Merge remote-tracking branch 'origin/master' into task1181723

Conflicts:
	Makefile
main
bol 12 years ago
commit 89f0c25360
  1. 35
      Makefile
  2. 14
      build_detect_platform
  3. 25
      build_detect_version
  4. 2
      db/builder.cc
  5. 61
      db/db_bench.cc
  6. 105
      db/db_impl.cc
  7. 18
      db/db_impl.h
  8. 72
      db/db_stats_logger.cc
  9. 27
      db/db_test.cc
  10. 18
      db/filename.cc
  11. 2
      db/filename.h
  12. 176
      db/version_set.cc
  13. 9
      db/version_set.h
  14. 7
      fbcode.sh
  15. 22
      hdfs/env_hdfs.h
  16. 21
      include/leveldb/env.h
  17. 26
      include/leveldb/options.h
  18. 1012
      scribe/if/gen-cpp/scribe.cpp
  19. 593
      scribe/if/gen-cpp/scribe.h
  20. 17
      scribe/if/gen-cpp/scribe_constants.cpp
  21. 25
      scribe/if/gen-cpp/scribe_constants.h
  22. 513
      scribe/if/gen-cpp/scribe_types.cpp
  23. 247
      scribe/if/gen-cpp/scribe_types.h
  24. 82
      scribe/if/scribe.thrift
  25. 90
      scribe/scribe_logger.cc
  26. 71
      scribe/scribe_logger.h
  27. 2
      thrift/README
  28. BIN
      thrift/bin/thrift
  29. 209
      thrift/gen-cpp/leveldb_types.cpp
  30. 70
      thrift/gen-cpp/leveldb_types.h
  31. 15
      thrift/if/leveldb.thrift
  32. 8
      thrift/server_options.h
  33. 23
      thrift/server_utils.cpp
  34. 71
      tools/manifest_dump.cc
  35. 14
      util/build_version.h
  36. 79
      util/env_posix.cc
  37. 57
      util/filelock_test.cc
  38. 27
      util/options.cc
  39. 23
      util/stats_logger.h

@ -33,8 +33,6 @@ MEMENVOBJECTS = $(MEMENV_SOURCES:.cc=.o)
TESTUTIL = ./util/testutil.o
TESTHARNESS = ./util/testharness.o $(TESTUTIL)
TOOLS = \
leveldb_shell
TESTS = \
arena_test \
@ -55,10 +53,17 @@ TESTS = \
table_test \
version_edit_test \
version_set_test \
write_batch_test
write_batch_test \
filelock_test
TOOLS = \
manifest_dump \
leveldb_shell
PROGRAMS = db_bench $(TESTS)
PROGRAMS = db_bench $(TESTS) $(TOOLS)
BENCHMARKS = db_bench_sqlite3 db_bench_tree_db
VERSIONFILE=util/build_version.cc
LIBRARY = libleveldb.a
MEMENVLIBRARY = libmemenv.a
@ -82,21 +87,20 @@ $(SHARED1): $(SHARED3)
ln -fs $(SHARED3) $(SHARED1)
endif
all: $(SHARED) $(LIBRARY) $(THRIFTSERVER) $(TOOLS)
check: all $(PROGRAMS) $(TESTS)
all: $(VERSIONFILE) $(SHARED) $(LIBRARY) $(THRIFTSERVER) $(TOOLS)
check: all $(PROGRAMS) $(TESTS) $(TOOLS)
for t in $(TESTS); do echo "***** Running $$t"; ./$$t || exit 1; done
clean:
-rm -f $(PROGRAMS) $(BENCHMARKS) $(LIBRARY) $(SHARED) $(MEMENVLIBRARY) $(THRIFTSERVER) */*.o */*/*.o ios-x86/*/*.o ios-arm/*/*.o build_config.mk
-rm -f $(PROGRAMS) $(BENCHMARKS) $(LIBRARY) $(SHARED) $(MEMENVLIBRARY) $(THRIFTSERVER) */*.o */*/*.o ios-x86/*/*.o ios-arm/*/*.o build_config.mk $(VERSIONFILE)
-rm -rf ios-x86/* ios-arm/*
$(LIBRARY): $(LIBOBJECTS)
rm -f $@
$(AR) -rs $@ $(LIBOBJECTS)
leveldb_shell: tools/shell/ShellContext.o tools/shell/ShellState.o tools/shell/LeveldbShell.o tools/shell/DBClientProxy.o tools/shell/ShellContext.h tools/shell/ShellState.h tools/shell/DBClientProxy.h $(LIBOBJECTS)
$(CXX) tools/shell/ShellContext.o tools/shell/ShellState.o tools/shell/LeveldbShell.o tools/shell/DBClientProxy.o $(LIBOBJECTS) -o $@ $(LDFLAGS)
db_bench: db/db_bench.o $(LIBOBJECTS) $(TESTUTIL)
$(CXX) db/db_bench.o $(LIBOBJECTS) $(TESTUTIL) $(EXEC_LDFLAGS) -o $@ $(LDFLAGS)
@ -174,9 +178,22 @@ leveldb_server: thrift/server.o $(LIBRARY)
leveldb_server_test: thrift/test/simpletest.o $(LIBRARY)
$(CXX) thrift/test/simpletest.o $(LIBRARY) $(EXEC_LDFLAGS) -o $@ $(LDFLAGS)
leveldb_shell: tools/shell/ShellContext.o tools/shell/ShellState.o tools/shell/LeveldbShell.o tools/shell/DBClientProxy.o tools/shell/ShellContext.h tools/shell/ShellState.h tools/shell/DBClientProxy.h $(LIBOBJECTS)
$(CXX) tools/shell/ShellContext.o tools/shell/ShellState.o tools/shell/LeveldbShell.o tools/shell/DBClientProxy.o $(LIBOBJECTS) -o $@ $(LDFLAGS)
DBClientProxy_test: tools/shell/test/DBClientProxyTest.o tools/shell/DBClientProxy.o $(LIBRARY)
$(CXX) tools/shell/test/DBClientProxyTest.o tools/shell/DBClientProxy.o $(LIBRARY) $(EXEC_LDFLAGS) -o $@ $(LDFLAGS)
manifest_dump: tools/manifest_dump.o $(LIBOBJECTS)
$(CXX) tools/manifest_dump.o $(LIBOBJECTS) -o $@ $(LDFLAGS)
filelock_test: util/filelock_test.o $(LIBOBJECTS) $(TESTHARNESS)
$(CXX) util/filelock_test.o $(LIBOBJECTS) $(TESTHARNESS) -o $@ $(LDFLAGS)
# recreate the version file with the latest git revision
$(VERSIONFILE): build_detect_version
$(shell ./build_detect_platform build_config.mk)
ifeq ($(PLATFORM), IOS)
# For iOS, create universal object files to be used on both the simulator and
# a device.

@ -96,15 +96,22 @@ case "$TARGET_OS" in
exit 1
esac
./build_detect_version
# We want to make a list of all cc files within util, db, table, and helpers
# except for the test and benchmark files. By default, find will output a list
# of all files matching either rule, so we need to append -print to make the
# prune take effect.
DIRS="util db table"
if test "$USE_THRIFT"; then
DIRS+=" thrift/gen-cpp thrift/server_utils.cpp "
DIRS+=" thrift/server_utils.cpp thrift/gen-cpp "
THRIFTSERVER=leveldb_server
fi
if test "$USE_SCRIBE"; then
DIRS+=" scribe "
fi
set -f # temporarily disable globbing so that our patterns aren't expanded
PRUNE_TEST="-name *test*.cc -prune"
PRUNE_BENCH="-name *_bench.cc -prune"
@ -197,6 +204,11 @@ if test "$USE_THRIFT"; then
PLATFORM_LDFLAGS+=$THRIFT_LDFLAGS
fi
#shall we build with scribe
if test "$USE_SCRIBE"; then
COMMON_FLAGS="$COMMON_FLAGS -I./thrift/lib/ -DUSE_SCRIBE"
fi
PLATFORM_CCFLAGS="$PLATFORM_CCFLAGS $COMMON_FLAGS"
PLATFORM_CXXFLAGS="$PLATFORM_CXXFLAGS $COMMON_FLAGS"

@ -0,0 +1,25 @@
#!/bin/sh
#
# Record the version of the source that we are compiling.
# We keep a record of the git revision in util/version.cc. This source file
# is then built as a regular source file as part of the compilation process.
# One can run "strings executable_filename | grep _build_" to find the version of
# the source that we used to build the executable file.
#
# create git version file
VFILE=util/build_version.cc
# check to see if git is in the path
which git > /dev/null
if [ "$?" = 0 ]; then
git rev-parse HEAD | awk ' BEGIN {print "#include \"build_version.h\""} {print "const char * leveldb_build_git_sha = \"leveldb_build_git_sha:" $0"\";"} END {}' > ${VFILE}
else
echo "git not found"| awk ' BEGIN {print "#include \"build_version.h\""} {print "const char * leveldb_build_git_sha = \"leveldb_build_git_sha:git not found\";"} END {}' > ${VFILE}
fi
date | awk 'BEGIN {} {print "const char * leveldb_build_git_datetime = \"leveldb_build_git_datetime:"$0"\";"} END {} ' >> ${VFILE}
echo "const char * leveldb_build_compile_date = __DATE__;" >> ${VFILE}
echo "const char * leveldb_build_compile_time = __TIME__;" >> ${VFILE}

@ -53,7 +53,7 @@ Status BuildTable(const std::string& dbname,
delete builder;
// Finish and check for file errors
if (s.ok()) {
if (s.ok() && !options.disableDataSync) {
s = file->Sync();
}
if (s.ok()) {

@ -120,9 +120,35 @@ static class leveldb::DBStatistics* dbstats = NULL;
// Number of write operations to do. If negative, do FLAGS_num reads.
static long FLAGS_writes = -1;
// These default values might change if the hardcoded
// Sync all writes to disk
static bool FLAGS_sync = false;
// If true, do not wait until data is synced to disk.
static bool FLAGS_disable_data_sync = false;
// If true, do not write WAL for write.
static bool FLAGS_disable_wal = false;
// Target level-0 file size for compaction
static int FLAGS_target_file_size_base = 2 * 1048576;
// A multiplier to compute targe level-N file size
static int FLAGS_target_file_size_multiplier = 1;
// Max bytes for level-0
static int FLAGS_max_bytes_for_level_base = 10 * 1048576;
// A multiplier to compute max bytes for level-N
static int FLAGS_max_bytes_for_level_multiplier = 10;
// Number of files in level-0 that will trigger put stop.
static int FLAGS_level0_stop_writes_trigger = 12;
// Number of files in level-0 that will slow down writes.
static int FLAGS_level0_slowdown_writes_trigger = 8;
// posix or hdfs environment
static leveldb::Env* FLAGS_env = leveldb::Env::Default();
@ -485,6 +511,8 @@ class Benchmark {
write_options_.sync = true;
}
write_options_.disableWAL = FLAGS_disable_wal;
void (Benchmark::*method)(ThreadState*) = NULL;
bool fresh_db = false;
int num_threads = FLAGS_threads;
@ -745,6 +773,15 @@ class Benchmark {
options.max_open_files = FLAGS_open_files;
options.statistics = dbstats;
options.env = FLAGS_env;
options.disableDataSync = FLAGS_disable_data_sync;
options.target_file_size_base = FLAGS_target_file_size_base;
options.target_file_size_multiplier = FLAGS_target_file_size_multiplier;
options.max_bytes_for_level_base = FLAGS_max_bytes_for_level_base;
options.max_bytes_for_level_multiplier =
FLAGS_max_bytes_for_level_multiplier;
options.level0_stop_writes_trigger = FLAGS_level0_stop_writes_trigger;
options.level0_slowdown_writes_trigger =
FLAGS_level0_slowdown_writes_trigger;
Status s = DB::Open(options, FLAGS_db, &db_);
if (!s.ok()) {
fprintf(stderr, "open error: %s\n", s.ToString().c_str());
@ -1030,8 +1067,32 @@ int main(int argc, char** argv) {
} else if (sscanf(argv[i], "--sync=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_sync = n;
} else if (sscanf(argv[i], "--disable_data_sync=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_disable_data_sync = n;
} else if (sscanf(argv[i], "--disable_wal=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_disable_wal = n;
} else if (sscanf(argv[i], "--hdfs=%s", &hdfsname) == 1) {
FLAGS_env = new leveldb::HdfsEnv(hdfsname);
} else if (sscanf(argv[i], "--target_file_size_base=%d%c",
&n, &junk) == 1) {
FLAGS_target_file_size_base = n;
} else if ( sscanf(argv[i], "--target_file_size_multiplier=%d%c",
&n, &junk) == 1) {
FLAGS_target_file_size_multiplier = n;
} else if (
sscanf(argv[i], "--max_bytes_for_level_base=%d%c", &n, &junk) == 1) {
FLAGS_max_bytes_for_level_base = n;
} else if (sscanf(argv[i], "--max_bytes_for_level_multiplier=%d%c",
&n, &junk) == 1) {
FLAGS_max_bytes_for_level_multiplier = n;
} else if (sscanf(argv[i],"--level0_stop_writes_trigger=%d%c",
&n, &junk) == 1) {
FLAGS_level0_stop_writes_trigger = n;
} else if (sscanf(argv[i],"--level0_slowdown_writes_trigger=%d%c",
&n, &junk) == 1) {
FLAGS_level0_slowdown_writes_trigger = n;
} else {
fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
exit(1);

