mirror of
https://github.com/facebook/rocksdb.git
synced 2024-11-30 22:41:48 +00:00
d79be3dca2
Summary: ## Option API updates * Add new CompressionOptions::max_compressed_bytes_per_kb, which corresponds to 1024.0 / min allowable compression ratio. This avoids the hard-coded minimum ratio of 8/7. * Remove unnecessary constructor for CompressionOptions. * Document undocumented CompressionOptions. Use idiom for default values shown clearly in one place (not precariously repeated). ## Stat API updates * Deprecate the BYTES_COMPRESSED, BYTES_DECOMPRESSED histograms. Histograms incur substantial extra space & time costs compared to tickers, and the distribution of uncompressed data block sizes tends to be uninteresting. If we're interested in that distribution, I don't see why it should be limited to blocks stored as compressed. * Deprecate the NUMBER_BLOCK_NOT_COMPRESSED ticker, because the name is very confusing. * New or existing tickers relevant to compression: * BYTES_COMPRESSED_FROM * BYTES_COMPRESSED_TO * BYTES_COMPRESSION_BYPASSED * BYTES_COMPRESSION_REJECTED * COMPACT_WRITE_BYTES + FLUSH_WRITE_BYTES (both existing) * NUMBER_BLOCK_COMPRESSED (existing) * NUMBER_BLOCK_COMPRESSION_BYPASSED * NUMBER_BLOCK_COMPRESSION_REJECTED * BYTES_DECOMPRESSED_FROM * BYTES_DECOMPRESSED_TO We can compute a number of things with these stats: * "Successful" compression ratio: BYTES_COMPRESSED_FROM / BYTES_COMPRESSED_TO * Compression ratio of data on which compression was attempted: (BYTES_COMPRESSED_FROM + BYTES_COMPRESSION_REJECTED) / (BYTES_COMPRESSED_TO + BYTES_COMPRESSION_REJECTED) * Compression ratio of data that could be eligible for compression: (BYTES_COMPRESSED_FROM + X) / (BYTES_COMPRESSED_TO + X) where X = BYTES_COMPRESSION_REJECTED + NUMBER_BLOCK_COMPRESSION_REJECTED * Overall SST compression ratio (compression disabled vs. actual): (Y - BYTES_COMPRESSED_TO + BYTES_COMPRESSED_FROM) / Y where Y = COMPACT_WRITE_BYTES + FLUSH_WRITE_BYTES Keeping _REJECTED separate from _BYPASSED helps us to understand "wasted" CPU time in compression. ## BlockBasedTableBuilder Various small refactorings, optimizations, and name clean-ups. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11388 Test Plan: unit tests added * `options_settable_test.cc`: use non-deprecated idiom for configuring CompressionOptions from string. The old idiom is tested elsewhere and does not need to be updated to support the new field. Reviewed By: ajkr Differential Revision: D45128202 Pulled By: pdillinger fbshipit-source-id: 5a652bf5c022b7ec340cf79018cccf0686962803
137 lines
4.4 KiB
C++
137 lines
4.4 KiB
C++
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
|
// This source code is licensed under both the GPLv2 (found in the
|
|
// COPYING file in the root directory) and Apache 2.0 License
|
|
// (found in the LICENSE.Apache file in the root directory).
|
|
//
|
|
#pragma once
|
|
#include "monitoring/statistics.h"
|
|
#include "rocksdb/system_clock.h"
|
|
|
|
namespace ROCKSDB_NAMESPACE {
|
|
// Auto-scoped.
|
|
// When statistics is not nullptr, records the measured time into any enabled
|
|
// histograms supplied to the constructor. A histogram argument may be omitted
|
|
// by setting it to Histograms::HISTOGRAM_ENUM_MAX. It is also saved into
|
|
// *elapsed if the pointer is not nullptr and overwrite is true, it will be
|
|
// added to *elapsed if overwrite is false.
