = "END TRANSACTION;";
// Check for synchronous flag in options
std::string sync_stmt = (write_sync) ? "PRAGMA synchronous = FULL" :
"PRAGMA synchronous = OFF";
status = sqlite3_exec(db_, sync_stmt.c_str(), NULL, NULL, &err_msg);
ExecErrorCheck(status, err_msg);
// Preparing sqlite3 statements
status = sqlite3_prepare_v2(db_, replace_str.c_str(), -1,
&replace_stmt, NULL);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, begin_trans_str.c_str(), -1,
&begin_trans_stmt, NULL);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, end_trans_str.c_str(), -1,
&end_trans_stmt, NULL);
ErrorCheck(status);
bool transaction = (entries_per_batch > 1);
for (int i = 0; i < num_entries; i += entries_per_batch) {
// Begin write transaction
if (FLAGS_transaction && transaction) {
status = sqlite3_step(begin_trans_stmt);
StepErrorCheck(status);
status = sqlite3_reset(begin_trans_stmt);
ErrorCheck(status);
}
// Create and execute SQL statements
for (int j = 0; j < entries_per_batch; j++) {
const char* value = gen_.Generate(value_size).data();
// Create values for key-value pair
const int k = (order == SEQUENTIAL) ? i + j :
(rand_.Next() % num_entries);
char key[100];
snprintf(key, sizeof(key), "%016d", k);
// Bind KV values into replace_stmt
status = sqlite3_bind_blob(replace_stmt, 1, key, 16, SQLITE_STATIC);
ErrorCheck(status);
status = sqlite3_bind_blob(replace_stmt, 2, value,
value_size, SQLITE_STATIC);
ErrorCheck(status);
// Execute replace_stmt
bytes_ += value_size + strlen(key);
status = sqlite3_step(replace_stmt);
StepErrorCheck(status);
// Reset SQLite statement for another use
status = sqlite3_clear_bindings(replace_stmt);
ErrorCheck(status);
status = sqlite3_reset(replace_stmt);
ErrorCheck(status);
FinishedSingleOp();
}
// End write transaction
if (FLAGS_transaction && transaction) {
status = sqlite3_step(end_trans_stmt);
StepErrorCheck(status);
status = sqlite3_reset(end_trans_stmt);
ErrorCheck(status);
}
}
status = sqlite3_finalize(replace_stmt);
ErrorCheck(status);
status = sqlite3_finalize(begin_trans_stmt);
ErrorCheck(status);
status = sqlite3_finalize(end_trans_stmt);
ErrorCheck(status);
}
void Read(Order order, int entries_per_batch) {
int status;
sqlite3_stmt *read_stmt, *begin_trans_stmt, *end_trans_stmt;
std::string read_str = "SELECT * FROM test WHERE key = ?";
std::string begin_trans_str = "BEGIN TRANSACTION;";
std::string end_trans_str = "END TRANSACTION;";
// Preparing sqlite3 statements
status = sqlite3_prepare_v2(db_, begin_trans_str.c_str(), -1,
&begin_trans_stmt, NULL);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, end_trans_str.c_str(), -1,
&end_trans_stmt, NULL);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, read_str.c_str(), -1, &read_stmt, NULL);
ErrorCheck(status);
bool transaction = (entries_per_batch > 1);
for (int i = 0; i < reads_; i += entries_per_batch) {
// Begin read transaction
if (FLAGS_transaction && transaction) {
status = sqlite3_step(begin_trans_stmt);
StepErrorCheck(status);
status = sqlite3_reset(begin_trans_stmt);
ErrorCheck(status);
}
// Create and execute SQL statements
for (int j = 0; j < entries_per_batch; j++) {
// Create key value
char key[100];
int k = (order == SEQUENTIAL) ? i + j : (rand_.Next() % reads_);
snprintf(key, sizeof(key), "%016d", k);
// Bind key value into read_stmt
status = sqlite3_bind_blob(read_stmt, 1, key, 16, SQLITE_STATIC);
ErrorCheck(status);
// Execute read statement
while ((status = sqlite3_step(read_stmt)) == SQLITE_ROW);
StepErrorCheck(status);
// Reset SQLite statement for another use
status = sqlite3_clear_bindings(read_stmt);
ErrorCheck(status);
status = sqlite3_reset(read_stmt);
ErrorCheck(status);
FinishedSingleOp();
}
// End read transaction
if (FLAGS_transaction && transaction) {
status = sqlite3_step(end_trans_stmt);
StepErrorCheck(status);
status = sqlite3_reset(end_trans_stmt);
ErrorCheck(status);
}
}
status = sqlite3_finalize(read_stmt);
ErrorCheck(status);
status = sqlite3_finalize(begin_trans_stmt);
ErrorCheck(status);
status = sqlite3_finalize(end_trans_stmt);
ErrorCheck(status);
}
void ReadSequential() {
int status;
sqlite3_stmt *pStmt;
std::string read_str = "SELECT * FROM test ORDER BY key";
status = sqlite3_prepare_v2(db_, read_str.c_str(), -1, &pStmt, NULL);
ErrorCheck(status);
for (int i = 0; i < reads_ && SQLITE_ROW == sqlite3_step(pStmt); i++) {
bytes_ += sqlite3_column_bytes(pStmt, 1) + sqlite3_column_bytes(pStmt, 2);
FinishedSingleOp();
}
status = sqlite3_finalize(pStmt);
ErrorCheck(status);
}
};
} // namespace rocksdb
int main(int argc, char** argv) {
std::string default_db_path;
for (int i = 1; i < argc; i++) {
double d;
int n;
char junk;
if (rocksdb::Slice(argv[i]).starts_with("--benchmarks=")) {
FLAGS_benchmarks = argv[i] + strlen("--benchmarks=");
} else if (sscanf(argv[i], "--histogram=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_histogram = n;
} else if (sscanf(argv[i], "--compression_ratio=%lf%c", &d, &junk) == 1) {
FLAGS_compression_ratio = d;
} else if (sscanf(argv[i], "--use_existing_db=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_use_existing_db = n;
} else if (sscanf(argv[i], "--num=%d%c", &n, &junk) == 1) {
FLAGS_num = n;
} else if (sscanf(argv[i], "--reads=%d%c", &n, &junk) == 1) {
FLAGS_reads = n;
} else if (sscanf(argv[i], "--value_size=%d%c", &n, &junk) == 1) {
FLAGS_value_size = n;
} else if (rocksdb::Slice(argv[i]) == rocksdb::Slice("--no_transaction")) {
FLAGS_transaction = false;
} else if (sscanf(argv[i], "--page_size=%d%c", &n, &junk) == 1) {
FLAGS_page_size = n;
} else if (sscanf(argv[i], "--num_pages=%d%c", &n, &junk) == 1) {
FLAGS_num_pages = n;
} else if (sscanf(argv[i], "--WAL_enabled=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_WAL_enabled = n;
} else if (strncmp(argv[i], "--db=", 5) == 0) {
FLAGS_db = argv[i] + 5;
} else {
fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
exit(1);
}
}
// Choose a location for the test database if none given with --db=<path>
if (FLAGS_db == NULL) {
rocksdb::Env::Default()->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
}
rocksdb::Benchmark benchmark;
benchmark.Run();
return 0;
}

528
doc/bench/db_bench_tree_db.cc
View file

@ -1,528 +0,0 @@
// 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 <stdio.h>
#include <stdlib.h>
#include <kcpolydb.h>
#include "util/histogram.h"
#include "util/random.h"
#include "util/testutil.h"
// Comma-separated list of operations to run in the specified order
// Actual benchmarks:
//
// fillseq -- write N values in sequential key order in async mode
// fillrandom -- write N values in random key order in async mode
// overwrite -- overwrite N values in random key order in async mode
// fillseqsync -- write N/100 values in sequential key order in sync mode
// fillrandsync -- write N/100 values in random key order in sync mode
// fillrand100K -- write N/1000 100K values in random order in async mode
// fillseq100K -- write N/1000 100K values in seq order in async mode
// readseq -- read N times sequentially
// readseq100K -- read N/1000 100K values in sequential order in async mode
// readrand100K -- read N/1000 100K values in sequential order in async mode
// readrandom -- read N times in random order
static const char* FLAGS_benchmarks =
"fillseq,"
"fillseqsync,"
"fillrandsync,"
"fillrandom,"
"overwrite,"
"readrandom,"
"readseq,"
"fillrand100K,"
"fillseq100K,"
"readseq100K,"
"readrand100K,"
;
// Number of key/values to place in database
static int FLAGS_num = 1000000;
// Number of read operations to do. If negative, do FLAGS_num reads.