@ -10,6 +10,7 @@
#include <stdint.h>
#include <stdio.h>
#include <vector>
#include <algorithm>
#include "db/builder.h"
#include "db/db_iter.h"
#include "db/dbformat.h"
@ -32,6 +33,7 @@
#include "util/coding.h"
#include "util/logging.h"
#include "util/mutexlock.h"
#include "util/build_version.h"
namespace leveldb {
@ -99,7 +101,8 @@ Options SanitizeOptions(const std::string& dbname,
if (result.info_log == NULL) {
// Open a log file in the same directory as the db
src.env->CreateDir(dbname); // In case it does not exist
src.env->RenameFile(InfoLogFileName(dbname), OldInfoLogFileName(dbname));
src.env->RenameFile(InfoLogFileName(dbname),
OldInfoLogFileName(dbname, src.env->NowMicros()));
Status s = src.env->NewLogger(InfoLogFileName(dbname), &result.info_log);
if (!s.ok()) {
// No place suitable for logging
@ -131,17 +134,34 @@ DBImpl::DBImpl(const Options& options, const std::string& dbname)
log_(NULL),
tmp_batch_(new WriteBatch),
bg_compaction_scheduled_(false),
manual_compaction_(NULL) {
manual_compaction_(NULL),
logger_(NULL) {
mem_->Ref();
has_imm_.Release_Store(NULL);
stats_ = new CompactionStats[options.num_levels];
// Reserve ten files or so for other uses and give the rest to TableCache.
const int table_cache_size = options.max_open_files - 10;
const int table_cache_size = options_.max_open_files - 10;
table_cache_ = new TableCache(dbname_, &options_, table_cache_size);
versions_ = new VersionSet(dbname_, &options_, table_cache_,
&internal_comparator_);
options_.Dump(options_.info_log);
#ifdef USE_SCRIBE
logger_ = new ScribeLogger("localhost", 1456);
#endif
char name[100];
Status st = env_->GetHostName(name, 100);
if(st.ok()) {
host_name_ = name;
} else {
Log(options_.info_log, "Can't get hostname, use localhost as host name.");
host_name_ = "localhost";
}
last_log_ts = 0;
}
DBImpl::~DBImpl() {
@ -224,6 +244,7 @@ void DBImpl::DeleteObsoleteFiles() {
env_->GetChildren(dbname_, &filenames); // Ignoring errors on purpose
uint64_t number;
FileType type;
std::vector<uint64_t> old_log_files_ts;
for (size_t i = 0; i < filenames.size(); i++) {
if (ParseFileName(filenames[i], &number, &type)) {
bool keep = true;
@ -245,9 +266,14 @@ void DBImpl::DeleteObsoleteFiles() {
// be recorded in pending_outputs_, which is inserted into "live"
keep = (live.find(number) != live.end());
break;
case kInfoLogFile:
keep = true;
if (number != 0) {
old_log_files_ts.push_back(number);
}
break;
case kCurrentFile:
case kDBLockFile:
case kInfoLogFile:
keep = true;
break;
}
@ -263,6 +289,20 @@ void DBImpl::DeleteObsoleteFiles() {
}
}
}
// Delete old log files.
int old_log_file_count = old_log_files_ts.size();
if (old_log_file_count >= KEEP_LOG_FILE_NUM) {
std::sort(old_log_files_ts.begin(), old_log_files_ts.end());
for (int i = 0; i >= (old_log_file_count - KEEP_LOG_FILE_NUM); i++) {
uint64_t ts = old_log_files_ts.at(i);
std::string to_delete = OldInfoLogFileName(dbname_, ts);
Log(options_.info_log, "Delete type=%d #%lld\n",
int(kInfoLogFile),
static_cast<unsigned long long>(ts));
env_->DeleteFile(dbname_ + "/" + to_delete);
}
}
}
Status DBImpl::Recover(VersionEdit* edit) {
@ -514,6 +554,7 @@ Status DBImpl::CompactMemTable() {
imm_ = NULL;
has_imm_.Release_Store(NULL);
DeleteObsoleteFiles();
MaybeScheduleLogDBDeployStats();
}
return s;
@ -537,15 +578,15 @@ void DBImpl::CompactRange(const Slice* begin, const Slice* end) {
}
int DBImpl::NumberLevels() {
return options_.num_levels;
return options_.num_levels;
}
int DBImpl::MaxMemCompactionLevel() {
return options_.max_mem_compaction_level;
return options_.max_mem_compaction_level;
}
int DBImpl::Level0StopWriteTrigger() {
return options_.level0_stop_writes_trigger;
return options_.level0_stop_writes_trigger;
}
Status DBImpl::Flush(const FlushOptions& options) {
@ -598,34 +639,33 @@ Status DBImpl::FlushMemTable(const FlushOptions& options) {
}
Status DBImpl::WaitForCompactMemTable() {
Status s;
// Wait until the compaction completes
MutexLock l(&mutex_);
while (imm_ != NULL && bg_error_.ok()) {
bg_cv_.Wait();
}
if (imm_ != NULL) {
s = bg_error_;
}
return s;
Status s;
// Wait until the compaction completes
MutexLock l(&mutex_);
while (imm_ != NULL && bg_error_.ok()) {
bg_cv_.Wait();
}
if (imm_ != NULL) {
s = bg_error_;
}
return s;
}
Status DBImpl::TEST_CompactMemTable() {
return FlushMemTable(FlushOptions());
}
Status DBImpl::TEST_WaitForCompactMemTable() {
return WaitForCompactMemTable();
return WaitForCompactMemTable();
}
Status DBImpl::TEST_WaitForCompact() {
// Wait until the compaction completes
MutexLock l(&mutex_);
while (bg_compaction_scheduled_ && bg_error_.ok()) {
bg_cv_.Wait();
}
return bg_error_;
// Wait until the compaction completes
MutexLock l(&mutex_);
while (bg_compaction_scheduled_ && bg_error_.ok()) {
bg_cv_.Wait();
}
return bg_error_;
}
void DBImpl::MaybeScheduleCompaction() {
@ -656,6 +696,8 @@ void DBImpl::BackgroundCall() {
}
bg_compaction_scheduled_ = false;
MaybeScheduleLogDBDeployStats();
// Previous compaction may have produced too many files in a level,
// so reschedule another compaction if needed.
MaybeScheduleCompaction();
@ -868,6 +910,9 @@ Status DBImpl::DoCompactionWork(CompactionState* compact) {
compact->compaction->level(),
compact->compaction->num_input_files(1),
compact->compaction->level() + 1);
char scratch[200];
compact->compaction->Summary(scratch, 256);
Log(options_.info_log, "Compaction start summary: %s\n", scratch);
assert(versions_->NumLevelFiles(compact->compaction->level()) > 0);
assert(compact->builder == NULL);
@ -1457,6 +1502,7 @@ Status DB::Open(const Options& options, const std::string& dbname,
if (s.ok()) {
impl->DeleteObsoleteFiles();
impl->MaybeScheduleCompaction();
impl->MaybeScheduleLogDBDeployStats();
}
}
impl->mutex_.Unlock();
@ -1502,4 +1548,13 @@ Status DestroyDB(const std::string& dbname, const Options& options) {
return result;
}
//
// A global method that can dump out the build version
void printLeveldbBuildVersion() {
printf("Git sha %s", leveldb_build_git_sha);
printf("Git datetime %s", leveldb_build_git_datetime);
printf("Compile time %s", leveldb_build_compile_time);
printf("Compile date %s", leveldb_build_compile_date);
}
} // namespace leveldb

@ -13,6 +13,13 @@
#include "leveldb/db.h"
#include "leveldb/env.h"
#include "port/port.h"
#include "util/stats_logger.h"
#ifdef USE_SCRIBE
#include "scribe/scribe_logger.h"
#endif
#include <boost/lexical_cast.hpp>
namespace leveldb {
@ -107,6 +114,9 @@ class DBImpl : public DB {
// Wait for memtable compaction
Status WaitForCompactMemTable();
void MaybeScheduleLogDBDeployStats();
static void LogDBDeployStats(void* db);
void MaybeScheduleCompaction();
static void BGWork(void* db);
void BackgroundCall();
@ -144,6 +154,8 @@ class DBImpl : public DB {
uint64_t logfile_number_;
log::Writer* log_;
std::string host_name_;
// Queue of writers.
std::deque<Writer*> writers_;
WriteBatch* tmp_batch_;
@ -172,6 +184,10 @@ class DBImpl : public DB {
// Have we encountered a background error in paranoid mode?
Status bg_error_;
StatsLogger* logger_;
int64_t volatile last_log_ts;
// Per level compaction stats. stats_[level] stores the stats for
// compactions that produced data for the specified "level".
struct CompactionStats {
@ -189,6 +205,8 @@ class DBImpl : public DB {
};
CompactionStats* stats_;
static const int KEEP_LOG_FILE_NUM = 1000;
// No copying allowed
DBImpl(const DBImpl&);
void operator=(const DBImpl&);

@ -0,0 +1,72 @@
// 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 "db/db_impl.h"
#include <string>
#include <stdint.h>
#include "db/version_set.h"
#include "leveldb/db.h"
#include "leveldb/env.h"
namespace leveldb {
void DBImpl::MaybeScheduleLogDBDeployStats() {
// There is a lock in the actual logger.
if (!logger_ || options_.db_stats_log_interval < 0
|| host_name_.empty()) {
return;
}
if (shutting_down_.Acquire_Load()) {
// Already scheduled
} else {
int64_t current_ts = 0;
Status st = env_->GetCurrentTime(&current_ts);
if (!st.ok()) {
return;
}
if ((current_ts - last_log_ts) < options_.db_stats_log_interval) {
return;
}
last_log_ts = current_ts;
env_->Schedule(&DBImpl::LogDBDeployStats, this);
}
}
void DBImpl::LogDBDeployStats(void* db) {
DBImpl* db_inst = reinterpret_cast<DBImpl*>(db);
if (db_inst->shutting_down_.Acquire_Load()) {
return;
}
std::string version_info;
version_info += boost::lexical_cast<std::string>(kMajorVersion);
version_info += ".";
version_info += boost::lexical_cast<std::string>(kMinorVersion);
std::string data_dir;
db_inst->env_->GetAbsolutePath(db_inst->dbname_, &data_dir);
uint64_t file_total_size = 0;
uint32_t file_total_num = 0;
for (int i = 0; i < db_inst->versions_->NumberLevels(); i++) {
file_total_num += db_inst->versions_->NumLevelFiles(i);
file_total_size += db_inst->versions_->NumLevelBytes(i);
}
VersionSet::LevelSummaryStorage scratch;
const char* file_num_summary = db_inst->versions_->LevelSummary(&scratch);
std::string file_num_per_level(file_num_summary);
const char* file_size_summary = db_inst->versions_->LevelDataSizeSummary(
&scratch);
std::string data_size_per_level(file_num_summary);
int64_t unix_ts;
db_inst->env_->GetCurrentTime(&unix_ts);
db_inst->logger_->Log_Deploy_Stats(version_info, db_inst->host_name_,
data_dir, file_total_size, file_total_num, file_num_per_level,
data_size_per_level, unix_ts);
}
}

@ -213,6 +213,10 @@ class DBTest {
ASSERT_OK(TryReopen(options));
}
Status PureReopen(Options* options, DB** db) {
return DB::Open(*options, dbname_, db);
}
Status TryReopen(Options* options) {
delete db_;
db_ = NULL;
@ -779,6 +783,22 @@ TEST(DBTest, Recover) {
} while (ChangeOptions());
}
TEST(DBTest, RollLog) {
do {
ASSERT_OK(Put("foo", "v1"));
ASSERT_OK(Put("baz", "v5"));
Reopen();
for (int i = 0; i < 10; i++) {
Reopen();
}
ASSERT_OK(Put("foo", "v4"));
for (int i = 0; i < 10; i++) {
Reopen();
}
} while (ChangeOptions());
}
TEST(DBTest, WAL) {
Options options = CurrentOptions();
WriteOptions writeOpt = WriteOptions();
@ -812,6 +832,13 @@ TEST(DBTest, WAL) {
ASSERT_EQ("v3", Get("foo"));
}
TEST(DBTest, CheckLock) {
DB* localdb;
Options options = CurrentOptions();
ASSERT_TRUE(TryReopen(&options).ok());
ASSERT_TRUE(!(PureReopen(&options, &localdb).ok())); // second open should fail
}
TEST(DBTest, FLUSH) {
Options options = CurrentOptions();
WriteOptions writeOpt = WriteOptions();

@ -60,8 +60,10 @@ std::string InfoLogFileName(const std::string& dbname) {
}
// Return the name of the old info log file for "dbname".
std::string OldInfoLogFileName(const std::string& dbname) {
return dbname + "/LOG.old";
std::string OldInfoLogFileName(const std::string& dbname, uint64_t ts) {
char buf[50];
snprintf(buf, sizeof(buf), "%llu", static_cast<unsigned long long>(ts));
return dbname + "/LOG.old." + buf;
}
@ -69,7 +71,7 @@ std::string OldInfoLogFileName(const std::string& dbname) {
// dbname/CURRENT
// dbname/LOCK
// dbname/LOG
// dbname/LOG.old
// dbname/LOG.old.[0-9]+
// dbname/MANIFEST-[0-9]+
// dbname/[0-9]+.(log|sst)
bool ParseFileName(const std::string& fname,
@ -82,9 +84,17 @@ bool ParseFileName(const std::string& fname,
} else if (rest == "LOCK") {
*number = 0;
*type = kDBLockFile;
} else if (rest == "LOG" || rest == "LOG.old") {
} else if (rest == "LOG") {
*number = 0;
*type = kInfoLogFile;
} else if (rest.starts_with("LOG.old.")) {
uint64_t ts_suffix;
rest.remove_prefix(sizeof("LOG.old."));
if (!ConsumeDecimalNumber(&rest, &ts_suffix)) {
return false;
}
*number = ts_suffix;
*type = kInfoLogFile;
} else if (rest.starts_with("MANIFEST-")) {
rest.remove_prefix(strlen("MANIFEST-"));
uint64_t num;

