rocksdb/file/file_prefetch_buffer.h
Hui Xiao 3093d98c78 Fix higher read qps during db open caused by pr 11406 (#11516)
Summary:
**Context:**
[PR11406](https://github.com/facebook/rocksdb/pull/11406/) caused more frequent read during db open reading files with no `tail_size` in the manifest as part of the upgrade to 11406. This is due to that PR introduced
- [smaller](https://github.com/facebook/rocksdb/pull/11406/files#diff-57ed8c49db2bdd4db7618646a177397674bbf25beacacecb104070071d30129fR833) prefetch tail buffer size compared to pre-11406 for small files (< 52 MB) when `tail_prefetch_stats` infers tail size to be 0 (usually happens when the stats does not have much historical data to infer early on)
-  more read (up to # of partitioned filter/index) when such small prefetch tail buffer does not contain all the partitioned filter/index needed in CacheDependencies() since the [fallback logic](https://github.com/facebook/rocksdb/pull/11406/files#diff-d98f1a83de24412ad7f3527725dae7e28851c7222622c3cdb832d3cdf24bbf9fR165-R179) that prefetches all partitions at once will be [skipped](url) when such a small prefetch tail buffer is passed in

**Summary:**
- Revert the fallback prefetch buffer size change to preserve existing behavior fully during upgrading in `BlockBasedTable::PrefetchTail()`
- Use passed-in prefetch tail buffer in `CacheDependencies()` only if it has a smaller offset than the the offset of first partition filter/index, that is, at least as good as the existing prefetching behavior

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11516

Test Plan:
- db bench

Create db with small files prior to PR 11406
```
./db_bench -db=/tmp/testdb/ --partition_index_and_filters=1 --statistics=1 -benchmarks=fillseq -key_size=3200 -value_size=5 -num=1000000 -write_buffer_size=6550000 -disable_auto_compactions=true -compression_type=zstd`
```
Read db to see if post-pr has lower read qps (i.e, rocksdb.file.read.db.open.micros count) during db open.
```
./db_bench -use_direct_reads=1 --file_opening_threads=1 --threads=1 --use_existing_db=1 --seed=1682546046158958 --partition_index_and_filters=1 --statistics=1 --db=/tmp/testdb/ --benchmarks=readrandom --key_size=3200 --value_size=5 --num=100 --disable_auto_compactions=true --compression_type=zstd
```
Pre-PR:
```
rocksdb.file.read.db.open.micros P50 : 3.399023 P95 : 5.924468 P99 : 12.408333 P100 : 29.000000 COUNT : 611 SUM : 2539
```

Post-PR:
```
rocksdb.file.read.db.open.micros P50 : 593.736842 P95 : 861.605263 P99 : 1212.868421 P100 : 2663.000000 COUNT : 585 SUM : 345349
```

_Note: To control the starting offset of the prefetch tail buffer easier, I manually override the following to eliminate the effect of alignment_
```
class PosixRandomAccessFile : public FSRandomAccessFile {
virtual size_t GetRequiredBufferAlignment() const override {
-    return logical_sector_size_;
+    return 1;
  }
 ```

- CI

Reviewed By: pdillinger

Differential Revision: D46472566

Pulled By: hx235

fbshipit-source-id: 2fe14ac8d489d15b0e08e6f8fe4f46d5f110978e
2023-06-06 17:42:43 -07:00