static int FLAGS_reads = -1;
// Size of each value
static int FLAGS_value_size = 100;
// Arrange to generate values that shrink to this fraction of
// their original size after compression
static double FLAGS_compression_ratio = 0.5;
// Print histogram of operation timings
static bool FLAGS_histogram = false;
// Cache size. Default 4 MB
static int FLAGS_cache_size = 4194304;
// Page size. Default 1 KB
static int FLAGS_page_size = 1024;
// If true, do not destroy the existing database. If you set this
// flag and also specify a benchmark that wants a fresh database, that
// benchmark will fail.
static bool FLAGS_use_existing_db = false;
// Compression flag. If true, compression is on. If false, compression
// is off.
static bool FLAGS_compression = true;
// Use the db with the following name.
static const char* FLAGS_db = NULL;
inline
static void DBSynchronize(kyotocabinet::TreeDB* db_)
{
// Synchronize will flush writes to disk
if (!db_->synchronize()) {
fprintf(stderr, "synchronize error: %s\n", db_->error().name());
}
}
namespace rocksdb {
// Helper for quickly generating random data.
namespace {
class RandomGenerator {
private:
std::string data_;
int pos_;
public:
RandomGenerator() {
// We use a limited amount of data over and over again and ensure
// that it is larger than the compression window (32KB), and also
// large enough to serve all typical value sizes we want to write.
Random rnd(301);
std::string piece;
while (data_.size() < 1048576) {
// Add a short fragment that is as compressible as specified
// by FLAGS_compression_ratio.
test::CompressibleString(&rnd, FLAGS_compression_ratio, 100, &piece);
data_.append(piece);
}
pos_ = 0;
}
Slice Generate(int len) {
if (pos_ + len > data_.size()) {
pos_ = 0;
assert(len < data_.size());
}
pos_ += len;
return Slice(data_.data() + pos_ - len, len);
}
};
static Slice TrimSpace(Slice s) {
int start = 0;
while (start < s.size() && isspace(s[start])) {
start++;
}
int limit = s.size();
while (limit > start && isspace(s[limit-1])) {
limit--;
}
return Slice(s.data() + start, limit - start);
}
} // namespace
class Benchmark {
private:
kyotocabinet::TreeDB* db_;
int db_num_;
int num_;
int reads_;
double start_;
double last_op_finish_;
int64_t bytes_;
std::string message_;
Histogram hist_;
RandomGenerator gen_;
Random rand_;
kyotocabinet::LZOCompressor<kyotocabinet::LZO::RAW> comp_;
// State kept for progress messages
int done_;
int next_report_; // When to report next
void PrintHeader() {
const int kKeySize = 16;
PrintEnvironment();
fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
fprintf(stdout, "Values: %d bytes each (%d bytes after compression)\n",
FLAGS_value_size,
static_cast<int>(FLAGS_value_size * FLAGS_compression_ratio + 0.5));
fprintf(stdout, "Entries: %d\n", num_);
fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_)
/ 1048576.0));
fprintf(stdout, "FileSize: %.1f MB (estimated)\n",
(((kKeySize + FLAGS_value_size * FLAGS_compression_ratio) * num_)
/ 1048576.0));
PrintWarnings();
fprintf(stdout, "------------------------------------------------\n");
}
void PrintWarnings() {
#if defined(__GNUC__) && !defined(__OPTIMIZE__)
fprintf(stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n"
);
#endif
#ifndef NDEBUG
fprintf(stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
#endif
}
void PrintEnvironment() {
fprintf(stderr, "Kyoto Cabinet: version %s, lib ver %d, lib rev %d\n",
kyotocabinet::VERSION, kyotocabinet::LIBVER, kyotocabinet::LIBREV);
#if defined(__linux)
time_t now = time(NULL);
fprintf(stderr, "Date: %s", ctime(&now)); // ctime() adds newline
FILE* cpuinfo = fopen("/proc/cpuinfo", "r");
if (cpuinfo != NULL) {
char line[1000];
int num_cpus = 0;
std::string cpu_type;
std::string cache_size;
while (fgets(line, sizeof(line), cpuinfo) != NULL) {
const char* sep = strchr(line, ':');
if (sep == NULL) {
continue;
}
Slice key = TrimSpace(Slice(line, sep - 1 - line));
Slice val = TrimSpace(Slice(sep + 1));
if (key == "model name") {
++num_cpus;
cpu_type = val.ToString();
} else if (key == "cache size") {
cache_size = val.ToString();
}
}
fclose(cpuinfo);
fprintf(stderr, "CPU: %d * %s\n", num_cpus, cpu_type.c_str());
fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
}
#endif
}
void Start() {
start_ = Env::Default()->NowMicros() * 1e-6;
bytes_ = 0;
message_.clear();
last_op_finish_ = start_;
hist_.Clear();
done_ = 0;
next_report_ = 100;
}
void FinishedSingleOp() {
if (FLAGS_histogram) {
double now = Env::Default()->NowMicros() * 1e-6;
double micros = (now - last_op_finish_) * 1e6;
hist_.Add(micros);
if (micros > 20000) {
fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
fflush(stderr);
}
last_op_finish_ = now;
}
done_++;
if (done_ >= next_report_) {
if (next_report_ < 1000) next_report_ += 100;
else if (next_report_ < 5000) next_report_ += 500;
else if (next_report_ < 10000) next_report_ += 1000;
else if (next_report_ < 50000) next_report_ += 5000;
else if (next_report_ < 100000) next_report_ += 10000;
else if (next_report_ < 500000) next_report_ += 50000;
else next_report_ += 100000;
fprintf(stderr, "... finished %d ops%30s\r", done_, "");
fflush(stderr);
}
}
void Stop(const Slice& name) {
double finish = Env::Default()->NowMicros() * 1e-6;
// Pretend at least one op was done in case we are running a benchmark
// that does not call FinishedSingleOp().