@ -60,7 +60,7 @@ extern std::string TempFileName(const std::string& dbname, uint64_t number);
extern std::string InfoLogFileName(const std::string& dbname);
// Return the name of the old info log file for "dbname".
extern std::string OldInfoLogFileName(const std::string& dbname);
extern std::string OldInfoLogFileName(const std::string& dbname, uint64_t ts);
// If filename is a leveldb file, store the type of the file in *type.
// The number encoded in the filename is stored in *number. If the

@ -702,15 +702,16 @@ VersionSet::VersionSet(const std::string& dbname,
compact_pointer_ = new std::string[options_->num_levels];
max_file_size_ = new uint64_t[options_->num_levels];
level_max_bytes_ = new uint64_t[options->num_levels];
max_file_size_[0] = options_->target_file_size_base;
level_max_bytes_[0] = options_->max_bytes_for_level_base;
int target_file_size_multiplier = options_->target_file_size_multiplier;
int max_bytes_multiplier = options_->max_bytes_for_level_multiplier;
int i = 1;
while (i < options_->num_levels) {
max_file_size_[i] = max_file_size_[i-1] * target_file_size_multiplier;
level_max_bytes_[i] = level_max_bytes_[i-1] * max_bytes_multiplier;
i++;
for (int i = 0; i < options_->num_levels; i++) {
if (i > 1) {
max_file_size_[i] = max_file_size_[i-1] * target_file_size_multiplier;
level_max_bytes_[i] = level_max_bytes_[i-1] * max_bytes_multiplier;
} else {
max_file_size_[i] = options_->target_file_size_base;
level_max_bytes_[i] = options_->max_bytes_for_level_base;
}
}
AppendVersion(new Version(this));
}
@ -939,6 +940,119 @@ Status VersionSet::Recover() {
return s;
}
Status VersionSet::DumpManifest(Options& options, std::string& dscname) {
struct LogReporter : public log::Reader::Reporter {
Status* status;
virtual void Corruption(size_t bytes, const Status& s) {
if (this->status->ok()) *this->status = s;
}
};
// Open the specified manifest file.
SequentialFile* file;
Status s = options.env->NewSequentialFile(dscname, &file);
if (!s.ok()) {
return s;
}
bool have_log_number = false;
bool have_prev_log_number = false;
bool have_next_file = false;
bool have_last_sequence = false;
uint64_t next_file = 0;
uint64_t last_sequence = 0;
uint64_t log_number = 0;
uint64_t prev_log_number = 0;
VersionSet::Builder builder(this, current_);
{
LogReporter reporter;
reporter.status = &s;
log::Reader reader(file, &reporter, true/*checksum*/, 0/*initial_offset*/);
Slice record;
std::string scratch;
while (reader.ReadRecord(&record, &scratch) && s.ok()) {
VersionEdit edit(NumberLevels());
s = edit.DecodeFrom(record);
if (s.ok()) {
if (edit.has_comparator_ &&
edit.comparator_ != icmp_.user_comparator()->Name()) {
s = Status::InvalidArgument(
edit.comparator_ + "does not match existing comparator ",
icmp_.user_comparator()->Name());
}
}
if (s.ok()) {
builder.Apply(&edit);
}
if (edit.has_log_number_) {
log_number = edit.log_number_;
have_log_number = true;
}
if (edit.has_prev_log_number_) {
prev_log_number = edit.prev_log_number_;
have_prev_log_number = true;
}
if (edit.has_next_file_number_) {
next_file = edit.next_file_number_;
have_next_file = true;
}
if (edit.has_last_sequence_) {
last_sequence = edit.last_sequence_;
have_last_sequence = true;
}
}
}
delete file;
file = NULL;
if (s.ok()) {
if (!have_next_file) {
s = Status::Corruption("no meta-nextfile entry in descriptor");
printf("no meta-nextfile entry in descriptor");
} else if (!have_log_number) {
s = Status::Corruption("no meta-lognumber entry in descriptor");
printf("no meta-lognumber entry in descriptor");
} else if (!have_last_sequence) {
printf("no last-sequence-number entry in descriptor");
s = Status::Corruption("no last-sequence-number entry in descriptor");
}
if (!have_prev_log_number) {
prev_log_number = 0;
}
MarkFileNumberUsed(prev_log_number);
MarkFileNumberUsed(log_number);
}
if (s.ok()) {
Version* v = new Version(this);
builder.SaveTo(v);
// Install recovered version
Finalize(v);
AppendVersion(v);
manifest_file_number_ = next_file;
next_file_number_ = next_file + 1;
last_sequence_ = last_sequence;
log_number_ = log_number;
prev_log_number_ = prev_log_number;
printf("manifest_file_number %d next_file_number %d last_sequence %d log_number %d prev_log_number %d\n",
manifest_file_number_, next_file_number_,
last_sequence, log_number, prev_log_number);
printf("%s \n", v->DebugString().c_str());
}
return s;
}
void VersionSet::MarkFileNumberUsed(uint64_t number) {
if (next_file_number_ <= number) {
next_file_number_ = number + 1;
@ -1032,6 +1146,21 @@ const char* VersionSet::LevelSummary(LevelSummaryStorage* scratch) const {
return scratch->buffer;
}
const char* VersionSet::LevelDataSizeSummary(
LevelSummaryStorage* scratch) const {
int len = snprintf(scratch->buffer, sizeof(scratch->buffer), "files_size[");
for (int i = 0; i < NumberLevels(); i++) {
int sz = sizeof(scratch->buffer) - len;
int ret = snprintf(scratch->buffer + len, sz, "%ld ",
NumLevelBytes(i));
if (ret < 0 || ret >= sz)
break;
len += ret;
}
snprintf(scratch->buffer + len, sizeof(scratch->buffer) - len, "]");
return scratch->buffer;
}
uint64_t VersionSet::ApproximateOffsetOf(Version* v, const InternalKey& ikey) {
uint64_t result = 0;
for (int level = 0; level < NumberLevels(); level++) {
@ -1443,4 +1572,37 @@ void Compaction::ReleaseInputs() {
}
}
static void InputSummary(std::vector<FileMetaData*>& files,
char* output,
int len) {
int write = 0;
for (int i = 0; i < files.size(); i++) {
int sz = len - write;
int ret = snprintf(output + write, sz, "%llu(%llu) ",
files.at(i)->number,
files.at(i)->file_size);
if (ret < 0 || ret >= sz)
break;
write += ret;
}
}
void Compaction::Summary(char* output, int len) {
int write = snprintf(output, len, "Base level %d, inputs:", level_);
if(write < 0 || write > len)
return;
char level_low_summary[100];
InputSummary(inputs_[0], level_low_summary, 100);
char level_up_summary[100];
if (inputs_[1].size()) {
InputSummary(inputs_[1], level_up_summary, 100);
} else {
level_up_summary[0] = '\0';
}
snprintf(output + write, len - write, "[%s],[%s]",
level_low_summary, level_up_summary);
}
} // namespace leveldb

@ -238,6 +238,13 @@ class VersionSet {
};
const char* LevelSummary(LevelSummaryStorage* scratch) const;
// printf contents (for debugging)
Status DumpManifest(Options& options, std::string& manifestFileName);
// Return a human-readable short (single-line) summary of the data size
// of files per level. Uses *scratch as backing store.
const char* LevelDataSizeSummary(LevelSummaryStorage* scratch) const;
private:
class Builder;
@ -344,6 +351,8 @@ class Compaction {
// is successful.
void ReleaseInputs();
void Summary(char* output, int len);
private:
friend class Version;
friend class VersionSet;

@ -8,6 +8,13 @@ TOOLCHAIN_REV=d28c90311ca14f9f0b2bb720f4e34b285513d4f4
TOOLCHAIN_EXECUTABLES="/mnt/gvfs/third-party/$TOOLCHAIN_REV/centos5.2-native"
TOOLCHAIN_LIB_BASE="/mnt/gvfs/third-party/$TOOLCHAIN_REV/gcc-4.6.2-glibc-2.13"
# always build thrift server
export USE_THRIFT=1
if ! test "$NO_SCRIBE"; then
export USE_SCRIBE=1
fi
# location of libhdfs libraries
if test "$USE_HDFS"; then
JAVA_HOME="/usr/local/jdk-6u22-64"

@ -122,6 +122,20 @@ class HdfsEnv : public Env {
posixEnv->SleepForMicroseconds(micros);
}
virtual Status GetHostName(char* name, uint len) {
return posixEnv->GetHostName(name, len);
}
virtual Status GetCurrentTime(int64_t* unix_time) {
return posixEnv->NowUnixTime(unix_time);
}
virtual Status GetAbsolutePath(const std::string& db_path,
std::string* output_path) {
return posixEnv->GetAbsolutePath(db_path, output_path);
}
static uint64_t gettid() {
assert(sizeof(pthread_t) <= sizeof(uint64_t));
return (uint64_t)pthread_self();
@ -245,6 +259,14 @@ class HdfsEnv : public Env {
virtual uint64_t NowMicros() {}
virtual void SleepForMicroseconds(int micros) {}
virtual Status GetHostName(char* name, uint len) {}
virtual Status GetCurrentTime(int64_t* unix_time) {}
virtual Status GetAbsolutePath(const std::string& db_path,
std::string* outputpath) {}
};
}

@ -145,6 +145,16 @@ class Env {
// Sleep/delay the thread for the perscribed number of micro-seconds.
virtual void SleepForMicroseconds(int micros) = 0;
// Get the current host name.
virtual Status GetHostName(char* name, uint len) = 0;
// Get the number of seconds since the Epoch, 1970-01-01 00:00:00 (UTC).
virtual Status GetCurrentTime(int64_t* unix_time) = 0;
// Get full directory name for this db.
virtual Status GetAbsolutePath(const std::string& db_path,
std::string* output_path) = 0;
private:
// No copying allowed
Env(const Env&);
@ -314,6 +324,17 @@ class EnvWrapper : public Env {
void SleepForMicroseconds(int micros) {
target_->SleepForMicroseconds(micros);
}
Status GetHostName(char* name, uint len) {
return target_->GetHostName(name, len);
}
Status GetCurrentTime(int64_t* unix_time) {
return target_->GetCurrentTime(unix_time);
}
Status GetAbsolutePath(const std::string& db_path,
std::string* output_path) {
return target_->GetAbsolutePath(db_path, output_path);
}
private:
Env* target_;
};

@ -152,17 +152,17 @@ struct Options {
// (max_bytes_for_level_base)^(max_bytes_for_level_multiplier).
int max_bytes_for_level_base;
int max_bytes_for_level_multiplier;
int max_bytes_for_level_multiplier;
// Maximum number of bytes in all compacted files. We avoid expanding
// the lower level file set of a compaction if it would make the
// total compaction cover more than
// (expanded_compaction_factor * targetFileSizeLevel()) many bytes.
int expanded_compaction_factor;
// Maximum number of bytes in all compacted files. We avoid expanding
// the lower level file set of a compaction if it would make the
// total compaction cover more than
// (expanded_compaction_factor * targetFileSizeLevel()) many bytes.
int expanded_compaction_factor;
// Control maximum bytes of overlaps in grandparent (i.e., level+2) before we
// stop building a single file in a level->level+1 compaction.
int max_grandparent_overlap_factor;
// Control maximum bytes of overlaps in grandparent (i.e., level+2) before we
// stop building a single file in a level->level+1 compaction.
int max_grandparent_overlap_factor;
// Compress blocks using the specified compression algorithm. This
// parameter can be changed dynamically.
@ -198,8 +198,16 @@ struct Options {
// Default: false
bool disableDataSync;
// This number controls how often a new scribe log about
// db deploy stats is written out.
// -1 indicates no logging at all.
// Default value is 1800 (half an hour).
int db_stats_log_interval;
// Create an Options object with default values for all fields.
Options();
void Dump(Logger * log) const;
};
// Options that control read operations