|
|
class StopWatch {
|
|
public:
|
|
StopWatch(SystemClock* clock, Statistics* statistics,
|
|
const uint32_t hist_type_1,
|
|
const uint32_t hist_type_2 = Histograms::HISTOGRAM_ENUM_MAX,
|
|
uint64_t* elapsed = nullptr, bool overwrite = true,
|
|
bool delay_enabled = false)
|
|
: clock_(clock),
|
|
statistics_(statistics),
|
|
hist_type_1_(statistics && statistics->HistEnabledForType(hist_type_1)
|
|
? hist_type_1
|
|
: Histograms::HISTOGRAM_ENUM_MAX),
|
|
hist_type_2_(statistics && statistics->HistEnabledForType(hist_type_2)
|
|
? hist_type_2
|
|
: Histograms::HISTOGRAM_ENUM_MAX),
|
|
elapsed_(elapsed),
|
|
overwrite_(overwrite),
|
|
stats_enabled_(statistics &&
|
|
statistics->get_stats_level() >=
|
|
StatsLevel::kExceptTimers &&
|
|
(hist_type_1_ != Histograms::HISTOGRAM_ENUM_MAX ||
|
|
hist_type_2_ != Histograms::HISTOGRAM_ENUM_MAX)),
|
|
delay_enabled_(delay_enabled),
|
|
total_delay_(0),
|
|
delay_start_time_(0),
|
|
start_time_((stats_enabled_ || elapsed != nullptr) ? clock->NowMicros()
|
|
: 0) {}
|
|
|
|
~StopWatch() {
|
|
if (elapsed_) {
|
|
if (overwrite_) {
|
|
*elapsed_ = clock_->NowMicros() - start_time_;
|
|
} else {
|
|
*elapsed_ += clock_->NowMicros() - start_time_;
|
|
}
|
|
}
|
|
if (elapsed_ && delay_enabled_) {
|
|
*elapsed_ -= total_delay_;
|
|
}
|
|
if (stats_enabled_) {
|
|
const auto time = (elapsed_ != nullptr)
|
|
? *elapsed_
|
|
: (clock_->NowMicros() - start_time_);
|
|
if (hist_type_1_ != Histograms::HISTOGRAM_ENUM_MAX) {
|
|
statistics_->reportTimeToHistogram(hist_type_1_, time);
|
|
}
|
|
if (hist_type_2_ != Histograms::HISTOGRAM_ENUM_MAX) {
|
|
statistics_->reportTimeToHistogram(hist_type_2_, time);
|
|
}
|
|
}
|
|
}
|
|
|
|
void DelayStart() {
|
|
// if delay_start_time_ is not 0, it means we are already tracking delay,
|
|
// so delay_start_time_ should not be overwritten
|
|
if (elapsed_ && delay_enabled_ && delay_start_time_ == 0) {
|
|
delay_start_time_ = clock_->NowMicros();
|
|
}
|
|
}
|
|
|
|
void DelayStop() {
|
|
if (elapsed_ && delay_enabled_ && delay_start_time_ != 0) {
|
|
total_delay_ += clock_->NowMicros() - delay_start_time_;
|
|
}
|
|
// reset to 0 means currently no delay is being tracked, so two consecutive
|
|
// calls to DelayStop will not increase total_delay_
|
|
delay_start_time_ = 0;
|
|
}
|
|
|
|
uint64_t GetDelay() const { return delay_enabled_ ? total_delay_ : 0; }
|
|
|
|
uint64_t start_time() const { return start_time_; }
|
|
|
|
private:
|
|
SystemClock* clock_;
|
|
Statistics* statistics_;
|
|
const uint32_t hist_type_1_;
|
|
const uint32_t hist_type_2_;
|
|
uint64_t* elapsed_;
|
|
bool overwrite_;
|
|
bool stats_enabled_;
|
|
bool delay_enabled_;
|
|
uint64_t total_delay_;
|
|
uint64_t delay_start_time_;
|
|
const uint64_t start_time_;
|
|
};
|
|
|
|
// a nano second precision stopwatch
|
|
class StopWatchNano {
|
|
public:
|
|
explicit StopWatchNano(SystemClock* clock, bool auto_start = false)
|
|
: clock_(clock), start_(0) {
|
|
if (auto_start) {
|
|
Start();
|
|
}
|
|
}
|
|
|
|
void Start() { start_ = clock_->NowNanos(); }
|
|
|
|
uint64_t ElapsedNanos(bool reset = false) {
|
|
auto now = clock_->NowNanos();
|
|
auto elapsed = now - start_;
|
|
if (reset) {
|
|
start_ = now;
|
|
}
|
|
return elapsed;
|
|
}
|
|
|
|
uint64_t ElapsedNanosSafe(bool reset = false) {
|
|
return (clock_ != nullptr) ? ElapsedNanos(reset) : 0U;
|
|
}
|
|
|
|
bool IsStarted() { return start_ != 0; }
|
|
|
|
private:
|
|
SystemClock* clock_;
|
|
uint64_t start_;
|
|
};
|
|
|
|
} // namespace ROCKSDB_NAMESPACE
|