if (done_ < 1) done_ = 1;
if (bytes_ > 0) {
char rate[100];
snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / (finish - start_));
if (!message_.empty()) {
message_ = std::string(rate) + " " + message_;
} else {
message_ = rate;
}
}
fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(),
(finish - start_) * 1e6 / done_,
(message_.empty() ? "" : " "),
message_.c_str());
if (FLAGS_histogram) {
fprintf(stdout, "Microseconds per op:\n%s\n", hist_.ToString().c_str());
}
fflush(stdout);
}
public:
enum Order {
SEQUENTIAL,
RANDOM
};
enum DBState {
FRESH,
EXISTING
};
Benchmark()
: db_(NULL),
num_(FLAGS_num),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
bytes_(0),
rand_(301) {
std::vector<std::string> files;
std::string test_dir;
Env::Default()->GetTestDirectory(&test_dir);
Env::Default()->GetChildren(test_dir.c_str(), &files);
if (!FLAGS_use_existing_db) {
for (int i = 0; i < files.size(); i++) {
if (Slice(files[i]).starts_with("dbbench_polyDB")) {
std::string file_name(test_dir);
file_name += "/";
file_name += files[i];
Env::Default()->DeleteFile(file_name.c_str());
}
}
}
}
~Benchmark() {
if (!db_->close()) {
fprintf(stderr, "close error: %s\n", db_->error().name());
}
}
void Run() {
PrintHeader();
Open(false);
const char* benchmarks = FLAGS_benchmarks;
while (benchmarks != NULL) {
const char* sep = strchr(benchmarks, ',');
Slice name;
if (sep == NULL) {
name = benchmarks;
benchmarks = NULL;
} else {
name = Slice(benchmarks, sep - benchmarks);
benchmarks = sep + 1;
}
Start();
bool known = true;
bool write_sync = false;
if (name == Slice("fillseq")) {
Write(write_sync, SEQUENTIAL, FRESH, num_, FLAGS_value_size, 1);
} else if (name == Slice("fillrandom")) {
Write(write_sync, RANDOM, FRESH, num_, FLAGS_value_size, 1);
DBSynchronize(db_);
} else if (name == Slice("overwrite")) {
Write(write_sync, RANDOM, EXISTING, num_, FLAGS_value_size, 1);
DBSynchronize(db_);
} else if (name == Slice("fillrandsync")) {
write_sync = true;
Write(write_sync, RANDOM, FRESH, num_ / 100, FLAGS_value_size, 1);
DBSynchronize(db_);
} else if (name == Slice("fillseqsync")) {
write_sync = true;
Write(write_sync, SEQUENTIAL, FRESH, num_ / 100, FLAGS_value_size, 1);
DBSynchronize(db_);
} else if (name == Slice("fillrand100K")) {
Write(write_sync, RANDOM, FRESH, num_ / 1000, 100 * 1000, 1);
DBSynchronize(db_);
} else if (name == Slice("fillseq100K")) {
Write(write_sync, SEQUENTIAL, FRESH, num_ / 1000, 100 * 1000, 1);
DBSynchronize(db_);
} else if (name == Slice("readseq")) {
ReadSequential();
} else if (name == Slice("readrandom")) {
ReadRandom();
} else if (name == Slice("readrand100K")) {
int n = reads_;
reads_ /= 1000;
ReadRandom();
reads_ = n;
} else if (name == Slice("readseq100K")) {
int n = reads_;
reads_ /= 1000;
ReadSequential();
reads_ = n;
} else {
known = false;
if (name != Slice()) { // No error message for empty name
fprintf(stderr, "unknown benchmark '%s'\n", name.ToString().c_str());
}
}
if (known) {
Stop(name);
}
}
}
private:
void Open(bool sync) {
assert(db_ == NULL);
// Initialize db_
db_ = new kyotocabinet::TreeDB();
char file_name[100];
db_num_++;
std::string test_dir;
Env::Default()->GetTestDirectory(&test_dir);
snprintf(file_name, sizeof(file_name),
"%s/dbbench_polyDB-%d.kct",
test_dir.c_str(),
db_num_);
// Create tuning options and open the database
int open_options = kyotocabinet::PolyDB::OWRITER |
kyotocabinet::PolyDB::OCREATE;
int tune_options = kyotocabinet::TreeDB::TSMALL |
kyotocabinet::TreeDB::TLINEAR;
if (FLAGS_compression) {
tune_options |= kyotocabinet::TreeDB::TCOMPRESS;
db_->tune_compressor(&comp_);
}
db_->tune_options(tune_options);
db_->tune_page_cache(FLAGS_cache_size);
db_->tune_page(FLAGS_page_size);
db_->tune_map(256LL<<20);
if (sync) {
open_options |= kyotocabinet::PolyDB::OAUTOSYNC;
}
if (!db_->open(file_name, open_options)) {
fprintf(stderr, "open error: %s\n", db_->error().name());
}
}
void Write(bool sync, Order order, DBState state,
int num_entries, int value_size, int entries_per_batch) {
// Create new database if state == FRESH
if (state == FRESH) {
if (FLAGS_use_existing_db) {
message_ = "skipping (--use_existing_db is true)";
return;
}
delete db_;
db_ = NULL;
Open(sync);
Start(); // Do not count time taken to destroy/open
}
if (num_entries != num_) {
char msg[100];
snprintf(msg, sizeof(msg), "(%d ops)", num_entries);
message_ = msg;
}
// Write to database
for (int i = 0; i < num_entries; i++)
{
const int k = (order == SEQUENTIAL) ? i : (rand_.Next() % num_entries);
char key[100];
snprintf(key, sizeof(key), "%016d", k);
bytes_ += value_size + strlen(key);
std::string cpp_key = key;
if (!db_->set(cpp_key, gen_.Generate(value_size).ToString())) {
fprintf(stderr, "set error: %s\n", db_->error().name());
}
FinishedSingleOp();
}
}
void ReadSequential() {
kyotocabinet::DB::Cursor* cur = db_->cursor();
cur->jump();
std::string ckey, cvalue;
while (cur->get(&ckey, &cvalue, true)) {
bytes_ += ckey.size() + cvalue.size();
FinishedSingleOp();
}
delete cur;
}
void ReadRandom() {
std::string value;
for (int i = 0; i < reads_; i++) {
char key[100];
const int k = rand_.Next() % reads_;
snprintf(key, sizeof(key), "%016d", k);
db_->get(key, &value);
FinishedSingleOp();
}
}
};
} // namespace rocksdb
int main(int argc, char** argv) {
std::string default_db_path;
for (int i = 1; i < argc; i++) {
double d;
int n;
char junk;
if (rocksdb::Slice(argv[i]).starts_with("--benchmarks=")) {
FLAGS_benchmarks = argv[i] + strlen("--benchmarks=");
} else if (sscanf(argv[i], "--compression_ratio=%lf%c", &d, &junk) == 1) {
FLAGS_compression_ratio = d;
} else if (sscanf(argv[i], "--histogram=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_histogram = n;
} else if (sscanf(argv[i], "--num=%d%c", &n, &junk) == 1) {
FLAGS_num = n;
} else if (sscanf(argv[i], "--reads=%d%c", &n, &junk) == 1) {
FLAGS_reads = n;
} else if (sscanf(argv[i], "--value_size=%d%c", &n, &junk) == 1) {
FLAGS_value_size = n;
} else if (sscanf(argv[i], "--cache_size=%d%c", &n, &junk) == 1) {
FLAGS_cache_size = n;
} else if (sscanf(argv[i], "--page_size=%d%c", &n, &junk) == 1) {
FLAGS_page_size = n;
} else if (sscanf(argv[i], "--compression=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_compression = (n == 1) ? true : false;
} else if (strncmp(argv[i], "--db=", 5) == 0) {
FLAGS_db = argv[i] + 5;
} else {
fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
exit(1);
}
}
// Choose a location for the test database if none given with --db=<path>
if (FLAGS_db == NULL) {
rocksdb::Env::Default()->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
}
rocksdb::Benchmark benchmark;
benchmark.Run();
return 0;
}

459
doc/benchmark.html
View file

@ -1,459 +0,0 @@
<!DOCTYPE html>
<html>
<head>
<title>LevelDB Benchmarks</title>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<style>
body {
font-family:Helvetica,sans-serif;
padding:20px;
}
h2 {
padding-top:30px;
}
table.bn {
width:800px;
border-collapse:collapse;
border:0;
padding:0;
}
table.bnbase {
width:650px;
}
table.bn td {
padding:2px 0;
}
table.bn td.c1 {
font-weight:bold;
width:150px;
}
table.bn td.c1 div.e {
float:right;
font-weight:normal;
}
table.bn td.c2 {
width:150px;
text-align:right;
padding:2px;
}
table.bn td.c3 {
width:350px;
}
table.bn td.c4 {
width:150px;
font-size:small;
padding-left:4px;
}
/* chart bars */
div.bldb {
background-color:#0255df;
}
div.bkct {
background-color:#df5555;
}
div.bsql {
background-color:#aadf55;
}
.code {
font-family:monospace;
font-size:large;
}
.todo {
color: red;
}
</style>
</head>
<body>
<h1>LevelDB Benchmarks</h1>
<p>Google, July 2011</p>
<hr>
<p>In order to test LevelDB's performance, we benchmark it against other well-established database implementations. We compare LevelDB (revision 39) against <a href="http://www.sqlite.org/">SQLite3</a> (version 3.7.6.3) and <a href="http://fallabs.com/kyotocabinet/spex.html">Kyoto Cabinet's</a> (version 1.2.67) TreeDB (a B+Tree based key-value store). We would like to acknowledge Scott Hess and Mikio Hirabayashi for their suggestions and contributions to the SQLite3 and Kyoto Cabinet benchmarks, respectively.</p>
<p>Benchmarks were all performed on a six-core Intel(R) Xeon(R) CPU X5650 @ 2.67GHz, with 12288 KB of total L3 cache and 12 GB of DDR3 RAM at 1333 MHz. (Note that LevelDB uses at most two CPUs since the benchmarks are single threaded: one to run the benchmark, and one for background compactions.) We ran the benchmarks on two machines (with identical processors), one with an Ext3 file system and one with an Ext4 file system. The machine with the Ext3 file system has a SATA Hitachi HDS721050CLA362 hard drive. The machine with the Ext4 file system has a SATA Samsung HD502HJ hard drive. Both hard drives spin at 7200 RPM and have hard drive write-caching enabled (using `hdparm -W 1 [device]`). The numbers reported below are the median of three measurements.</p>
<h4>Benchmark Source Code</h4>
<p>We wrote benchmark tools for SQLite and Kyoto TreeDB based on LevelDB's <span class="code">db_bench</span>. The code for each of the benchmarks resides here:</p>
<ul>
<li> <b>LevelDB:</b> <a href="http://code.google.com/p/leveldb/source/browse/trunk/db/db_bench.cc">db/db_bench.cc</a>.</li>
<li> <b>SQLite:</b> <a href="http://code.google.com/p/leveldb/source/browse/#svn%2Ftrunk%2Fdoc%2Fbench%2Fdb_bench_sqlite3.cc">doc/bench/db_bench_sqlite3.cc</a>.</li>
<li> <b>Kyoto TreeDB:</b> <a href="http://code.google.com/p/leveldb/source/browse/#svn%2Ftrunk%2Fdoc%2Fbench%2Fdb_bench_tree_db.cc">doc/bench/db_bench_tree_db.cc</a>.</li>
</ul>
<h4>Custom Build Specifications</h4>
<ul>
<li>LevelDB: LevelDB was compiled with the <a href="http://code.google.com/p/google-perftools">tcmalloc</a> library and the <a href="http://code.google.com/p/snappy/">Snappy</a> compression library (revision 33). Assertions were disabled.</li>
<li>TreeDB: TreeDB was compiled using the <a href="http://www.oberhumer.com/opensource/lzo/">LZO</a> compression library (version 2.03). Furthermore, we enabled the TSMALL and TLINEAR options when opening the database in order to reduce the footprint of each record.</li>
<li>SQLite: We tuned SQLite's performance, by setting its locking mode to exclusive. We also enabled SQLite's <a href="http://www.sqlite.org/draft/wal.html">write-ahead logging</a>.</li>
</ul>
<h2>1. Baseline Performance</h2>
<p>This section gives the baseline performance of all the
databases. Following sections show how performance changes as various
parameters are varied. For the baseline:</p>
<ul>
<li> Each database is allowed 4 MB of cache memory.</li>
<li> Databases are opened in <em>asynchronous</em> write mode.