472 lines
19 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).
//
// 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.
#pragma once
#include <algorithm>
#include <atomic>
#include <sstream>
#include <string>
#include "file/readahead_file_info.h"
#include "monitoring/statistics_impl.h"
#include "port/port.h"
#include "rocksdb/env.h"
#include "rocksdb/file_system.h"
#include "rocksdb/options.h"
#include "util/aligned_buffer.h"
#include "util/autovector.h"
#include "util/stop_watch.h"
namespace ROCKSDB_NAMESPACE {
#define DEFAULT_DECREMENT 8 * 1024
struct IOOptions;
class RandomAccessFileReader;
struct BufferInfo {
AlignedBuffer buffer_;
uint64_t offset_ = 0;
// Below parameters are used in case of async read flow.
// Length requested for in ReadAsync.
size_t async_req_len_ = 0;
// async_read_in_progress can be used as mutex. Callback can update the buffer
// and its size but async_read_in_progress is only set by main thread.
bool async_read_in_progress_ = false;
// io_handle is allocated and used by underlying file system in case of
// asynchronous reads.
void* io_handle_ = nullptr;
IOHandleDeleter del_fn_ = nullptr;
// pos represents the index of this buffer in vector of BufferInfo.
uint32_t pos_ = 0;
};
enum class FilePrefetchBufferUsage {
kTableOpenPrefetchTail,
kUnknown,
};
// FilePrefetchBuffer is a smart buffer to store and read data from a file.
class FilePrefetchBuffer {
public:
// Constructor.
//
// All arguments are optional.
// readahead_size : the initial readahead size.
// max_readahead_size : the maximum readahead size.
// If max_readahead_size > readahead_size, the readahead size will be
// doubled on every IO until max_readahead_size is hit.
// Typically this is set as a multiple of readahead_size.
// max_readahead_size should be greater than equal to readahead_size.
// enable : controls whether reading from the buffer is enabled.
// If false, TryReadFromCache() always return false, and we only take stats
// for the minimum offset if track_min_offset = true.
// track_min_offset : Track the minimum offset ever read and collect stats on
// it. Used for adaptable readahead of the file footer/metadata.
// implicit_auto_readahead : Readahead is enabled implicitly by rocksdb after
// doing sequential scans for two times.
//
// Automatic readhead is enabled for a file if readahead_size
// and max_readahead_size are passed in.
// A user can construct a FilePrefetchBuffer without any arguments, but use
// `Prefetch` to load data into the buffer.
FilePrefetchBuffer(
size_t readahead_size = 0, size_t max_readahead_size = 0,
bool enable = true, bool track_min_offset = false,
bool implicit_auto_readahead = false, uint64_t num_file_reads = 0,
uint64_t num_file_reads_for_auto_readahead = 0, FileSystem* fs = nullptr,
SystemClock* clock = nullptr, Statistics* stats = nullptr,
FilePrefetchBufferUsage usage = FilePrefetchBufferUsage::kUnknown)
: curr_(0),
readahead_size_(readahead_size),
initial_auto_readahead_size_(readahead_size),
max_readahead_size_(max_readahead_size),
min_offset_read_(std::numeric_limits<size_t>::max()),
enable_(enable),
track_min_offset_(track_min_offset),
implicit_auto_readahead_(implicit_auto_readahead),
prev_offset_(0),
prev_len_(0),
num_file_reads_for_auto_readahead_(num_file_reads_for_auto_readahead),
num_file_reads_(num_file_reads),
explicit_prefetch_submitted_(false),
fs_(fs),
clock_(clock),
stats_(stats),
usage_(usage) {
assert((num_file_reads_ >= num_file_reads_for_auto_readahead_ + 1) ||
(num_file_reads_ == 0));
// If ReadOptions.async_io is enabled, data is asynchronously filled in
// second buffer while curr_ is being consumed. If data is overlapping in
// two buffers, data is copied to third buffer to return continuous buffer.
bufs_.resize(3);
for (uint32_t i = 0; i < 2; i++) {
bufs_[i].pos_ = i;
}
}
~FilePrefetchBuffer() {
// Abort any pending async read request before destroying the class object.