File diff suppressed because it is too large Load Diff

@ -0,0 +1,593 @@
/**
* Autogenerated by Thrift
*
* DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
* @generated
*/
#ifndef _Tleveldb_scribe_H
#define _Tleveldb_scribe_H
#include <TDispatchProcessor.h>
#include "scribe_types.h"
namespace Tleveldb {
class scribeIf {
public:
virtual ~scribeIf() {}
virtual ResultCode Log(const std::vector<LogEntry> & messages) = 0;
virtual void LogMulti(std::vector<ResultCode> & _return, const std::vector<LogEntry> & messages) = 0;
virtual ResultCode LogCompressedMsg(const std::string& compressedMessages) = 0;
};
class scribeIfFactory {
public:
typedef scribeIf Handler;
virtual ~scribeIfFactory() {}
virtual scribeIf* getHandler(::apache::thrift::server::TConnectionContext* ctx) = 0;
virtual void releaseHandler(scribeIf* handler) = 0;
};
class scribeIfSingletonFactory : virtual public scribeIfFactory {
public:
scribeIfSingletonFactory(const boost::shared_ptr<scribeIf>& iface) : iface_(iface) {}
virtual ~scribeIfSingletonFactory() {}
virtual scribeIf* getHandler(::apache::thrift::server::TConnectionContext*) {
return iface_.get();
}
virtual void releaseHandler(scribeIf* handler) {}
protected:
boost::shared_ptr<scribeIf> iface_;
};
class scribeNull : virtual public scribeIf {
public:
virtual ~scribeNull() {}
ResultCode Log(const std::vector<LogEntry> & /* messages */) {
ResultCode _return = (ResultCode)0;
return _return;
}
void LogMulti(std::vector<ResultCode> & /* _return */, const std::vector<LogEntry> & /* messages */) {
return;
}
ResultCode LogCompressedMsg(const std::string& /* compressedMessages */) {
ResultCode _return = (ResultCode)0;
return _return;
}
};
class scribe_Log_args {
public:
static const uint64_t _reflection_id = 5902265217339133004U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
scribe_Log_args() {
}
scribe_Log_args(const scribe_Log_args&) = default;
scribe_Log_args& operator=(const scribe_Log_args&) = default;
scribe_Log_args(scribe_Log_args&&) = default;
scribe_Log_args& operator=(scribe_Log_args&&) = default;
void __clear() {
messages.clear();
__isset.__clear();
}
virtual ~scribe_Log_args() throw() {}
std::vector<LogEntry> messages;
struct __isset {
__isset() { __clear(); }
void __clear() {
messages = false;
}
bool messages;
} __isset;
bool operator == (const scribe_Log_args & rhs) const
{
if (!(this->messages == rhs.messages))
return false;
return true;
}
bool operator != (const scribe_Log_args &rhs) const {
return !(*this == rhs);
}
bool operator < (const scribe_Log_args & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_Log_pargs {
public:
static const uint64_t _reflection_id = 5555604010648986412U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
virtual ~scribe_Log_pargs() throw() {}
const std::vector<LogEntry> * messages;
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_Log_result {
public:
static const uint64_t _reflection_id = 18205781396971565932U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
scribe_Log_result() : success(static_cast<ResultCode>(0)) {
}
scribe_Log_result(const scribe_Log_result&) = default;
scribe_Log_result& operator=(const scribe_Log_result&) = default;
scribe_Log_result(scribe_Log_result&&) = default;
scribe_Log_result& operator=(scribe_Log_result&&) = default;
void __clear() {
success = static_cast<ResultCode>(0);
__isset.__clear();
}
virtual ~scribe_Log_result() throw() {}
ResultCode success;
struct __isset {
__isset() { __clear(); }
void __clear() {
success = false;
}
bool success;
} __isset;
bool operator == (const scribe_Log_result & rhs) const
{
if (!(this->success == rhs.success))
return false;
return true;
}
bool operator != (const scribe_Log_result &rhs) const {
return !(*this == rhs);
}
bool operator < (const scribe_Log_result & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_Log_presult {
public:
static const uint64_t _reflection_id = 12945584136895385836U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
virtual ~scribe_Log_presult() throw() {}
ResultCode* success;
struct __isset {
__isset() { __clear(); }
void __clear() {
success = false;
}
bool success;
} __isset;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
};
class scribe_LogMulti_args {
public:
static const uint64_t _reflection_id = 7590876486278061516U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
scribe_LogMulti_args() {
}
scribe_LogMulti_args(const scribe_LogMulti_args&) = default;
scribe_LogMulti_args& operator=(const scribe_LogMulti_args&) = default;
scribe_LogMulti_args(scribe_LogMulti_args&&) = default;
scribe_LogMulti_args& operator=(scribe_LogMulti_args&&) = default;
void __clear() {
messages.clear();
__isset.__clear();
}
virtual ~scribe_LogMulti_args() throw() {}
std::vector<LogEntry> messages;
struct __isset {
__isset() { __clear(); }
void __clear() {
messages = false;
}
bool messages;
} __isset;
bool operator == (const scribe_LogMulti_args & rhs) const
{
if (!(this->messages == rhs.messages))
return false;
return true;
}
bool operator != (const scribe_LogMulti_args &rhs) const {
return !(*this == rhs);
}
bool operator < (const scribe_LogMulti_args & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_LogMulti_pargs {
public:
static const uint64_t _reflection_id = 9124384543551655628U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
virtual ~scribe_LogMulti_pargs() throw() {}
const std::vector<LogEntry> * messages;
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_LogMulti_result {
public:
static const uint64_t _reflection_id = 4828367046341273164U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
scribe_LogMulti_result() {
}
scribe_LogMulti_result(const scribe_LogMulti_result&) = default;
scribe_LogMulti_result& operator=(const scribe_LogMulti_result&) = default;
scribe_LogMulti_result(scribe_LogMulti_result&&) = default;
scribe_LogMulti_result& operator=(scribe_LogMulti_result&&) = default;
void __clear() {
success.clear();
__isset.__clear();
}
virtual ~scribe_LogMulti_result() throw() {}
std::vector<ResultCode> success;
struct __isset {
__isset() { __clear(); }
void __clear() {
success = false;
}
bool success;
} __isset;
bool operator == (const scribe_LogMulti_result & rhs) const
{
if (!(this->success == rhs.success))
return false;
return true;
}
bool operator != (const scribe_LogMulti_result &rhs) const {
return !(*this == rhs);
}
bool operator < (const scribe_LogMulti_result & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_LogMulti_presult {
public:
static const uint64_t _reflection_id = 5642041737363050316U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
virtual ~scribe_LogMulti_presult() throw() {}
std::vector<ResultCode> * success;
struct __isset {
__isset() { __clear(); }
void __clear() {
success = false;
}
bool success;
} __isset;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
};
class scribe_LogCompressedMsg_args {
public:
static const uint64_t _reflection_id = 12705053036625273964U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
scribe_LogCompressedMsg_args() : compressedMessages("") {
}
scribe_LogCompressedMsg_args(const scribe_LogCompressedMsg_args&) = default;
scribe_LogCompressedMsg_args& operator=(const scribe_LogCompressedMsg_args&) = default;
scribe_LogCompressedMsg_args(scribe_LogCompressedMsg_args&&) = default;
scribe_LogCompressedMsg_args& operator=(scribe_LogCompressedMsg_args&&) = default;
void __clear() {
compressedMessages = "";
__isset.__clear();
}
virtual ~scribe_LogCompressedMsg_args() throw() {}
std::string compressedMessages;
struct __isset {
__isset() { __clear(); }
void __clear() {
compressedMessages = false;
}
bool compressedMessages;
} __isset;
bool operator == (const scribe_LogCompressedMsg_args & rhs) const
{
if (!(this->compressedMessages == rhs.compressedMessages))
return false;
return true;
}
bool operator != (const scribe_LogCompressedMsg_args &rhs) const {
return !(*this == rhs);
}
bool operator < (const scribe_LogCompressedMsg_args & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_LogCompressedMsg_pargs {
public:
static const uint64_t _reflection_id = 13645577436870531500U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
virtual ~scribe_LogCompressedMsg_pargs() throw() {}
const std::string* compressedMessages;
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_LogCompressedMsg_result {
public:
static const uint64_t _reflection_id = 15026639991904524972U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
scribe_LogCompressedMsg_result() : success(static_cast<ResultCode>(0)) {
}
scribe_LogCompressedMsg_result(const scribe_LogCompressedMsg_result&) = default;
scribe_LogCompressedMsg_result& operator=(const scribe_LogCompressedMsg_result&) = default;
scribe_LogCompressedMsg_result(scribe_LogCompressedMsg_result&&) = default;
scribe_LogCompressedMsg_result& operator=(scribe_LogCompressedMsg_result&&) = default;
void __clear() {
success = static_cast<ResultCode>(0);
__isset.__clear();
}
virtual ~scribe_LogCompressedMsg_result() throw() {}
ResultCode success;
struct __isset {
__isset() { __clear(); }
void __clear() {
success = false;
}
bool success;
} __isset;
bool operator == (const scribe_LogCompressedMsg_result & rhs) const
{
if (!(this->success == rhs.success))
return false;
return true;
}
bool operator != (const scribe_LogCompressedMsg_result &rhs) const {
return !(*this == rhs);
}
bool operator < (const scribe_LogCompressedMsg_result & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class scribe_LogCompressedMsg_presult {
public:
static const uint64_t _reflection_id = 5311776576442573772U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
virtual ~scribe_LogCompressedMsg_presult() throw() {}
ResultCode* success;
struct __isset {
__isset() { __clear(); }
void __clear() {
success = false;
}
bool success;
} __isset;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
};
class scribeClient : virtual public scribeIf, virtual public apache::thrift::TClientBase {
public:
scribeClient(boost::shared_ptr<apache::thrift::protocol::TProtocol> prot) :
checkSeqid_(true),
nextSendSequenceId_(1),
nextRecvSequenceId_(1),
piprot_(prot),
poprot_(prot) {
iprot_ = prot.get();
oprot_ = prot.get();
}
scribeClient(boost::shared_ptr<apache::thrift::protocol::TProtocol> iprot, boost::shared_ptr<apache::thrift::protocol::TProtocol> oprot) :
checkSeqid_(true),
nextSendSequenceId_(1),
nextRecvSequenceId_(1),
piprot_(iprot),
poprot_(oprot) {
iprot_ = iprot.get();
oprot_ = oprot.get();
}
boost::shared_ptr<apache::thrift::protocol::TProtocol> getInputProtocol() {
return piprot_;
}
boost::shared_ptr<apache::thrift::protocol::TProtocol> getOutputProtocol() {
return poprot_;
}
ResultCode Log(const std::vector<LogEntry> & messages);
void send_Log(const std::vector<LogEntry> & messages);
ResultCode recv_Log();
void LogMulti(std::vector<ResultCode> & _return, const std::vector<LogEntry> & messages);
void send_LogMulti(const std::vector<LogEntry> & messages);
void recv_LogMulti(std::vector<ResultCode> & _return);
ResultCode LogCompressedMsg(const std::string& compressedMessages);
void send_LogCompressedMsg(const std::string& compressedMessages);
ResultCode recv_LogCompressedMsg();
/**
* Disable checking the seqid field in server responses.
*
* This should only be used with broken servers that return incorrect seqid values.
*/
void _disableSequenceIdChecks() {
checkSeqid_ = false;
}
protected:
bool checkSeqid_;
int32_t nextSendSequenceId_;
int32_t nextRecvSequenceId_;
int32_t getNextSendSequenceId();
int32_t getNextRecvSequenceId();
boost::shared_ptr<apache::thrift::protocol::TProtocol> piprot_;
boost::shared_ptr<apache::thrift::protocol::TProtocol> poprot_;
apache::thrift::protocol::TProtocol* iprot_;
apache::thrift::protocol::TProtocol* oprot_;
};
class scribeProcessor : public ::apache::thrift::TDispatchProcessor {
protected:
boost::shared_ptr<scribeIf> iface_;
virtual bool dispatchCall(apache::thrift::protocol::TProtocol* iprot, apache::thrift::protocol::TProtocol* oprot, const std::string& fname, int32_t seqid, apache::thrift::server::TConnectionContext* connectionContext);
private:
typedef void (scribeProcessor::*ProcessFunction)(int32_t, apache::thrift::protocol::TProtocol*, apache::thrift::protocol::TProtocol*, apache::thrift::server::TConnectionContext*);
typedef std::map<std::string, ProcessFunction> ProcessMap;
ProcessMap processMap_;
void process_Log(int32_t seqid, apache::thrift::protocol::TProtocol* iprot, apache::thrift::protocol::TProtocol* oprot, apache::thrift::server::TConnectionContext* connectionContext);
void process_LogMulti(int32_t seqid, apache::thrift::protocol::TProtocol* iprot, apache::thrift::protocol::TProtocol* oprot, apache::thrift::server::TConnectionContext* connectionContext);
void process_LogCompressedMsg(int32_t seqid, apache::thrift::protocol::TProtocol* iprot, apache::thrift::protocol::TProtocol* oprot, apache::thrift::server::TConnectionContext* connectionContext);
public:
scribeProcessor(boost::shared_ptr<scribeIf> iface) :
iface_(iface) {
processMap_["Log"] = &scribeProcessor::process_Log;
processMap_["LogMulti"] = &scribeProcessor::process_LogMulti;
processMap_["LogCompressedMsg"] = &scribeProcessor::process_LogCompressedMsg;
}
virtual ~scribeProcessor() {}
boost::shared_ptr<std::set<std::string> > getProcessFunctions() {
boost::shared_ptr<std::set<std::string> > rSet(new std::set<std::string>());
rSet->insert("scribe.Log");
rSet->insert("scribe.LogMulti");
rSet->insert("scribe.LogCompressedMsg");
return rSet;
}
};
class scribeProcessorFactory : public ::apache::thrift::TProcessorFactory {
public:
scribeProcessorFactory(const ::boost::shared_ptr< scribeIfFactory >& handlerFactory) :
handlerFactory_(handlerFactory) {}
::boost::shared_ptr< ::apache::thrift::TProcessor > getProcessor(::apache::thrift::server::TConnectionContext* ctx);
protected:
::boost::shared_ptr< scribeIfFactory > handlerFactory_;
};
class scribeMultiface : virtual public scribeIf {
public:
scribeMultiface(std::vector<boost::shared_ptr<scribeIf> >& ifaces) : ifaces_(ifaces) {
}
virtual ~scribeMultiface() {}
protected:
std::vector<boost::shared_ptr<scribeIf> > ifaces_;
scribeMultiface() {}
void add(boost::shared_ptr<scribeIf> iface) {
ifaces_.push_back(iface);
}
public:
ResultCode Log(const std::vector<LogEntry> & messages) {
uint32_t i;
uint32_t sz = ifaces_.size();
for (i = 0; i < sz - 1; ++i) {
ifaces_[i]->Log(messages);
}
return ifaces_[i]->Log(messages);
}
void LogMulti(std::vector<ResultCode> & _return, const std::vector<LogEntry> & messages) {
uint32_t i;
uint32_t sz = ifaces_.size();
for (i = 0; i < sz; ++i) {
ifaces_[i]->LogMulti(_return, messages);
}
}
ResultCode LogCompressedMsg(const std::string& compressedMessages) {
uint32_t i;
uint32_t sz = ifaces_.size();
for (i = 0; i < sz - 1; ++i) {
ifaces_[i]->LogCompressedMsg(compressedMessages);
}
return ifaces_[i]->LogCompressedMsg(compressedMessages);
}
};
} // namespace
#endif