(LevelDB's sync option, TreeDB's OAUTOSYNC option, and
SQLite3's synchronous options are all turned off). I.e.,
every write is pushed to the operating system, but the
benchmark does not wait for the write to reach the disk.</li>
<li> Keys are 16 bytes each.</li>
<li> Value are 100 bytes each (with enough redundancy so that
a simple compressor shrinks them to 50% of their original
size).</li>
<li> Sequential reads/writes traverse the key space in increasing order.</li>
<li> Random reads/writes traverse the key space in random order.</li>
</ul>
<h3>A. Sequential Reads</h3>
<table class="bn bnbase">
<tr><td class="c1">LevelDB</td>
<td class="c2">4,030,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">1,010,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:95px">&nbsp;</div></td>
<tr><td class="c1">SQLite3</td>
<td class="c2">383,000 ops/sec</td>
<td class="c3"><div class="bsql" style="width:33px">&nbsp;</div></td>
</table>
<h3>B. Random Reads</h3>
<table class="bn bnbase">
<tr><td class="c1">LevelDB</td>
<td class="c2">129,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:298px">&nbsp;</div></td>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">151,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:350px">&nbsp;</div></td>
<tr><td class="c1">SQLite3</td>
<td class="c2">134,000 ops/sec</td>
<td class="c3"><div class="bsql" style="width:310px">&nbsp;</div></td>
</table>
<h3>C. Sequential Writes</h3>
<table class="bn bnbase">
<tr><td class="c1">LevelDB</td>
<td class="c2">779,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">342,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:154px">&nbsp;</div></td>
<tr><td class="c1">SQLite3</td>
<td class="c2">48,600 ops/sec</td>
<td class="c3"><div class="bsql" style="width:22px">&nbsp;</div></td>
</table>
<h3>D. Random Writes</h3>
<table class="bn bnbase">
<tr><td class="c1">LevelDB</td>
<td class="c2">164,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">88,500 ops/sec</td>
<td class="c3"><div class="bkct" style="width:188px">&nbsp;</div></td>
<tr><td class="c1">SQLite3</td>
<td class="c2">9,860 ops/sec</td>
<td class="c3"><div class="bsql" style="width:21px">&nbsp;</div></td>
</table>
<p>LevelDB outperforms both SQLite3 and TreeDB in sequential and random write operations and sequential read operations. Kyoto Cabinet has the fastest random read operations.</p>
<h2>2. Write Performance under Different Configurations</h2>
<h3>A. Large Values </h3>
<p>For this benchmark, we start with an empty database, and write 100,000 byte values (~50% compressible). To keep the benchmark running time reasonable, we stop after writing 1000 values.</p>
<h4>Sequential Writes</h4>
<table class="bn bnbase">
<tr><td class="c1">LevelDB</td>
<td class="c2">1,100 ops/sec</td>
<td class="c3"><div class="bldb" style="width:234px">&nbsp;</div></td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">1,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:224px">&nbsp;</div></td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">1,600 ops/sec</td>
<td class="c3"><div class="bsql" style="width:350px">&nbsp;</div></td></tr>
</table>
<h4>Random Writes</h4>
<table class="bn bnbase">
<tr><td class="c1">LevelDB</td>
<td class="c2">480 ops/sec</td>
<td class="c3"><div class="bldb" style="width:105px">&nbsp;</div></td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">1,100 ops/sec</td>
<td class="c3"><div class="bkct" style="width:240px">&nbsp;</div></td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">1,600 ops/sec</td>
<td class="c3"><div class="bsql" style="width:350px">&nbsp;</div></td></tr>
</table>
<p>LevelDB doesn't perform as well with large values of 100,000 bytes each. This is because LevelDB writes keys and values at least twice: first time to the transaction log, and second time (during a compaction) to a sorted file.
With larger values, LevelDB's per-operation efficiency is swamped by the
cost of extra copies of large values.</p>
<h3>B. Batch Writes</h3>
<p>A batch write is a set of writes that are applied atomically to the underlying database. A single batch of N writes may be significantly faster than N individual writes. The following benchmark writes one thousand batches where each batch contains one thousand 100-byte values. TreeDB does not support batch writes and is omitted from this benchmark.</p>
<h4>Sequential Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">840,000 entries/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(1.08x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">124,000 entries/sec</td>
<td class="c3"><div class="bsql" style="width:52px">&nbsp;</div></td>
<td class="c4">(2.55x baseline)</td></tr>
</table>
<h4>Random Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">221,000 entries/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(1.35x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">22,000 entries/sec</td>
<td class="c3"><div class="bsql" style="width:34px">&nbsp;</div></td>
<td class="c4">(2.23x baseline)</td></tr>
</table>
<p>Because of the way LevelDB persistent storage is organized, batches of
random writes are not much slower (only a factor of 4x) than batches
of sequential writes.</p>
<h3>C. Synchronous Writes</h3>
<p>In the following benchmark, we enable the synchronous writing modes
of all of the databases. Since this change significantly slows down the
benchmark, we stop after 10,000 writes. For synchronous write tests, we've
disabled hard drive write-caching (using `hdparm -W 0 [device]`).</p>
<ul>
<li>For LevelDB, we set WriteOptions.sync = true.</li>
<li>In TreeDB, we enabled TreeDB's OAUTOSYNC option.</li>
<li>For SQLite3, we set "PRAGMA synchronous = FULL".</li>
</ul>
<h4>Sequential Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">100 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(0.003x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">7 ops/sec</td>
<td class="c3"><div class="bkct" style="width:27px">&nbsp;</div></td>
<td class="c4">(0.0004x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">88 ops/sec</td>
<td class="c3"><div class="bsql" style="width:315px">&nbsp;</div></td>
<td class="c4">(0.002x baseline)</td></tr>
</table>
<h4>Random Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">100 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(0.015x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">8 ops/sec</td>
<td class="c3"><div class="bkct" style="width:29px">&nbsp;</div></td>
<td class="c4">(0.001x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">88 ops/sec</td>
<td class="c3"><div class="bsql" style="width:314px">&nbsp;</div></td>
<td class="c4">(0.009x baseline)</td></tr>
</table>
<p>Also see the <code>ext4</code> performance numbers below
since synchronous writes behave significantly differently
on <code>ext3</code> and <code>ext4</code>.</p>
<h3>D. Turning Compression Off</h3>
<p>In the baseline measurements, LevelDB and TreeDB were using
light-weight compression
(<a href="http://code.google.com/p/snappy/">Snappy</a> for LevelDB,
and <a href="http://www.oberhumer.com/opensource/lzo/">LZO</a> for
TreeDB). SQLite3, by default does not use compression. The
experiments below show what happens when compression is disabled in
all of the databases (the SQLite3 numbers are just a copy of
its baseline measurements):</p>
<h4>Sequential Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">594,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(0.76x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">485,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:239px">&nbsp;</div></td>
<td class="c4">(1.42x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">48,600 ops/sec</td>
<td class="c3"><div class="bsql" style="width:29px">&nbsp;</div></td>
<td class="c4">(1.00x baseline)</td></tr>
</table>
<h4>Random Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">135,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:296px">&nbsp;</div></td>
<td class="c4">(0.82x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">159,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:350px">&nbsp;</div></td>
<td class="c4">(1.80x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">9,860 ops/sec</td>
<td class="c3"><div class="bsql" style="width:22px">&nbsp;</div></td>
<td class="c4">(1.00x baseline)</td></tr>
</table>
<p>LevelDB's write performance is better with compression than without
since compression decreases the amount of data that has to be written
to disk. Therefore LevelDB users can leave compression enabled in
most scenarios without having worry about a tradeoff between space
usage and performance. TreeDB's performance on the other hand is