if (fs_ != nullptr) {
std::vector<void*> handles;
for (uint32_t i = 0; i < 2; i++) {
if (bufs_[i].async_read_in_progress_ &&
bufs_[i].io_handle_ != nullptr) {
handles.emplace_back(bufs_[i].io_handle_);
}
}
if (!handles.empty()) {
StopWatch sw(clock_, stats_, ASYNC_PREFETCH_ABORT_MICROS);
Status s = fs_->AbortIO(handles);
assert(s.ok());
}
}
// Prefetch buffer bytes discarded.
uint64_t bytes_discarded = 0;
// Iterated over 2 buffers.
for (int i = 0; i < 2; i++) {
int first = i;
int second = i ^ 1;
if (DoesBufferContainData(first)) {
// If last block was read completely from first and some bytes in
// first buffer are still unconsumed.
if (prev_offset_ >= bufs_[first].offset_ &&
prev_offset_ + prev_len_ <
bufs_[first].offset_ + bufs_[first].buffer_.CurrentSize()) {
bytes_discarded += bufs_[first].buffer_.CurrentSize() -
(prev_offset_ + prev_len_ - bufs_[first].offset_);
}
// If data was in second buffer and some/whole block bytes were read
// from second buffer.
else if (prev_offset_ < bufs_[first].offset_ &&
!DoesBufferContainData(second)) {
// If last block read was completely from different buffer, this
// buffer is unconsumed.
if (prev_offset_ + prev_len_ <= bufs_[first].offset_) {
bytes_discarded += bufs_[first].buffer_.CurrentSize();
}
// If last block read overlaps with this buffer and some data is
// still unconsumed and previous buffer (second) is not cleared.
else if (prev_offset_ + prev_len_ > bufs_[first].offset_ &&
bufs_[first].offset_ + bufs_[first].buffer_.CurrentSize() ==
bufs_[second].offset_) {
bytes_discarded += bufs_[first].buffer_.CurrentSize() -
(/*bytes read from this buffer=*/prev_len_ -
(bufs_[first].offset_ - prev_offset_));
}
}
}
}
for (uint32_t i = 0; i < 2; i++) {
// Release io_handle.
DestroyAndClearIOHandle(i);
}
RecordInHistogram(stats_, PREFETCHED_BYTES_DISCARDED, bytes_discarded);
}
bool Enabled() const { return enable_; }
// Load data into the buffer from a file.
// reader : the file reader.
// offset : the file offset to start reading from.
// n : the number of bytes to read.
// rate_limiter_priority : rate limiting priority, or `Env::IO_TOTAL` to
// bypass.
Status Prefetch(const IOOptions& opts, RandomAccessFileReader* reader,
uint64_t offset, size_t n,
Env::IOPriority rate_limiter_priority);
// Request for reading the data from a file asynchronously.
// If data already exists in the buffer, result will be updated.
// reader : the file reader.
// offset : the file offset to start reading from.
// n : the number of bytes to read.
// result : if data already exists in the buffer, result will
// be updated with the data.
//
// If data already exist in the buffer, it will return Status::OK, otherwise
// it will send asynchronous request and return Status::TryAgain.
Status PrefetchAsync(const IOOptions& opts, RandomAccessFileReader* reader,
uint64_t offset, size_t n, Slice* result);
// Tries returning the data for a file read from this buffer if that data is
// in the buffer.
// It handles tracking the minimum read offset if track_min_offset = true.
// It also does the exponential readahead when readahead_size is set as part
// of the constructor.
//
// opts : the IO options to use.
// reader : the file reader.
// offset : the file offset.
// n : the number of bytes.
// result : output buffer to put the data into.
// s : output status.
// rate_limiter_priority : rate limiting priority, or `Env::IO_TOTAL` to
// bypass.
// for_compaction : true if cache read is done for compaction read.
bool TryReadFromCache(const IOOptions& opts, RandomAccessFileReader* reader,
uint64_t offset, size_t n, Slice* result, Status* s,
Env::IOPriority rate_limiter_priority,
bool for_compaction = false);
bool TryReadFromCacheAsync(const IOOptions& opts,
RandomAccessFileReader* reader, uint64_t offset,
size_t n, Slice* result, Status* status,
Env::IOPriority rate_limiter_priority);