@ -0,0 +1,17 @@
/**
* Autogenerated by Thrift
*
* DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
* @generated
*/
#include "scribe_constants.h"
namespace Tleveldb {
const scribeConstants g_scribe_constants;
scribeConstants::scribeConstants() {
SCRIBE_MAX_MESSAGE_LENGTH = 26214400;
}
} // namespace

@ -0,0 +1,25 @@
/**
* Autogenerated by Thrift
*
* DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
* @generated
*/
#ifndef scribe_CONSTANTS_H
#define scribe_CONSTANTS_H
#include "scribe_types.h"
namespace Tleveldb {
class scribeConstants {
public:
scribeConstants();
int32_t SCRIBE_MAX_MESSAGE_LENGTH;
};
extern const scribeConstants g_scribe_constants;
} // namespace
#endif

@ -0,0 +1,513 @@
/**
* Autogenerated by Thrift
*
* DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
* @generated
*/
#include "scribe_types.h"
#include <thrift/lib/cpp/Reflection.h>
#include <algorithm>
#include <string.h>
namespace Tleveldb {
int _kResultCodeValues[] = {
OK,
TRY_LATER,
ERROR_DECOMPRESS
};
const char* _kResultCodeNames[] = {
"OK",
"TRY_LATER",
"ERROR_DECOMPRESS"
};
const std::map<int, const char*> _ResultCode_VALUES_TO_NAMES(apache::thrift::TEnumIterator<int>(3, _kResultCodeValues, _kResultCodeNames), apache::thrift::TEnumIterator<int>(-1, NULL, NULL));
const std::map<const char*, int, apache::thrift::ltstr> _ResultCode_NAMES_TO_VALUES(apache::thrift::TEnumInverseIterator<int>(3, _kResultCodeValues, _kResultCodeNames), apache::thrift::TEnumInverseIterator<int>(-1, NULL, NULL));
} // namespace
namespace apache { namespace thrift {
template<>
const char* TEnumTraits< ::Tleveldb::ResultCode>::findName( ::Tleveldb::ResultCode value) {
return findName( ::Tleveldb::_ResultCode_VALUES_TO_NAMES, value);
}
template<>
bool TEnumTraits< ::Tleveldb::ResultCode>::findValue(const char* name, ::Tleveldb::ResultCode* out) {
return findValue( ::Tleveldb::_ResultCode_NAMES_TO_VALUES, name, out);
}
}} // apache::thrift
namespace Tleveldb {
// Reflection initializer for struct scribe.SourceInfo
namespace {
void reflectionInitializer_16557823557777806572(::apache::thrift::reflection::Schema& schema) {
const uint64_t id = 16557823557777806572U;
if (schema.dataTypes.count(id)) return;
::apache::thrift::reflection::DataType dt;
dt.name = "struct scribe.SourceInfo";
dt.__isset.fields = true;
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 1U;
f.name = "host";
dt.fields[1] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "port";
dt.fields[2] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 6U;
f.name = "timestamp";
dt.fields[3] = f;
}
schema.dataTypes[id] = dt;
schema.names[dt.name] = id;
}
} // namespace
const uint64_t SourceInfo::_reflection_id;
void SourceInfo::_reflection_register(::apache::thrift::reflection::Schema& schema) {
reflectionInitializer_16557823557777806572(schema);
}
uint32_t SourceInfo::read(apache::thrift::protocol::TProtocol* iprot) {
uint32_t xfer = 0;
std::string fname;
apache::thrift::protocol::TType ftype;
int16_t fid;
xfer += iprot->readStructBegin(fname);
using apache::thrift::protocol::TProtocolException;
while (true)
{
xfer += iprot->readFieldBegin(fname, ftype, fid);
if (ftype == apache::thrift::protocol::T_STOP) {
break;
}
switch (fid)
{
case 1:
if (ftype == apache::thrift::protocol::T_STRING) {
xfer += iprot->readBinary(this->host);
this->__isset.host = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 2:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->port);
this->__isset.port = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 3:
if (ftype == apache::thrift::protocol::T_I64) {
xfer += iprot->readI64(this->timestamp);
this->__isset.timestamp = true;
} else {
xfer += iprot->skip(ftype);
}
break;
default:
xfer += iprot->skip(ftype);
break;
}
xfer += iprot->readFieldEnd();
}
xfer += iprot->readStructEnd();
return xfer;
}
uint32_t SourceInfo::write(apache::thrift::protocol::TProtocol* oprot) const {
uint32_t xfer = 0;
xfer += oprot->writeStructBegin("SourceInfo");
xfer += oprot->writeFieldBegin("host", apache::thrift::protocol::T_STRING, 1);
xfer += oprot->writeBinary(this->host);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("port", apache::thrift::protocol::T_I32, 2);
xfer += oprot->writeI32(this->port);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("timestamp", apache::thrift::protocol::T_I64, 3);
xfer += oprot->writeI64(this->timestamp);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldStop();
xfer += oprot->writeStructEnd();
return xfer;
}
void swap(SourceInfo &a, SourceInfo &b) {
using ::std::swap;
(void)a;
(void)b;
swap(a.host, b.host);
swap(a.port, b.port);
swap(a.timestamp, b.timestamp);
swap(a.__isset, b.__isset);
}
// Reflection initializer for map<string, string>
namespace {
void reflectionInitializer_9246346592659763371(::apache::thrift::reflection::Schema& schema) {
const uint64_t id = 9246346592659763371U;
if (schema.dataTypes.count(id)) return;
::apache::thrift::reflection::DataType dt;
dt.name = "map<string, string>";
dt.__isset.mapKeyType = true;
dt.mapKeyType = 1U;
dt.__isset.valueType = true;
dt.valueType = 1U;
schema.dataTypes[id] = dt;
schema.names[dt.name] = id;
}
} // namespace
// Reflection initializer for struct scribe.LogEntry
namespace {
void reflectionInitializer_15053466696968532300(::apache::thrift::reflection::Schema& schema) {
const uint64_t id = 15053466696968532300U;
if (schema.dataTypes.count(id)) return;
reflectionInitializer_16557823557777806572(schema); // struct scribe.SourceInfo
reflectionInitializer_9246346592659763371(schema); // map<string, string>
::apache::thrift::reflection::DataType dt;
dt.name = "struct scribe.LogEntry";
dt.__isset.fields = true;
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 1U;
f.name = "category";
dt.fields[1] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 1U;
f.name = "message";
dt.fields[2] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = false;
f.type = 9246346592659763371U;
f.name = "metadata";
dt.fields[3] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = false;
f.type = 5U;
f.name = "checksum";
dt.fields[4] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = false;
f.type = 16557823557777806572U;
f.name = "source";
dt.fields[5] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = false;
f.type = 5U;
f.name = "bucket";
dt.fields[6] = f;
}
schema.dataTypes[id] = dt;
schema.names[dt.name] = id;
}
} // namespace
const uint64_t LogEntry::_reflection_id;
void LogEntry::_reflection_register(::apache::thrift::reflection::Schema& schema) {
reflectionInitializer_15053466696968532300(schema);
}
uint32_t LogEntry::read(apache::thrift::protocol::TProtocol* iprot) {
uint32_t xfer = 0;
std::string fname;
apache::thrift::protocol::TType ftype;
int16_t fid;
xfer += iprot->readStructBegin(fname);
using apache::thrift::protocol::TProtocolException;
while (true)
{
xfer += iprot->readFieldBegin(fname, ftype, fid);
if (ftype == apache::thrift::protocol::T_STOP) {
break;
}
switch (fid)
{
case 1:
if (ftype == apache::thrift::protocol::T_STRING) {
xfer += iprot->readBinary(this->category);
this->__isset.category = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 2:
if (ftype == apache::thrift::protocol::T_STRING) {
xfer += iprot->readBinary(this->message);
this->__isset.message = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 3:
if (ftype == apache::thrift::protocol::T_MAP) {
{
this->metadata.clear();
uint32_t _size0;
apache::thrift::protocol::TType _ktype1;
apache::thrift::protocol::TType _vtype2;
xfer += iprot->readMapBegin(_ktype1, _vtype2, _size0);
uint32_t _i4;
for (_i4 = 0; _i4 < _size0; ++_i4)
{
std::string _key5;
xfer += iprot->readString(_key5);
std::string& _val6 = this->metadata[_key5];
xfer += iprot->readString(_val6);
}
xfer += iprot->readMapEnd();
}
this->__isset.metadata = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 4:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->checksum);
this->__isset.checksum = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 5:
if (ftype == apache::thrift::protocol::T_STRUCT) {
xfer += this->source.read(iprot);
this->__isset.source = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 6:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->bucket);
this->__isset.bucket = true;
} else {
xfer += iprot->skip(ftype);
}
break;
default:
xfer += iprot->skip(ftype);
break;
}
xfer += iprot->readFieldEnd();
}
xfer += iprot->readStructEnd();
return xfer;
}
uint32_t LogEntry::write(apache::thrift::protocol::TProtocol* oprot) const {
uint32_t xfer = 0;
xfer += oprot->writeStructBegin("LogEntry");
xfer += oprot->writeFieldBegin("category", apache::thrift::protocol::T_STRING, 1);
xfer += oprot->writeBinary(this->category);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("message", apache::thrift::protocol::T_STRING, 2);
xfer += oprot->writeBinary(this->message);
xfer += oprot->writeFieldEnd();
if (this->__isset.metadata) {
xfer += oprot->writeFieldBegin("metadata", apache::thrift::protocol::T_MAP, 3);
{
xfer += oprot->writeMapBegin(apache::thrift::protocol::T_STRING, apache::thrift::protocol::T_STRING, this->metadata.size());
std::map<std::string, std::string> ::const_iterator _iter7;
for (_iter7 = this->metadata.begin(); _iter7 != this->metadata.end(); ++_iter7)
{
xfer += oprot->writeString(_iter7->first);
xfer += oprot->writeString(_iter7->second);
}
xfer += oprot->writeMapEnd();
}
xfer += oprot->writeFieldEnd();
}
if (this->__isset.checksum) {
xfer += oprot->writeFieldBegin("checksum", apache::thrift::protocol::T_I32, 4);
xfer += oprot->writeI32(this->checksum);
xfer += oprot->writeFieldEnd();
}
if (this->__isset.source) {
xfer += oprot->writeFieldBegin("source", apache::thrift::protocol::T_STRUCT, 5);
xfer += this->source.write(oprot);
xfer += oprot->writeFieldEnd();
}
if (this->__isset.bucket) {
xfer += oprot->writeFieldBegin("bucket", apache::thrift::protocol::T_I32, 6);
xfer += oprot->writeI32(this->bucket);
xfer += oprot->writeFieldEnd();
}
xfer += oprot->writeFieldStop();
xfer += oprot->writeStructEnd();
return xfer;
}
void swap(LogEntry &a, LogEntry &b) {
using ::std::swap;
(void)a;
(void)b;
swap(a.category, b.category);
swap(a.message, b.message);
swap(a.metadata, b.metadata);
swap(a.checksum, b.checksum);
swap(a.source, b.source);
swap(a.bucket, b.bucket);
swap(a.__isset, b.__isset);
}
// Reflection initializer for list<struct scribe.LogEntry>
namespace {
void reflectionInitializer_10251729064312664553(::apache::thrift::reflection::Schema& schema) {
const uint64_t id = 10251729064312664553U;
if (schema.dataTypes.count(id)) return;
reflectionInitializer_15053466696968532300(schema); // struct scribe.LogEntry
::apache::thrift::reflection::DataType dt;
dt.name = "list<struct scribe.LogEntry>";
dt.__isset.valueType = true;
dt.valueType = 15053466696968532300U;
schema.dataTypes[id] = dt;
schema.names[dt.name] = id;
}
} // namespace
// Reflection initializer for struct scribe.MessageList
namespace {
void reflectionInitializer_5674270912483072844(::apache::thrift::reflection::Schema& schema) {
const uint64_t id = 5674270912483072844U;
if (schema.dataTypes.count(id)) return;
reflectionInitializer_10251729064312664553(schema); // list<struct scribe.LogEntry>
::apache::thrift::reflection::DataType dt;
dt.name = "struct scribe.MessageList";
dt.__isset.fields = true;
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 10251729064312664553U;
f.name = "messages";
dt.fields[1] = f;
}
schema.dataTypes[id] = dt;
schema.names[dt.name] = id;
}
} // namespace
const uint64_t MessageList::_reflection_id;
void MessageList::_reflection_register(::apache::thrift::reflection::Schema& schema) {
reflectionInitializer_5674270912483072844(schema);
}
uint32_t MessageList::read(apache::thrift::protocol::TProtocol* iprot) {
uint32_t xfer = 0;
std::string fname;
apache::thrift::protocol::TType ftype;
int16_t fid;
xfer += iprot->readStructBegin(fname);
using apache::thrift::protocol::TProtocolException;
while (true)
{
xfer += iprot->readFieldBegin(fname, ftype, fid);
if (ftype == apache::thrift::protocol::T_STOP) {
break;
}
switch (fid)
{
case 1:
if (ftype == apache::thrift::protocol::T_LIST) {
{
this->messages.clear();
uint32_t _size8;
apache::thrift::protocol::TType _etype11;
xfer += iprot->readListBegin(_etype11, _size8);
this->messages.resize(_size8);
uint32_t _i12;
for (_i12 = 0; _i12 < _size8; ++_i12)
{
xfer += this->messages[_i12].read(iprot);
}
xfer += iprot->readListEnd();
}
this->__isset.messages = true;
} else {
xfer += iprot->skip(ftype);
}
break;
default:
xfer += iprot->skip(ftype);
break;
}
xfer += iprot->readFieldEnd();
}
xfer += iprot->readStructEnd();
return xfer;
}
uint32_t MessageList::write(apache::thrift::protocol::TProtocol* oprot) const {
uint32_t xfer = 0;
xfer += oprot->writeStructBegin("MessageList");
xfer += oprot->writeFieldBegin("messages", apache::thrift::protocol::T_LIST, 1);
{
xfer += oprot->writeListBegin(apache::thrift::protocol::T_STRUCT, this->messages.size());
std::vector<LogEntry> ::const_iterator _iter13;
for (_iter13 = this->messages.begin(); _iter13 != this->messages.end(); ++_iter13)
{
xfer += (*_iter13).write(oprot);
}
xfer += oprot->writeListEnd();
}
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldStop();
xfer += oprot->writeStructEnd();
return xfer;
}
void swap(MessageList &a, MessageList &b) {
using ::std::swap;
(void)a;
(void)b;
swap(a.messages, b.messages);
swap(a.__isset, b.__isset);
}
} // namespace