better without compression than with compression. Presumably this is
because TreeDB's compression library (LZO) is more expensive than
LevelDB's compression library (Snappy).<p>
<h3>E. Using More Memory</h3>
<p>We increased the overall cache size for each database to 128 MB. For LevelDB, we partitioned 128 MB into a 120 MB write buffer and 8 MB of cache (up from 2 MB of write buffer and 2 MB of cache). For SQLite3, we kept the page size at 1024 bytes, but increased the number of pages to 131,072 (up from 4096). For TreeDB, we also kept the page size at 1024 bytes, but increased the cache size to 128 MB (up from 4 MB).</p>
<h4>Sequential Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">812,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(1.04x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">321,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:138px">&nbsp;</div></td>
<td class="c4">(0.94x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">48,500 ops/sec</td>
<td class="c3"><div class="bsql" style="width:21px">&nbsp;</div></td>
<td class="c4">(1.00x baseline)</td></tr>
</table>
<h4>Random Writes</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">355,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(2.16x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">284,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:280px">&nbsp;</div></td>
<td class="c4">(3.21x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">9,670 ops/sec</td>
<td class="c3"><div class="bsql" style="width:10px">&nbsp;</div></td>
<td class="c4">(0.98x baseline)</td></tr>
</table>
<p>SQLite's performance does not change substantially when compared to
the baseline, but the random write performance for both LevelDB and
TreeDB increases significantly. LevelDB's performance improves
because a larger write buffer reduces the need to merge sorted files
(since it creates a smaller number of larger sorted files). TreeDB's
performance goes up because the entire database is available in memory
for fast in-place updates.</p>
<h2>3. Read Performance under Different Configurations</h2>
<h3>A. Larger Caches</h3>
<p>We increased the overall memory usage to 128 MB for each database.
For LevelDB, we allocated 8 MB to LevelDB's write buffer and 120 MB
to LevelDB's cache. The other databases don't differentiate between a
write buffer and a cache, so we simply set their cache size to 128
MB.</p>
<h4>Sequential Reads</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">5,210,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(1.29x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">1,070,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:72px">&nbsp;</div></td>
<td class="c4">(1.06x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">609,000 ops/sec</td>
<td class="c3"><div class="bsql" style="width:41px">&nbsp;</div></td>
<td class="c4">(1.59x baseline)</td></tr>
</table>
<h4>Random Reads</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">190,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:144px">&nbsp;</div></td>
<td class="c4">(1.47x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">463,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:350px">&nbsp;</div></td>
<td class="c4">(3.07x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">186,000 ops/sec</td>
<td class="c3"><div class="bsql" style="width:141px">&nbsp;</div></td>
<td class="c4">(1.39x baseline)</td></tr>
</table>
<p>As expected, the read performance of all of the databases increases
when the caches are enlarged. In particular, TreeDB seems to make
very effective use of a cache that is large enough to hold the entire
database.</p>
<h3>B. No Compression Reads </h3>
<p>For this benchmark, we populated a database with 1 million entries consisting of 16 byte keys and 100 byte values. We compiled LevelDB and Kyoto Cabinet without compression support, so results that are read out from the database are already uncompressed. We've listed the SQLite3 baseline read performance as a point of comparison.</p>
<h4>Sequential Reads</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">4,880,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:350px">&nbsp;</div></td>
<td class="c4">(1.21x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">1,230,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:88px">&nbsp;</div></td>
<td class="c4">(3.60x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">383,000 ops/sec</td>
<td class="c3"><div class="bsql" style="width:27px">&nbsp;</div></td>
<td class="c4">(1.00x baseline)</td></tr>
</table>
<h4>Random Reads</h4>
<table class="bn">
<tr><td class="c1">LevelDB</td>
<td class="c2">149,000 ops/sec</td>
<td class="c3"><div class="bldb" style="width:300px">&nbsp;</div></td>
<td class="c4">(1.16x baseline)</td></tr>
<tr><td class="c1">Kyoto TreeDB</td>
<td class="c2">175,000 ops/sec</td>
<td class="c3"><div class="bkct" style="width:350px">&nbsp;</div></td>
<td class="c4">(1.16x baseline)</td></tr>
<tr><td class="c1">SQLite3</td>
<td class="c2">134,000 ops/sec</td>
<td class="c3"><div class="bsql" style="width:268px">&nbsp;</div></td>
<td class="c4">(1.00x baseline)</td></tr>
</table>
<p>Performance of both LevelDB and TreeDB improves a small amount when
compression is disabled. Note however that under different workloads,
performance may very well be better with compression if it allows more
of the working set to fit in memory.</p>
<h2>Note about Ext4 Filesystems</h2>
<p>The preceding numbers are for an ext3 file system. Synchronous writes are much slower under <a href="http://en.wikipedia.org/wiki/Ext4">ext4</a> (LevelDB drops to ~31 writes / second and TreeDB drops to ~5 writes / second; SQLite3's synchronous writes do not noticeably drop) due to ext4's different handling of <span class="code">fsync</span> / <span class="code">msync</span> calls. Even LevelDB's asynchronous write performance drops somewhat since it spreads its storage across multiple files and issues <span class="code">fsync</span> calls when switching to a new file.</p>
<h2>Acknowledgements</h2>
<p>Jeff Dean and Sanjay Ghemawat wrote LevelDB. Kevin Tseng wrote and compiled these benchmarks. Mikio Hirabayashi, Scott Hess, and Gabor Cselle provided help and advice.</p>
</body>
</html>

213
doc/impl.html
View file

@ -1,213 +0,0 @@
<!DOCTYPE html>
<html>
<head>
<link rel="stylesheet" type="text/css" href="doc.css" />
<title>Leveldb file layout and compactions</title>
</head>
<body>
<h1>Files</h1>
The implementation of leveldb is similar in spirit to the
representation of a single
<a href="http://labs.google.com/papers/bigtable.html">
Bigtable tablet (section 5.3)</a>.
However the organization of the files that make up the representation
is somewhat different and is explained below.
<p>
Each database is represented by a set of files stored in a directory.
There are several different types of files as documented below:
<p>
<h2>Log files</h2>
<p>
A log file (*.log) stores a sequence of recent updates. Each update
is appended to the current log file. When the log file reaches a
pre-determined size (approximately 4MB by default), it is converted
to a sorted table (see below) and a new log file is created for future
updates.
<p>
A copy of the current log file is kept in an in-memory structure (the
<code>memtable</code>). This copy is consulted on every read so that read
operations reflect all logged updates.
<p>
<h2>Sorted tables</h2>
<p>
A sorted table (*.sst) stores a sequence of entries sorted by key.
Each entry is either a value for the key, or a deletion marker for the
key. (Deletion markers are kept around to hide obsolete values
present in older sorted tables).
<p>
The set of sorted tables are organized into a sequence of levels. The
sorted table generated from a log file is placed in a special <code>young</code>
level (also called level-0). When the number of young files exceeds a
certain threshold (currently four), all of the young files are merged
together with all of the overlapping level-1 files to produce a
sequence of new level-1 files (we create a new level-1 file for every
2MB of data.)
<p>
Files in the young level may contain overlapping keys. However files
in other levels have distinct non-overlapping key ranges. Consider
level number L where L >= 1. When the combined size of files in
level-L exceeds (10^L) MB (i.e., 10MB for level-1, 100MB for level-2,
...), one file in level-L, and all of the overlapping files in
level-(L+1) are merged to form a set of new files for level-(L+1).
These merges have the effect of gradually migrating new updates from
the young level to the largest level using only bulk reads and writes
(i.e., minimizing expensive seeks).
<h2>Manifest</h2>
<p>
A MANIFEST file lists the set of sorted tables that make up each
level, the corresponding key ranges, and other important metadata.
A new MANIFEST file (with a new number embedded in the file name)
is created whenever the database is reopened. The MANIFEST file is
formatted as a log, and changes made to the serving state (as files
are added or removed) are appended to this log.
<p>
<h2>Current</h2>
<p>
CURRENT is a simple text file that contains the name of the latest
MANIFEST file.
<p>
<h2>Info logs</h2>
<p>
Informational messages are printed to files named LOG and LOG.old.
<p>
<h2>Others</h2>
<p>
Other files used for miscellaneous purposes may also be present
(LOCK, *.dbtmp).
<h1>Level 0</h1>
When the log file grows above a certain size (1MB by default):
<ul>
<li>Create a brand new memtable and log file and direct future updates here
<li>In the background:
<ul>
<li>Write the contents of the previous memtable to an sstable
<li>Discard the memtable
<li>Delete the old log file and the old memtable
<li>Add the new sstable to the young (level-0) level.
</ul>
</ul>
<h1>Compactions</h1>
<p>
When the size of level L exceeds its limit, we compact it in a
background thread. The compaction picks a file from level L and all
overlapping files from the next level L+1. Note that if a level-L
file overlaps only part of a level-(L+1) file, the entire file at
level-(L+1) is used as an input to the compaction and will be
discarded after the compaction. Aside: because level-0 is special
(files in it may overlap each other), we treat compactions from
level-0 to level-1 specially: a level-0 compaction may pick more than
one level-0 file in case some of these files overlap each other.
<p>
A compaction merges the contents of the picked files to produce a
sequence of level-(L+1) files. We switch to producing a new
level-(L+1) file after the current output file has reached the target
file size (2MB). We also switch to a new output file when the key
range of the current output file has grown enough to overlap more then
ten level-(L+2) files. This last rule ensures that a later compaction
of a level-(L+1) file will not pick up too much data from level-(L+2).
<p>
The old files are discarded and the new files are added to the serving
state.
<p>
Compactions for a particular level rotate through the key space. In
more detail, for each level L, we remember the ending key of the last
compaction at level L. The next compaction for level L will pick the
first file that starts after this key (wrapping around to the
beginning of the key space if there is no such file).
<p>
Compactions drop overwritten values. They also drop deletion markers
if there are no higher numbered levels that contain a file whose range
overlaps the current key.
<h2>Timing</h2>
Level-0 compactions will read up to four 1MB files from level-0, and
at worst all the level-1 files (10MB). I.e., we will read 14MB and
write 14MB.
<p>
Other than the special level-0 compactions, we will pick one 2MB file
from level L. In the worst case, this will overlap ~ 12 files from
level L+1 (10 because level-(L+1) is ten times the size of level-L,
and another two at the boundaries since the file ranges at level-L
will usually not be aligned with the file ranges at level-L+1). The
compaction will therefore read 26MB and write 26MB. Assuming a disk
IO rate of 100MB/s (ballpark range for modern drives), the worst
compaction cost will be approximately 0.5 second.
<p>
If we throttle the background writing to something small, say 10% of
the full 100MB/s speed, a compaction may take up to 5 seconds. If the
user is writing at 10MB/s, we might build up lots of level-0 files
(~50 to hold the 5*10MB). This may signficantly increase the cost of
reads due to the overhead of merging more files together on every
read.
<p>
Solution 1: To reduce this problem, we might want to increase the log
switching threshold when the number of level-0 files is large. Though
the downside is that the larger this threshold, the more memory we will
need to hold the corresponding memtable.
<p>
Solution 2: We might want to decrease write rate artificially when the
number of level-0 files goes up.
<p>
Solution 3: We work on reducing the cost of very wide merges.
Perhaps most of the level-0 files will have their blocks sitting
uncompressed in the cache and we will only need to worry about the
O(N) complexity in the merging iterator.
<h2>Number of files</h2>
Instead of always making 2MB files, we could make larger files for
larger levels to reduce the total file count, though at the expense of
more bursty compactions. Alternatively, we could shard the set of
files into multiple directories.
<p>
An experiment on an <code>ext3</code> filesystem on Feb 04, 2011 shows
the following timings to do 100K file opens in directories with
varying number of files:
<table class="datatable">
<tr><th>Files in directory</th><th>Microseconds to open a file</th></tr>
<tr><td>1000</td><td>9</td>
<tr><td>10000</td><td>10</td>
<tr><td>100000</td><td>16</td>
</table>
So maybe even the sharding is not necessary on modern filesystems?
<h1>Recovery</h1>
<ul>
<li> Read CURRENT to find name of the latest committed MANIFEST
<li> Read the named MANIFEST file
<li> Clean up stale files
<li> We could open all sstables here, but it is probably better to be lazy...
<li> Convert log chunk to a new level-0 sstable
<li> Start directing new writes to a new log file with recovered sequence#
</ul>
<h1>Garbage collection of files</h1>
<code>DeleteObsoleteFiles()</code> is called at the end of every
compaction and at the end of recovery. It finds the names of all
files in the database. It deletes all log files that are not the
current log file. It deletes all table files that are not referenced
from some level and are not the output of an active compaction.
</body>
</html>

246
doc/index.html
View file

@ -2,28 +2,30 @@
<html>
<head>
<link rel="stylesheet" type="text/css" href="doc.css" />
<title>Leveldb</title>
<title>RocksDB</title>
</head>
<body>
<h1>Leveldb</h1>
<address>Jeff Dean, Sanjay Ghemawat</address>
<h1>RocksDB</h1>
<address>The Facebook Database Engineering Team</address>
<address>Build on earlier work on leveldb by Sanjay Ghemawat
(sanjay@google.com) and Jeff Dean (jeff@google.com)</address>
<p>
The <code>leveldb</code> library provides a persistent key value store. Keys and
The <code>rocksdb</code> library provides a persistent key value store. Keys and
values are arbitrary byte arrays. The keys are ordered within the key
value store according to a user-specified comparator function.
<p>
<h1>Opening A Database</h1>
<p>
A <code>leveldb</code> database has a name which corresponds to a file system
A <code>rocksdb</code> database has a name which corresponds to a file system
directory. All of the contents of database are stored in this
directory. The following example shows how to open a database,
creating it if necessary:
<p>
<pre>
#include &lt;assert&gt;
#include "leveldb/db.h"
#include "rocksdb/db.h"
rocksdb::DB* db;
rocksdb::Options options;
@ -40,7 +42,7 @@ the following line before the <code>rocksdb::DB::Open</code> call:
<h1>Status</h1>
<p>
You may have noticed the <code>rocksdb::Status</code> type above. Values of this
type are returned by most functions in <code>leveldb</code> that may encounter an
type are returned by most functions in <code>rocksdb</code> that may encounter an
error. You can check if such a result is ok, and also print an
associated error message:
<p>
@ -138,11 +140,30 @@ applied together using a synchronous write (i.e.,
the synchronous write will be amortized across all of the writes in
the batch.
<p>
We also provide a way to completely disable Write Ahead Log for a
particular write. If you set write_option.disableWAL to true, the
write will not go to the log at all and may be lost in an event of
process crash.
<p>
When opening a DB, you can disable syncing of data files by setting
Options::disableDataSync to true. This can be useful when doing
bulk-loading or big idempotent operations. Once the operation is
finished, you can manually call sync() to flush all dirty buffers
to stable storage.
<p>
RocksDB by default uses faster fdatasync() to sync files. If you want
to use fsync(), you can set Options::use_fsync to true. You should set
this to true on filesystems like ext3 that can lose files after a
reboot.
<p>
<h1>Concurrency</h1>
<p>
A database may only be opened by one process at a time.
The <code>leveldb</code> implementation acquires a lock from the
The <code>rocksdb</code> implementation acquires a lock from the
operating system to prevent misuse. Within a single process, the
same <code>rocksdb::DB</code> object may be safely shared by multiple
concurrent threads. I.e., different threads may write into or fetch
@ -153,6 +174,19 @@ automatically do the required synchronization). However other objects
If two threads share such an object, they must protect access to it
using their own locking protocol. More details are available in
the public header files.
<p>
<h1>Merge operators</h1>
<p>
Merge operators provide efficient support for read-modify-write operation.