@ -0,0 +1,247 @@
/**
* Autogenerated by Thrift
*
* DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
* @generated
*/
#ifndef scribe_TYPES_H
#define scribe_TYPES_H
#include <Thrift.h>
#include <TApplicationException.h>
#include <protocol/TProtocol.h>
#include <transport/TTransport.h>
namespace apache { namespace thrift { namespace reflection {
class Schema;
}}}
namespace Tleveldb {
enum ResultCode {
OK = 0,
TRY_LATER = 1,
ERROR_DECOMPRESS = 2
};
extern const std::map<int, const char*> _ResultCode_VALUES_TO_NAMES;
extern const std::map<const char*, int, apache::thrift::ltstr> _ResultCode_NAMES_TO_VALUES;
} // namespace
namespace apache { namespace thrift {
template<>
inline constexpr ::Tleveldb::ResultCode TEnumTraits< ::Tleveldb::ResultCode>::min() {
return ::Tleveldb::ResultCode::OK;
}
template<>
inline constexpr ::Tleveldb::ResultCode TEnumTraits< ::Tleveldb::ResultCode>::max() {
return ::Tleveldb::ResultCode::ERROR_DECOMPRESS;
}
}} // apache:thrift
namespace Tleveldb {
class SourceInfo {
public:
static const uint64_t _reflection_id = 16557823557777806572U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
SourceInfo() : host(""), port(0), timestamp(0) {
}
SourceInfo(const SourceInfo&) = default;
SourceInfo& operator=(const SourceInfo&) = default;
SourceInfo(SourceInfo&&) = default;
SourceInfo& operator=(SourceInfo&&) = default;
void __clear() {
host = "";
port = 0;
timestamp = 0;
__isset.__clear();
}
virtual ~SourceInfo() throw() {}
std::string host;
int32_t port;
int64_t timestamp;
struct __isset {
__isset() { __clear(); }
void __clear() {
host = false;
port = false;
timestamp = false;
}
bool host;
bool port;
bool timestamp;
} __isset;
bool operator == (const SourceInfo & rhs) const
{
if (!(this->host == rhs.host))
return false;
if (!(this->port == rhs.port))
return false;
if (!(this->timestamp == rhs.timestamp))
return false;
return true;
}
bool operator != (const SourceInfo &rhs) const {
return !(*this == rhs);
}
bool operator < (const SourceInfo & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class SourceInfo;
void swap(SourceInfo &a, SourceInfo &b);
class LogEntry {
public:
static const uint64_t _reflection_id = 15053466696968532300U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
LogEntry() : category(""), message(""), checksum(0), bucket(0) {
}
LogEntry(const LogEntry&) = default;
LogEntry& operator=(const LogEntry&) = default;
LogEntry(LogEntry&&) = default;
LogEntry& operator=(LogEntry&&) = default;
void __clear() {
category = "";
message = "";
metadata.clear();
checksum = 0;
source.__clear();
bucket = 0;
__isset.__clear();
}
virtual ~LogEntry() throw() {}
std::string category;
std::string message;
std::map<std::string, std::string> metadata;
int32_t checksum;
SourceInfo source;
int32_t bucket;
struct __isset {
__isset() { __clear(); }
void __clear() {
category = false;
message = false;
metadata = false;
checksum = false;
source = false;
bucket = false;
}
bool category;
bool message;
bool metadata;
bool checksum;
bool source;
bool bucket;
} __isset;
bool operator == (const LogEntry & rhs) const
{
if (!(this->category == rhs.category))
return false;
if (!(this->message == rhs.message))
return false;
if (__isset.metadata != rhs.__isset.metadata)
return false;
else if (__isset.metadata && !(metadata == rhs.metadata))
return false;
if (__isset.checksum != rhs.__isset.checksum)
return false;
else if (__isset.checksum && !(checksum == rhs.checksum))
return false;
if (__isset.source != rhs.__isset.source)
return false;
else if (__isset.source && !(source == rhs.source))
return false;
if (__isset.bucket != rhs.__isset.bucket)
return false;
else if (__isset.bucket && !(bucket == rhs.bucket))
return false;
return true;
}
bool operator != (const LogEntry &rhs) const {
return !(*this == rhs);
}
bool operator < (const LogEntry & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class LogEntry;
void swap(LogEntry &a, LogEntry &b);
class MessageList {
public:
static const uint64_t _reflection_id = 5674270912483072844U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
MessageList() {
}
MessageList(const MessageList&) = default;
MessageList& operator=(const MessageList&) = default;
MessageList(MessageList&&) = default;
MessageList& operator=(MessageList&&) = default;
void __clear() {
messages.clear();
__isset.__clear();
}
virtual ~MessageList() throw() {}
std::vector<LogEntry> messages;
struct __isset {
__isset() { __clear(); }
void __clear() {
messages = false;
}
bool messages;
} __isset;
bool operator == (const MessageList & rhs) const
{
if (!(this->messages == rhs.messages))
return false;
return true;
}
bool operator != (const MessageList &rhs) const {
return !(*this == rhs);
}
bool operator < (const MessageList & ) const;
uint32_t read(apache::thrift::protocol::TProtocol* iprot);
uint32_t write(apache::thrift::protocol::TProtocol* oprot) const;
};
class MessageList;
void swap(MessageList &a, MessageList &b);
} // namespace
#endif

@ -0,0 +1,82 @@
#!/usr/local/bin/thrift --cpp --php
## Copyright (c) 2007-2012 Facebook
##
## Licensed under the Apache License, Version 2.0 (the "License");
## you may not use this file except in compliance with the License.
## You may obtain a copy of the License at
##
## http://www.apache.org/licenses/LICENSE-2.0
##
## Unless required by applicable law or agreed to in writing, software
## distributed under the License is distributed on an "AS IS" BASIS,
## WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
## See the License for the specific language governing permissions and
## limitations under the License.
##
## See accompanying file LICENSE or visit the Scribe site at:
## http://developers.facebook.com/scribe/
namespace cpp Tleveldb
namespace java Tleveldb
// Max message length allowed to log through scribe
const i32 SCRIBE_MAX_MESSAGE_LENGTH = 26214400;
enum ResultCode
{
OK,
TRY_LATER,
ERROR_DECOMPRESS
}
struct SourceInfo
{
1: binary host,
2: i32 port,
3: i64 timestamp
}
struct LogEntry
{
1: binary category,
2: binary message,
3: optional map<string, string> metadata,
4: optional i32 checksum,
5: optional SourceInfo source,
6: optional i32 bucket
}
struct MessageList
{
1: list<LogEntry> messages
}
service scribe
{
#
# Delivers a list of LogEntry messages to the Scribe server.
# A returned ResultCode of anything other than OK indicates that the
# whole batch was unable to be delivered to the server.
# If data loss is a concern, the caller should buffer and retry the messages.
#
ResultCode Log(1: list<LogEntry> messages);
#
# NOTE: FOR INTERNAL USE ONLY!
#
# Delivers a list of LogEntry messages to the Scribe server, but
# allows partial successes. A list of ResultCodes will be returned to
# indicate the success or failure of each message at the corresponding index.
# If data loss is a concern, the caller should retry only the failed messages.
#
list<ResultCode> LogMulti(1: list<LogEntry> messages);
#
# NOTE: FOR INTERNAL USE ONLY!
#
# The same as Log(...) except that the list of messages must first be
# serialized and compressed in some internal format.
#
ResultCode LogCompressedMsg(1: binary compressedMessages);
}

@ -0,0 +1,90 @@
#include "scribe_logger.h"
namespace leveldb {
const std::string ScribeLogger::COL_SEPERATOR = "\x1";
const std::string ScribeLogger::DEPLOY_STATS_CATEGORY = "leveldb_deploy_stats";
ScribeLogger::ScribeLogger(const std::string& host, int port,
int retry_times, uint32_t retry_intervals, int batch_size)
: host_(host),
port_(port),
retry_times_(retry_times),
retry_intervals_ (retry_intervals),
batch_size_ (batch_size) {
shared_ptr<TSocket> socket(new TSocket(host_, port_));
shared_ptr<TFramedTransport> framedTransport(new TFramedTransport(socket));
framedTransport->open();
shared_ptr<TBinaryProtocol> protocol(new TBinaryProtocol(framedTransport));
scribe_client_ = new scribeClient(protocol);
}
void ScribeLogger::Log(const std::string& category,
const std::string& message) {
LogEntry entry;
entry.category = category;
entry.message = message;
logger_mutex_.Lock();
logs_.push_back(entry);
if (logs_.size() >= batch_size_) {
ResultCode ret = scribe_client_->Log(logs_);
int retries_left = retry_times_;
while (ret == TRY_LATER && retries_left > 0) {
Env::Default()->SleepForMicroseconds(retry_intervals_);
ret = scribe_client_->Log(logs_);
retries_left--;
}
// Clear the local messages if either successfully write out
// or has failed in the last 10 calls.
if (ret == OK || logs_.size() > batch_size_ * 5) {
logs_.clear();
}
}
logger_mutex_.Unlock();
}
void ScribeLogger::MakeScribeMessage(std::string& output,
std::vector<std::string>& cols) {
int sz = cols.size();
int i = 0;
for (; i < sz - 1; i++) {
std::string& col = cols.at(i);
output += col;
output += ScribeLogger::COL_SEPERATOR;
}
std::string& col = cols.at(i);
output+=col;
}
void ScribeLogger::Log_Deploy_Stats(
const std::string& db_version,
const std::string& machine_info,
const std::string& data_dir,
const uint64_t data_size,
const uint32_t file_number,
const std::string& data_size_per_level,
const std::string& file_number_per_level,
const int64_t& ts_unix) {
std::string message;
std::vector<std::string> cols;
cols.push_back(db_version);
cols.push_back(machine_info);
cols.push_back(data_dir);
cols.push_back(boost::lexical_cast<std::string>(data_size));
cols.push_back(boost::lexical_cast<std::string>(file_number));
cols.push_back(data_size_per_level);
cols.push_back(file_number_per_level);
cols.push_back(boost::lexical_cast<std::string>(ts_unix));
MakeScribeMessage(message, cols);
return Log(ScribeLogger::DEPLOY_STATS_CATEGORY, message);
}
ScribeLogger::~ScribeLogger(){
delete scribe_client_;
}
}

@ -0,0 +1,71 @@
#ifndef SCRIBE_LOGGER_H_
#define SCRIBE_LOGGER_H_
#include "scribe/if/gen-cpp/scribe.h"
#include "scribe/if/gen-cpp/scribe_types.h"
#include "thrift/lib/cpp/protocol/TProtocol.h"
#include "thrift/lib/cpp/transport/TSocket.h"
#include "thrift/lib/cpp/protocol/TBinaryProtocol.h"
#include "thrift/lib/cpp/transport/TBufferTransports.h"
#include "leveldb/env.h"
#include "port/port.h"
#include "util/stats_logger.h"
#include "boost/lexical_cast.hpp"
using namespace Tleveldb;
using Tleveldb::scribeClient;
using namespace apache::thrift;
using namespace apache::thrift::protocol;
using namespace apache::thrift::transport;
using boost::shared_ptr;
using namespace ::Tleveldb;
namespace leveldb {
class ScribeLogger : public StatsLogger{
private:
std::string host_;
int port_;
int batch_size_;
scribeClient* scribe_client_;
std::vector<LogEntry> logs_;
port::Mutex logger_mutex_;
int retry_times_;
uint32_t retry_intervals_;
void MakeScribeMessage(std::string& output, std::vector<std::string>& cols);
public:
static const std::string COL_SEPERATOR;
static const std::string DEPLOY_STATS_CATEGORY;
ScribeLogger(const std::string& host, int port,
int retry_times=3, uint32_t retry_intervals=1000000,
int batch_size=1);
virtual ~ScribeLogger();
virtual void Log(const std::string& category, const std::string& message);
virtual void Log_Deploy_Stats(
const std::string& db_version,
const std::string& machine_info,
const std::string& data_dir,
const uint64_t data_size,
const uint32_t file_number,
const std::string& data_size_per_level,
const std::string& file_number_per_level,
const int64_t& ts_unix
);
};
}
#endif /* SCRIBE_LOGGER_H_ */

@ -21,5 +21,5 @@ You can run the leveldb server unit tests by
You can regenerate the thrift cpp files by doing the following
cd ./thrift
thrift --gen cpp if/leveldb.thrift
bin/thrift --gen cpp if/leveldb.thrift

Binary file not shown.