More on the interface and implementation can be found on:
<p>
<a href="https://github.com/facebook/rocksdb/wiki/Merge-Operator">
Merge Operator</a>
<p>
<a href="https://github.com/facebook/rocksdb/wiki/Merge-Operator-Implementation">
Merge Operator Implementation</a>
<p>
<h1>Iteration</h1>
<p>
@ -215,8 +249,8 @@ are instances of the <code>rocksdb::Slice</code> type. <code>Slice</code> is a
structure that contains a length and a pointer to an external byte
array. Returning a <code>Slice</code> is a cheaper alternative to returning a
<code>std::string</code> since we do not need to copy potentially large keys and
values. In addition, <code>leveldb</code> methods do not return null-terminated
C-style strings since <code>leveldb</code> keys and values are allowed to
values. In addition, <code>rocksdb</code> methods do not return null-terminated
C-style strings since <code>rocksdb</code> keys and values are allowed to
contain '\0' bytes.
<p>
C++ strings and null-terminated C-style strings can be easily converted
@ -309,16 +343,42 @@ third part to the keys processed by <code>TwoPartComparator</code>),
(a) keep the same comparator name (b) increment the version number
for new keys (c) change the comparator function so it uses the
version numbers found in the keys to decide how to interpret them.
<p>
<h1>MemTable and Table factories</h1>
<p>
By default, we keep the data in memory in skiplist memtable and the data
on disk in a table format described here:
<a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Table-Format">
RocksDB Table Format</a>.
<p>
Since one of the goals of RocksDB is to have
different parts of the system easily pluggable, we support different
implementations of both memtable and table format. You can supply
your own memtable factory by setting <code>Options::memtable_factory</code>
and your own table factory by setting <code>Options::table_factory</code>.
For available memtable factories, please refer to
<code>rocksdb/memtablerep.h</code> and for table factores to
<code>rocksdb/table.h</code>. These features are both in active development
and please be wary of any API changes that might break your application
going forward.
<p>
You can also read more about memtables here:
<a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Architecture-Guide#memtables">
Memtables wiki
</a>
<p>
<h1>Performance</h1>
<p>
Performance can be tuned by changing the default values of the
types defined in <code>include/leveldb/options.h</code>.
types defined in <code>include/rocksdb/options.h</code>.
<p>
<h2>Block size</h2>
<p>
<code>leveldb</code> groups adjacent keys together into the same block and such a
<code>rocksdb</code> groups adjacent keys together into the same block and such a
block is the unit of transfer to and from persistent storage. The
default block size is approximately 4096 uncompressed bytes.
Applications that mostly do bulk scans over the contents of the
@ -329,6 +389,29 @@ much benefit in using blocks smaller than one kilobyte, or larger than
a few megabytes. Also note that compression will be more effective
with larger block sizes.
<p>
<h2>Write buffer</h2>
<p>
<code>Options::write_buffer_size</code> specifies the amount of data
to build up in memory before converting to a sorted on-disk file.
Larger values increase performance, especially during bulk loads.
Up to max_write_buffer_number write buffers may be held in memory
at the same time,
so you may wish to adjust this parameter to control memory usage.
Also, a larger write buffer will result in a longer recovery time
the next time the database is opened.
Related option is
<code>Options::max_write_buffer_number</code>, which is maximum number
of write buffers that are built up in memory. The default is 2, so that
when 1 write buffer is being flushed to storage, new writes can continue
to the other write buffer.
<code>Options::min_write_buffer_number_to_merge</code> is the minimum number
of write buffers that will be merged together before writing to storage.
If set to 1, then all write buffers are fushed to L0 as individual files and
this increases read amplification because a get request has to check in all
of these files. Also, an in-memory merge may result in writing lesser
data to storage if there are duplicate records in each of these
individual write buffers. Default: 1
<p>
<h2>Compression</h2>
<p>
Each block is individually compressed before being written to
@ -347,25 +430,26 @@ performance improvement:
<p>
The contents of the database are stored in a set of files in the
filesystem and each file stores a sequence of compressed blocks. If
<code>options.cache</code> is non-NULL, it is used to cache frequently used
uncompressed block contents.
<code>options.block_cache</code> is non-NULL, it is used to cache frequently
used uncompressed block contents. If <code>options.block_cache_compressed</code>
is non-NULL, it is used to cache frequently used compressed blocks. Compressed
cache is an alternative to OS cache, which also caches compressed blocks. If
compressed cache is used, you should disable OS cache by setting
<code>options.allow_os_buffer</code> to false.
<p>
<pre>
#include "leveldb/cache.h"
#include "rocksdb/cache.h"
rocksdb::Options options;
options.cache = rocksdb::NewLRUCache(100 * 1048576); // 100MB cache
options.block_cache = rocksdb::NewLRUCache(100 * 1048576); // 100MB uncompressed cache
options.block_cache_compressed = rocksdb::NewLRUCache(100 * 1048576); // 100MB compressed cache
rocksdb::DB* db;
rocksdb::DB::Open(options, name, &db);
... use the db ...
delete db
delete options.cache;
delete options.block_cache;
delete options.block_cache_compressed;
</pre>
Note that the cache holds uncompressed data, and therefore it should
be sized according to application level data sizes, without any
reduction from compression. (Caching of compressed blocks is left to
the operating system buffer cache, or any custom <code>Env</code>
implementation provided by the client.)
<p>
When performing a bulk read, the application may wish to disable
caching so that the data processed by the bulk read does not end up
@ -380,6 +464,9 @@ used to achieve this:
...
}
</pre>
<p>
You can also disable block cache by setting <code>options.no_block_cache</code>
to true.
<h2>Key Layout</h2>
<p>
Note that the unit of disk transfer and caching is a block. Adjacent
@ -389,7 +476,7 @@ by placing keys that are accessed together near each other and placing
infrequently used keys in a separate region of the key space.
<p>
For example, suppose we are implementing a simple file system on top
of <code>leveldb</code>. The types of entries we might wish to store are:
of <code>rocksdb</code>. The types of entries we might wish to store are:
<p>
<pre>
filename -&gt; permission-bits, length, list of file_block_ids
@ -402,7 +489,7 @@ contents.
<p>
<h2>Filters</h2>
<p>
Because of the way <code>leveldb</code> data is organized on disk,
Because of the way <code>rocksdb</code> data is organized on disk,
a single <code>Get()</code> call may involve multiple reads from disk.
The optional <code>FilterPolicy</code> mechanism can be used to reduce
the number of disk reads substantially.
@ -461,11 +548,11 @@ also ignores trailing spaces. For example:
<p>
Advanced applications may provide a filter policy that does not use
a bloom filter but uses some other mechanism for summarizing a set
of keys. See <code>leveldb/filter_policy.h</code> for detail.
of keys. See <code>rocksdb/filter_policy.h</code> for detail.
<p>
<h1>Checksums</h1>
<p>
<code>leveldb</code> associates checksums with all data it stores in the file system.
<code>rocksdb</code> associates checksums with all data it stores in the file system.
There are two separate controls provided over how aggressively these
checksums are verified:
<p>
@ -485,9 +572,52 @@ checksums are verified:
<p>
If a database is corrupted (perhaps it cannot be opened when
paranoid checking is turned on), the <code>rocksdb::RepairDB</code> function
may be used to recover as much of the data as possible
may be used to recover as much of the data as possible.
<p>
</ul>
<p>
<h1>Compaction</h1>
<p>
You can read more on Compactions here:
<a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Architecture-Guide#multi-threaded-compactions">
Multi-threaded compactions
</a>
<p>
Here we give overview of the options that impact behavior of Compactions:
<ul>
<p>
<li><code>Options::compaction_style</code> - RocksDB currently supports two
compaction algorithms - Compaction style and Level style. This option switches
between the two. Can be kCompactionStyleUniversal or kCompactionStyleLevel.
If this is kCompactionStyleUniversal, then you can configure universal style
parameters with <code>Options::compaction_options_universal</code>.
<p>
<li><code>Options::disable_auto_compactions</code> - Disable automatic compactions.
Manual compactions can still be issued on this database.
<p>
<li><code>Options::compaction_filter</code> - Allows an application to modify/delete
a key-value during background compaction. The client must provide
compaction_filter_factory if it requires a new compaction filter to be used
for different compaction processes. Client should specify only one of filter
or factory.
<p>
<li><code>Options::compaction_filter_factory</code> - a factory that provides
compaction filter objects which allow an application to modify/delete a
key-value during background compaction.