@ -422,6 +422,76 @@ void reflectionInitializer_6731746507948871532(::apache::thrift::reflection::Sch
f.name = "compression";
dt.fields[7] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "num_levels";
dt.fields[8] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "level0_file_num_compaction_trigger";
dt.fields[9] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "level0_slowdown_writes_trigger";
dt.fields[10] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "level0_stop_writes_trigger";
dt.fields[11] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "target_file_size_base";
dt.fields[12] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "target_file_size_multiplier";
dt.fields[13] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "max_bytes_for_level_base";
dt.fields[14] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "max_bytes_for_level_multiplier";
dt.fields[15] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 5U;
f.name = "max_grandparent_overlap_factor";
dt.fields[16] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 2U;
f.name = "disableDataSync";
dt.fields[17] = f;
}
schema.dataTypes[id] = dt;
schema.names[dt.name] = id;
}
@ -509,6 +579,86 @@ uint32_t DBOptions::read(apache::thrift::protocol::TProtocol* iprot) {
xfer += iprot->skip(ftype);
}
break;
case 8:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->num_levels);
this->__isset.num_levels = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 9:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->level0_file_num_compaction_trigger);
this->__isset.level0_file_num_compaction_trigger = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 10:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->level0_slowdown_writes_trigger);
this->__isset.level0_slowdown_writes_trigger = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 11:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->level0_stop_writes_trigger);
this->__isset.level0_stop_writes_trigger = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 12:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->target_file_size_base);
this->__isset.target_file_size_base = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 13:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->target_file_size_multiplier);
this->__isset.target_file_size_multiplier = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 14:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->max_bytes_for_level_base);
this->__isset.max_bytes_for_level_base = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 15:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->max_bytes_for_level_multiplier);
this->__isset.max_bytes_for_level_multiplier = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 16:
if (ftype == apache::thrift::protocol::T_I32) {
xfer += iprot->readI32(this->max_grandparent_overlap_factor);
this->__isset.max_grandparent_overlap_factor = true;
} else {
xfer += iprot->skip(ftype);
}
break;
case 17:
if (ftype == apache::thrift::protocol::T_BOOL) {
xfer += iprot->readBool(this->disableDataSync);
this->__isset.disableDataSync = true;
} else {
xfer += iprot->skip(ftype);
}
break;
default:
xfer += iprot->skip(ftype);
break;
@ -545,6 +695,36 @@ uint32_t DBOptions::write(apache::thrift::protocol::TProtocol* oprot) const {
xfer += oprot->writeFieldBegin("compression", apache::thrift::protocol::T_I32, 7);
xfer += oprot->writeI32((int32_t)this->compression);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("num_levels", apache::thrift::protocol::T_I32, 8);
xfer += oprot->writeI32(this->num_levels);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("level0_file_num_compaction_trigger", apache::thrift::protocol::T_I32, 9);
xfer += oprot->writeI32(this->level0_file_num_compaction_trigger);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("level0_slowdown_writes_trigger", apache::thrift::protocol::T_I32, 10);
xfer += oprot->writeI32(this->level0_slowdown_writes_trigger);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("level0_stop_writes_trigger", apache::thrift::protocol::T_I32, 11);
xfer += oprot->writeI32(this->level0_stop_writes_trigger);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("target_file_size_base", apache::thrift::protocol::T_I32, 12);
xfer += oprot->writeI32(this->target_file_size_base);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("target_file_size_multiplier", apache::thrift::protocol::T_I32, 13);
xfer += oprot->writeI32(this->target_file_size_multiplier);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("max_bytes_for_level_base", apache::thrift::protocol::T_I32, 14);
xfer += oprot->writeI32(this->max_bytes_for_level_base);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("max_bytes_for_level_multiplier", apache::thrift::protocol::T_I32, 15);
xfer += oprot->writeI32(this->max_bytes_for_level_multiplier);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("max_grandparent_overlap_factor", apache::thrift::protocol::T_I32, 16);
xfer += oprot->writeI32(this->max_grandparent_overlap_factor);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("disableDataSync", apache::thrift::protocol::T_BOOL, 17);
xfer += oprot->writeBool(this->disableDataSync);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldStop();
xfer += oprot->writeStructEnd();
return xfer;
@ -561,6 +741,16 @@ void swap(DBOptions &a, DBOptions &b) {
swap(a.block_size, b.block_size);
swap(a.block_restart_interval, b.block_restart_interval);
swap(a.compression, b.compression);
swap(a.num_levels, b.num_levels);
swap(a.level0_file_num_compaction_trigger, b.level0_file_num_compaction_trigger);
swap(a.level0_slowdown_writes_trigger, b.level0_slowdown_writes_trigger);
swap(a.level0_stop_writes_trigger, b.level0_stop_writes_trigger);
swap(a.target_file_size_base, b.target_file_size_base);
swap(a.target_file_size_multiplier, b.target_file_size_multiplier);
swap(a.max_bytes_for_level_base, b.max_bytes_for_level_base);
swap(a.max_bytes_for_level_multiplier, b.max_bytes_for_level_multiplier);
swap(a.max_grandparent_overlap_factor, b.max_grandparent_overlap_factor);
swap(a.disableDataSync, b.disableDataSync);
swap(a.__isset, b.__isset);
}
@ -579,6 +769,13 @@ void reflectionInitializer_8830325115029814540(::apache::thrift::reflection::Sch
f.name = "sync";
dt.fields[1] = f;
}
{
::apache::thrift::reflection::StructField f;
f.isRequired = true;
f.type = 2U;
f.name = "disableWAL";
dt.fields[2] = f;
}
schema.dataTypes[id] = dt;
schema.names[dt.name] = id;
}
@ -616,6 +813,14 @@ uint32_t WriteOptions::read(apache::thrift::protocol::TProtocol* iprot) {
xfer += iprot->skip(ftype);
}
break;
case 2:
if (ftype == apache::thrift::protocol::T_BOOL) {
xfer += iprot->readBool(this->disableWAL);
this->__isset.disableWAL = true;
} else {
xfer += iprot->skip(ftype);
}
break;
default:
xfer += iprot->skip(ftype);
break;
@ -634,6 +839,9 @@ uint32_t WriteOptions::write(apache::thrift::protocol::TProtocol* oprot) const {
xfer += oprot->writeFieldBegin("sync", apache::thrift::protocol::T_BOOL, 1);
xfer += oprot->writeBool(this->sync);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldBegin("disableWAL", apache::thrift::protocol::T_BOOL, 2);
xfer += oprot->writeBool(this->disableWAL);
xfer += oprot->writeFieldEnd();
xfer += oprot->writeFieldStop();
xfer += oprot->writeStructEnd();
return xfer;
@ -644,6 +852,7 @@ void swap(WriteOptions &a, WriteOptions &b) {
(void)a;
(void)b;
swap(a.sync, b.sync);
swap(a.disableWAL, b.disableWAL);
swap(a.__isset, b.__isset);
}

@ -238,7 +238,7 @@ class DBOptions {
static const uint64_t _reflection_id = 6731746507948871532U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
DBOptions() : create_if_missing(0), error_if_exists(0), write_buffer_size(0), max_open_files(0), block_size(0), block_restart_interval(0), compression(static_cast<CompressionType>(0)) {
DBOptions() : create_if_missing(0), error_if_exists(0), write_buffer_size(0), max_open_files(0), block_size(0), block_restart_interval(0), compression(static_cast<CompressionType>(0)), num_levels(0), level0_file_num_compaction_trigger(0), level0_slowdown_writes_trigger(0), level0_stop_writes_trigger(0), target_file_size_base(0), target_file_size_multiplier(0), max_bytes_for_level_base(0), max_bytes_for_level_multiplier(0), max_grandparent_overlap_factor(0), disableDataSync(0) {
}
DBOptions(const DBOptions&) = default;
@ -254,6 +254,16 @@ class DBOptions {
block_size = 0;
block_restart_interval = 0;
compression = static_cast<CompressionType>(0);
num_levels = 0;
level0_file_num_compaction_trigger = 0;
level0_slowdown_writes_trigger = 0;
level0_stop_writes_trigger = 0;
target_file_size_base = 0;
target_file_size_multiplier = 0;
max_bytes_for_level_base = 0;
max_bytes_for_level_multiplier = 0;
max_grandparent_overlap_factor = 0;
disableDataSync = 0;
__isset.__clear();
}
@ -266,6 +276,16 @@ class DBOptions {
int32_t block_size;
int32_t block_restart_interval;
CompressionType compression;
int32_t num_levels;
int32_t level0_file_num_compaction_trigger;
int32_t level0_slowdown_writes_trigger;
int32_t level0_stop_writes_trigger;
int32_t target_file_size_base;
int32_t target_file_size_multiplier;
int32_t max_bytes_for_level_base;
int32_t max_bytes_for_level_multiplier;
int32_t max_grandparent_overlap_factor;
bool disableDataSync;
struct __isset {
__isset() { __clear(); }
@ -277,6 +297,16 @@ class DBOptions {
block_size = false;
block_restart_interval = false;
compression = false;
num_levels = false;
level0_file_num_compaction_trigger = false;
level0_slowdown_writes_trigger = false;
level0_stop_writes_trigger = false;
target_file_size_base = false;
target_file_size_multiplier = false;
max_bytes_for_level_base = false;
max_bytes_for_level_multiplier = false;
max_grandparent_overlap_factor = false;
disableDataSync = false;
}
bool create_if_missing;
bool error_if_exists;
@ -285,6 +315,16 @@ class DBOptions {
bool block_size;
bool block_restart_interval;
bool compression;
bool num_levels;
bool level0_file_num_compaction_trigger;
bool level0_slowdown_writes_trigger;
bool level0_stop_writes_trigger;
bool target_file_size_base;
bool target_file_size_multiplier;
bool max_bytes_for_level_base;
bool max_bytes_for_level_multiplier;
bool max_grandparent_overlap_factor;
bool disableDataSync;
} __isset;
bool operator == (const DBOptions & rhs) const
@ -303,6 +343,26 @@ class DBOptions {
return false;
if (!(this->compression == rhs.compression))
return false;
if (!(this->num_levels == rhs.num_levels))
return false;
if (!(this->level0_file_num_compaction_trigger == rhs.level0_file_num_compaction_trigger))
return false;
if (!(this->level0_slowdown_writes_trigger == rhs.level0_slowdown_writes_trigger))
return false;
if (!(this->level0_stop_writes_trigger == rhs.level0_stop_writes_trigger))
return false;
if (!(this->target_file_size_base == rhs.target_file_size_base))
return false;
if (!(this->target_file_size_multiplier == rhs.target_file_size_multiplier))
return false;
if (!(this->max_bytes_for_level_base == rhs.max_bytes_for_level_base))
return false;
if (!(this->max_bytes_for_level_multiplier == rhs.max_bytes_for_level_multiplier))
return false;
if (!(this->max_grandparent_overlap_factor == rhs.max_grandparent_overlap_factor))
return false;
if (!(this->disableDataSync == rhs.disableDataSync))
return false;
return true;
}
bool operator != (const DBOptions &rhs) const {
@ -324,7 +384,7 @@ class WriteOptions {
static const uint64_t _reflection_id = 8830325115029814540U;
static void _reflection_register(::apache::thrift::reflection::Schema&);
WriteOptions() : sync(0) {
WriteOptions() : sync(0), disableWAL(0) {
}
WriteOptions(const WriteOptions&) = default;
@ -334,25 +394,31 @@ class WriteOptions {
void __clear() {
sync = 0;
disableWAL = 0;
__isset.__clear();
}
virtual ~WriteOptions() throw() {}
bool sync;
bool disableWAL;
struct __isset {
__isset() { __clear(); }
void __clear() {
sync = false;
disableWAL = false;
}
bool sync;
bool disableWAL;
} __isset;
bool operator == (const WriteOptions & rhs) const
{
if (!(this->sync == rhs.sync))
return false;
if (!(this->disableWAL == rhs.disableWAL))
return false;
return true;
}
bool operator != (const WriteOptions &rhs) const {