</ul>
<p>
Other options impacting performance of compactions and when they get triggered
are: <code>access_hint_on_compaction_start</code>,
<code>level0_file_num_compaction_trigger</code>,
<code>max_mem_compaction_level</code>, <code>target_file_size_base</code>,
<code>target_file_size_multiplier</code>,
<code>expanded_compaction_factor</code>, <code>source_compaction_factor</code>,
<code>max_grandparent_overlap_factor</code>,
<code>disable_seek_compaction</code>, <code>max_background_compactions</code>.
<p>
You can learn more about all of those options in <code>rocksdb/options.h</code>
<h1>Approximate Sizes</h1>
<p>
The <code>GetApproximateSizes</code> method can used to get the approximate
@ -508,11 +638,11 @@ bytes of file system space used by the key range <code>[a..c)</code> and
<h1>Environment</h1>
<p>
All file operations (and other operating system calls) issued by the
<code>leveldb</code> implementation are routed through a <code>rocksdb::Env</code> object.
<code>rocksdb</code> implementation are routed through a <code>rocksdb::Env</code> object.
Sophisticated clients may wish to provide their own <code>Env</code>
implementation to get better control. For example, an application may
introduce artificial delays in the file IO paths to limit the impact
of <code>leveldb</code> on other activities in the system.
of <code>rocksdb</code> on other activities in the system.
<p>
<pre>
class SlowEnv : public rocksdb::Env {
@ -526,22 +656,66 @@ of <code>leveldb</code> on other activities in the system.
</pre>
<h1>Porting</h1>
<p>
<code>leveldb</code> may be ported to a new platform by providing platform
<code>rocksdb</code> may be ported to a new platform by providing platform
specific implementations of the types/methods/functions exported by
<code>leveldb/port/port.h</code>. See <code>leveldb/port/port_example.h</code> for more
<code>rocksdb/port/port.h</code>. See <code>rocksdb/port/port_example.h</code> for more
details.
<p>
In addition, the new platform may need a new default <code>rocksdb::Env</code>
implementation. See <code>leveldb/util/env_posix.h</code> for an example.
implementation. See <code>rocksdb/util/env_posix.h</code> for an example.
<h1>Statistics</h1>
<p>
To be able to efficiently tune your application, it is always helpful if you
have access to usage statistics. You can collect those statistics by setting
<code>Options::table_stats_collectors</code> or
<code>Options::statistics</code>. For more information, refer to
<code>rocksdb/table_stats.h</code> and <code>rocksdb/statistics.h</code>.
These should not add significant overhead to your application and we
recommend exporting them to other monitoring tools.
<h1>Purging WAL files</h1>
<p>
By default, old write-ahead logs are deleted automatically when they fall out
of scope and application doesn't need them anymore. There are options that
enable the user to archive the logs and then delete them lazily, either in
TTL fashion or based on size limit.
The options are <code>Options::WAL_ttl_seconds</code> and
<code>Options::WAL_size_limit_MB</code>. Here is how they can be used:
<ul>
<li>
<p>
If both set to 0, logs will be deleted asap and will never get into the archive.
<li>
<p>
If <code>WAL_ttl_seconds</code> is 0 and WAL_size_limit_MB is not 0, WAL
files will be checked every 10 min and if total size is greater then
<code>WAL_size_limit_MB</code>, they will be deleted starting with the
earliest until size_limit is met. All empty files will be deleted.
<li>
<p>
If <code>WAL_ttl_seconds</code> is not 0 and WAL_size_limit_MB is 0, then
WAL files will be checked every <code>WAL_ttl_seconds / 2</code> and those
that are older than WAL_ttl_seconds will be deleted.
<li>
<p>
If both are not 0, WAL files will be checked every 10 min and both
checks will be performed with ttl being first.
</ul>
You can completely disable WAL file purging by setting
<code>Options::purge_log_after_memtable_flush</code> to false.
<h1>Other Information</h1>
<p>
Details about the <code>leveldb</code> implementation may be found in
Details about the <code>rocksdb</code> implementation may be found in
the following documents:
<ul>
<li> <a href="impl.html">Implementation notes</a>
<li> <a href="table_format.txt">Format of an immutable Table file</a>
<li> <a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Architecture-Guide">
RocksDB Architecture Guide</a>
<li> <a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Table-Format">
Format of an immutable Table file</a>
<li> <a href="log_format.txt">Format of a log file</a>
</ul>

102
doc/table_format.txt
View file

@ -1,102 +0,0 @@
File format
===========
<beginning_of_file>
[data block 1]
[data block 2]
...
[data block N]
[meta block 1]
...
[meta block K]
[metaindex block]
[index block]
[Footer] (fixed size; starts at file_size - sizeof(Footer))
<end_of_file>
The file contains internal pointers. Each such pointer is called
a BlockHandle and contains the following information:
offset: varint64
size: varint64
(1) The sequence of key/value pairs in the file are stored in sorted
order and partitioned into a sequence of data blocks. These blocks
come one after another at the beginning of the file. Each data block
is formatted according to the code in block_builder.cc, and then
optionally compressed.
(2) After the data blocks we store a bunch of meta blocks. The
supported meta block types are described below. More meta block types
may be added in the future. Each meta block is again formatted using
block_builder.cc and then optionally compressed.
(3) A "metaindex" block. It contains one entry for every other meta
block where the key is the name of the meta block and the value is a
BlockHandle pointing to that meta block.
(4) An "index" block. This block contains one entry per data block,
where the key is a string >= last key in that data block and before
the first key in the successive data block. The value is the
BlockHandle for the data block.
(6) At the very end of the file is a fixed length footer that contains
the BlockHandle of the metaindex and index blocks as well as a magic number.
metaindex_handle: char[p]; // Block handle for metaindex
index_handle: char[q]; // Block handle for index
padding: char[40-p-q]; // 0 bytes to make fixed length
// (40==2*BlockHandle::kMaxEncodedLength)
magic: fixed64; // == 0xdb4775248b80fb57
"filter" Meta Block
-------------------
If a "FilterPolicy" was specified when the database was opened, a
filter block is stored in each table. The "metaindex" block contains
an entry that maps from "filter.<N>" to the BlockHandle for the filter
block where "<N>" is the string returned by the filter policy's
"Name()" method.
The filter block stores a sequence of filters, where filter i contains
the output of FilterPolicy::CreateFilter() on all keys that are stored
in a block whose file offset falls within the range
[ i*base ... (i+1)*base-1 ]
Currently, "base" is 2KB. So for example, if blocks X and Y start in
the range [ 0KB .. 2KB-1 ], all of the keys in X and Y will be
converted to a filter by calling FilterPolicy::CreateFilter(), and the
resulting filter will be stored as the first filter in the filter
block.
The filter block is formatted as follows:
[filter 0]
[filter 1]
[filter 2]
...
[filter N-1]
[offset of filter 0] : 4 bytes
[offset of filter 1] : 4 bytes
[offset of filter 2] : 4 bytes
...
[offset of filter N-1] : 4 bytes
[offset of beginning of offset array] : 4 bytes
lg(base) : 1 byte
The offset array at the end of the filter block allows efficient
mapping from a data block offset to the corresponding filter.
"stats" Meta Block
------------------
This meta block contains a bunch of stats. The key is the name
of the statistic. The value contains the statistic.
TODO(postrelease): record following stats.
data size
index size
key size (uncompressed)
value size (uncompressed)
number of entries
number of data blocks

6
include/rocksdb/options.h
View file

@ -355,7 +355,7 @@ struct Options {
// If true, then every store to stable storage will issue a fsync.
// If false, then every store to stable storage will issue a fdatasync.
// This parameter should be set to true while storing data to
// filesystem like ext3 which can lose files after a reboot.
// filesystem like ext3 that can lose files after a reboot.
// Default: false
bool use_fsync;
@ -504,7 +504,7 @@ struct Options {
// then WAL_size_limit_MB, they will be deleted starting with the
// earliest until size_limit is met. All empty files will be deleted.
// 3. If WAL_ttl_seconds is not 0 and WAL_size_limit_MB is 0, then
// WAL files will be checked every WAL_ttl_secondsi / 2 and those which
// WAL files will be checked every WAL_ttl_secondsi / 2 and those that
// are older than WAL_ttl_seconds will be deleted.
// 4. If both are not 0, WAL files will be checked every 10 min and both
// checks will be performed with ttl being first.
@ -723,7 +723,7 @@ struct WriteOptions {
// In other words, a DB write with sync==false has similar
// crash semantics as the "write()" system call. A DB write
// with sync==true has similar crash semantics to a "write()"
// system call followed by "fsync()".
// system call followed by "fdatasync()".
//
// Default: false
bool sync;