@ -47,12 +47,23 @@ struct DBOptions {
4:i32 max_open_files;
5:i32 block_size;
6:i32 block_restart_interval;
7:CompressionType compression
7:CompressionType compression,
8:i32 num_levels,
9:i32 level0_file_num_compaction_trigger,
10:i32 level0_slowdown_writes_trigger,
11:i32 level0_stop_writes_trigger,
12:i32 target_file_size_base,
13:i32 target_file_size_multiplier,
14:i32 max_bytes_for_level_base,
15:i32 max_bytes_for_level_multiplier,
16:i32 max_grandparent_overlap_factor,
17:bool disableDataSync
}
// Options for writing
struct WriteOptions {
1:bool sync
1:bool sync,
2:bool disableWAL
}
struct Snapshot {

@ -73,10 +73,10 @@ public:
cache_size_ = l;
} else if (sscanf(argv[i], "--cache_numshardbits=%d%c", &n, &junk) == 1) {
cache_numshardbits_ = n;
} else if (strncmp(argv[i], "--hostname=", 10) == 0) {
hostname_ = argv[i] + 10;
} else if (strncmp(argv[i], "--rootdir=", 9) == 0) {
rootdir_ = argv[i] + 9;
} else if (strncmp(argv[i], "--hostname=", 11) == 0) {
hostname_ = argv[i] + 11;
} else if (strncmp(argv[i], "--rootdir=", 10) == 0) {
rootdir_ = argv[i] + 10;
} else {
fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
return false;

@ -75,6 +75,26 @@ class DBHandler : virtual public DBIf {
} else if (dboptions.compression == kSnappyCompression) {
options.compression = leveldb::kSnappyCompression;
}
if (dboptions.num_levels > 0)
options.num_levels = dboptions.num_levels;
if (dboptions.level0_file_num_compaction_trigger > 0)
options.level0_file_num_compaction_trigger = dboptions.level0_file_num_compaction_trigger;
if (dboptions.level0_slowdown_writes_trigger > 0)
options.level0_slowdown_writes_trigger = dboptions.level0_slowdown_writes_trigger;
if (dboptions.level0_stop_writes_trigger)
options.level0_stop_writes_trigger = dboptions.level0_stop_writes_trigger;
if (dboptions.target_file_size_base > 0)
options.target_file_size_base = dboptions.target_file_size_base;
if (dboptions.target_file_size_multiplier > 0)
options.target_file_size_multiplier = dboptions.target_file_size_multiplier;
if (dboptions.max_bytes_for_level_base)
options.max_bytes_for_level_base = dboptions.max_bytes_for_level_base;
if (dboptions.max_bytes_for_level_multiplier)
options.max_bytes_for_level_multiplier = dboptions.max_bytes_for_level_multiplier;
if (dboptions.max_grandparent_overlap_factor)
options.max_grandparent_overlap_factor = dboptions.max_grandparent_overlap_factor;
if (dboptions.disableDataSync)
options.disableDataSync = dboptions.disableDataSync;
openHandles->add(options, dbname, dbdir);
_return.dbname = dbname;
}
@ -91,6 +111,7 @@ class DBHandler : virtual public DBIf {
const WriteOptions& options) {
leveldb::WriteOptions woptions;
woptions.sync = options.sync;
woptions.disableWAL = options.disableWAL;
leveldb::Slice key, value;
key.data_ = kv.key.data.data();
key.size_ = kv.key.size;
@ -111,6 +132,7 @@ class DBHandler : virtual public DBIf {
const WriteOptions& options) {
leveldb::WriteOptions woptions;
woptions.sync = options.sync;
woptions.disableWAL = options.disableWAL;
leveldb::Slice key;
key.data_ = kv.data.data();
key.size_ = kv.size;
@ -130,6 +152,7 @@ class DBHandler : virtual public DBIf {
leveldb::WriteOptions woptions;
leveldb::WriteBatch lbatch;
woptions.sync = options.sync;
woptions.disableWAL = options.disableWAL;
leveldb::Slice key, value;
for (unsigned int i = 0; i < batch.size(); i++) {
kv one = batch[i];

@ -0,0 +1,71 @@
// 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 "db/version_set.h"
#include <algorithm>
#include <stdio.h>
#include "db/filename.h"
#include "db/log_reader.h"
#include "db/log_writer.h"
#include "db/memtable.h"
#include "db/table_cache.h"
#include "leveldb/env.h"
#include "leveldb/table_builder.h"
#include "table/merger.h"
#include "table/two_level_iterator.h"
#include "util/coding.h"
#include "util/logging.h"
static int verbose = 0;
using namespace leveldb;
//
// Takes a manifest file and dumps out all metedata
//
int main(int argc, char** argv) {
// parse command line options
int n;
char junk;
int foundfile = 0;
std::string manifestfile;
for (int i = 1; i < argc; i++) {
std::string param(argv[i]);
if ((n = param.find("--file=")) != std::string::npos) {
manifestfile = param.substr(strlen("--file="));
foundfile = 1;
} else if (sscanf(argv[i], "--verbose=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
verbose = n;
}
}
if (!foundfile) {
fprintf(stderr, "%s [--verbose=0|1] [--file=pathname of manifest file\n",
argv[0]);
abort();
}
if (verbose) {
printf("Processing Manifest file %s\n", manifestfile.c_str());
}
Options options;
std::string file(manifestfile);
std::string dbname("dummy");
TableCache* tc = new TableCache(dbname, &options, 10);
const InternalKeyComparator* cmp = new InternalKeyComparator(options.comparator);
VersionSet* versions = new VersionSet(dbname, &options,
tc, cmp);
Status s = versions->DumpManifest(options, file);
if (!s.ok()) {
printf("Error in processing file %s %s\n", manifestfile.c_str(),
s.ToString().c_str());
}
if (verbose) {
printf("Processing Manifest file %s done\n", manifestfile.c_str());
}
}

@ -0,0 +1,14 @@
/*version.h*/
#ifndef VERSION_H_
#define VERSION_H_
// these variables tell us about the git config and time
extern const char* leveldb_build_git_sha;
extern const char* leveldb_build_git_datetime;
// these variables tell us when the compilation occured
extern const char* leveldb_build_compile_time;
extern const char* leveldb_build_compile_date;
#endif /* VERSION_H_ */

@ -3,6 +3,7 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include <deque>
#include <set>
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
@ -31,6 +32,10 @@ namespace leveldb {
namespace {
// list of pathnames that are locked
static std::set<std::string> lockedFiles;
static port::Mutex mutex_lockedFiles;
static Status IOError(const std::string& context, int err_number) {
return Status::IOError(context, strerror(err_number));
}
@ -291,7 +296,28 @@ class PosixMmapFile : public WritableFile {
}
};
static int LockOrUnlock(int fd, bool lock) {
static int LockOrUnlock(const std::string& fname, int fd, bool lock) {
mutex_lockedFiles.Lock();
if (lock) {
// If it already exists in the lockedFiles set, then it is already locked,
// and fail this lock attempt. Otherwise, insert it into lockedFiles.
// This check is needed because fcntl() does not detect lock conflict
// if the fcntl is issued by the same thread that earlier acquired
// this lock.
if (lockedFiles.insert(fname).second == false) {
mutex_lockedFiles.Unlock();
errno = ENOLCK;
return -1;
}
} else {
// If we are unlocking, then verify that we had locked it earlier,
// it should already exist in lockedFiles. Remove it from lockedFiles.
if (lockedFiles.erase(fname) != 1) {
mutex_lockedFiles.Unlock();
errno = ENOLCK;
return -1;
}
}
errno = 0;
struct flock f;
memset(&f, 0, sizeof(f));
@ -299,12 +325,19 @@ static int LockOrUnlock(int fd, bool lock) {
f.l_whence = SEEK_SET;
f.l_start = 0;
f.l_len = 0; // Lock/unlock entire file
return fcntl(fd, F_SETLK, &f);
int value = fcntl(fd, F_SETLK, &f);
if (value == -1 && lock) {
// if there is an error in locking, then remove the pathname from lockedfiles
lockedFiles.erase(fname);
}
mutex_lockedFiles.Unlock();
return value;
}
class PosixFileLock : public FileLock {
public:
int fd_;
std::string filename;
};
class PosixEnv : public Env {
@ -435,12 +468,13 @@ class PosixEnv : public Env {
int fd = open(fname.c_str(), O_RDWR | O_CREAT, 0644);
if (fd < 0) {
result = IOError(fname, errno);
} else if (LockOrUnlock(fd, true) == -1) {
} else if (LockOrUnlock(fname, fd, true) == -1) {
result = IOError("lock " + fname, errno);
close(fd);
} else {
PosixFileLock* my_lock = new PosixFileLock;
my_lock->fd_ = fd;
my_lock->filename = fname;
*lock = my_lock;
}
return result;
@ -449,7 +483,7 @@ class PosixEnv : public Env {
virtual Status UnlockFile(FileLock* lock) {
PosixFileLock* my_lock = reinterpret_cast<PosixFileLock*>(lock);
Status result;
if (LockOrUnlock(my_lock->fd_, false) == -1) {
if (LockOrUnlock(my_lock->filename, my_lock->fd_, false) == -1) {
result = IOError("unlock", errno);
}
close(my_lock->fd_);
@ -503,6 +537,43 @@ class PosixEnv : public Env {
usleep(micros);
}
virtual Status GetHostName(char* name, uint len) {
int ret = gethostname(name, len);
if (ret < 0) {
if (errno == EFAULT || errno == EINVAL)
return Status::InvalidArgument(strerror(errno));
else
return IOError("GetHostName", errno);
}
return Status::OK();
}
virtual Status GetCurrentTime(int64_t* unix_time) {
time_t ret = time(NULL);
if (ret == (time_t) -1) {
return IOError("GetCurrentTime", errno);
}
*unix_time = (int64_t) ret;
return Status::OK();
}
virtual Status GetAbsolutePath(const std::string& db_path,
std::string* output_path) {
if (db_path.find('/') == 0) {
*output_path = db_path;
return Status::OK();
}
char the_path[256];
char* ret = getcwd(the_path, 256);
if (ret == NULL) {
return Status::IOError(strerror(errno));
}
*output_path = ret;
return Status::OK();
}
private:
void PthreadCall(const char* label, int result) {
if (result != 0) {

@ -0,0 +1,57 @@
// Copyright (c) 2012 Facebook. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "leveldb/status.h"
#include "leveldb/env.h"
#include <vector>
#include "util/coding.h"
#include "util/testharness.h"
namespace leveldb {
class LockTest {
public:
static LockTest* current_;
std::string file_;
leveldb::Env* env_;
LockTest() : file_(test::TmpDir() + "/db_testlock_file"),
env_(leveldb::Env::Default()) {
current_ = this;
}
~LockTest() {
}
Status LockFile(FileLock** db_lock) {
return env_->LockFile(file_, db_lock);
}
Status UnlockFile(FileLock* db_lock) {
return env_->UnlockFile(db_lock);
}
};
LockTest* LockTest::current_;
TEST(LockTest, LockBySameThread) {
FileLock* lock1;
FileLock* lock2;
// acquire a lock on a file
ASSERT_OK(LockFile(&lock1));
// re-acquire the lock on the same file. This should fail.
ASSERT_TRUE(LockFile(&lock2).IsIOError());
// release the lock
ASSERT_OK(UnlockFile(lock1));
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

@ -6,6 +6,7 @@
#include "leveldb/comparator.h"
#include "leveldb/env.h"
#include "leveldb/filter_policy.h"
namespace leveldb {
@ -28,14 +29,36 @@ Options::Options()
level0_stop_writes_trigger(12),
max_mem_compaction_level(2),
target_file_size_base(2 * 1048576),
target_file_size_multiplier(10),
target_file_size_multiplier(1),
max_bytes_for_level_base(10 * 1048576),
max_bytes_for_level_multiplier(10),
expanded_compaction_factor(25),
max_grandparent_overlap_factor(10),
filter_policy(NULL),
statistics(NULL),
disableDataSync(false) {
disableDataSync(false),
db_stats_log_interval(1800) {
}
void
Options::Dump(
Logger * log) const
{
Log(log," Options.comparator: %s", comparator->Name());
Log(log," Options.create_if_missing: %d", create_if_missing);
Log(log," Options.error_if_exists: %d", error_if_exists);
Log(log," Options.paranoid_checks: %d", paranoid_checks);
Log(log," Options.env: %p", env);
Log(log," Options.info_log: %p", info_log);
Log(log," Options.write_buffer_size: %zd", write_buffer_size);
Log(log," Options.max_open_files: %d", max_open_files);
Log(log," Options.block_cache: %p", block_cache);
Log(log," Options.block_size: %zd", block_size);
Log(log,"Options.block_restart_interval: %d", block_restart_interval);
Log(log," Options.compression: %d", compression);
Log(log," Options.filter_policy: %s", filter_policy == NULL ? "NULL" : filter_policy->Name());
} // Options::Dump
} // namespace leveldb

@ -0,0 +1,23 @@
#ifndef STATS_LOGGER_H_
#define STATS_LOGGER_H_
namespace leveldb {
class StatsLogger {
public:
virtual void Log_Deploy_Stats(const std::string& db_version,
const std::string& machine_info,
const std::string& data_dir,
const uint64_t data_size,
const uint32_t file_number,
const std::string& data_size_per_level,
const std::string& file_number_per_level,
const int64_t& ts_unix) = 0;
};
}
#endif
Loading…
Cancel
Save