// The minimum `offset` ever passed to TryReadFromCache(). This will nly be
// tracked if track_min_offset = true.
size_t min_offset_read() const { return min_offset_read_; }
size_t GetPrefetchOffset() const { return bufs_[curr_].offset_; }
// Called in case of implicit auto prefetching.
void UpdateReadPattern(const uint64_t& offset, const size_t& len,
bool decrease_readaheadsize) {
if (decrease_readaheadsize) {
// Since this block was eligible for prefetch but it was found in
// cache, so check and decrease the readahead_size by 8KB (default)
// if eligible.
DecreaseReadAheadIfEligible(offset, len);
}
prev_offset_ = offset;
prev_len_ = len;
explicit_prefetch_submitted_ = false;
}
void GetReadaheadState(ReadaheadFileInfo::ReadaheadInfo* readahead_info) {
readahead_info->readahead_size = readahead_size_;
readahead_info->num_file_reads = num_file_reads_;
}
void DecreaseReadAheadIfEligible(uint64_t offset, size_t size,
size_t value = DEFAULT_DECREMENT) {
// Decrease the readahead_size if
// - its enabled internally by RocksDB (implicit_auto_readahead_) and,
// - readahead_size is greater than 0 and,
// - this block would have called prefetch API if not found in cache for
// which conditions are:
// - few/no bytes are in buffer and,
// - block is sequential with the previous read and,
// - num_file_reads_ + 1 (including this read) >
// num_file_reads_for_auto_readahead_
size_t curr_size = bufs_[curr_].async_read_in_progress_
? bufs_[curr_].async_req_len_
: bufs_[curr_].buffer_.CurrentSize();
if (implicit_auto_readahead_ && readahead_size_ > 0) {
if ((offset + size > bufs_[curr_].offset_ + curr_size) &&
IsBlockSequential(offset) &&
(num_file_reads_ + 1 > num_file_reads_for_auto_readahead_)) {
readahead_size_ =
std::max(initial_auto_readahead_size_,
(readahead_size_ >= value ? readahead_size_ - value : 0));
}
}
}
// Callback function passed to underlying FS in case of asynchronous reads.
void PrefetchAsyncCallback(const FSReadRequest& req, void* cb_arg);
private:
// Calculates roundoff offset and length to be prefetched based on alignment
// and data present in buffer_. It also allocates new buffer or refit tail if
// required.
void CalculateOffsetAndLen(size_t alignment, uint64_t offset,
size_t roundup_len, uint32_t index,
bool refit_tail, uint64_t& chunk_len);
void AbortIOIfNeeded(uint64_t offset);
void AbortAllIOs();
void UpdateBuffersIfNeeded(uint64_t offset);
// It calls Poll API if any there is any pending asynchronous request. It then
// checks if data is in any buffer. It clears the outdated data and swaps the
// buffers if required.
void PollAndUpdateBuffersIfNeeded(uint64_t offset);
Status PrefetchAsyncInternal(const IOOptions& opts,
RandomAccessFileReader* reader, uint64_t offset,
size_t length, size_t readahead_size,
Env::IOPriority rate_limiter_priority,
bool& copy_to_third_buffer);
Status Read(const IOOptions& opts, RandomAccessFileReader* reader,
Env::IOPriority rate_limiter_priority, uint64_t read_len,
uint64_t chunk_len, uint64_t rounddown_start, uint32_t index);
Status ReadAsync(const IOOptions& opts, RandomAccessFileReader* reader,
uint64_t read_len, uint64_t rounddown_start, uint32_t index);
// Copy the data from src to third buffer.
void CopyDataToBuffer(uint32_t src, uint64_t& offset, size_t& length);
bool IsBlockSequential(const size_t& offset) {
return (prev_len_ == 0 || (prev_offset_ + prev_len_ == offset));
}
// Called in case of implicit auto prefetching.
void ResetValues() {
num_file_reads_ = 1;
readahead_size_ = initial_auto_readahead_size_;
}
// Called in case of implicit auto prefetching.
bool IsEligibleForPrefetch(uint64_t offset, size_t n) {
// Prefetch only if this read is sequential otherwise reset readahead_size_
// to initial value.
if (!IsBlockSequential(offset)) {
UpdateReadPattern(offset, n, false /*decrease_readaheadsize*/);
ResetValues();
return false;
}
num_file_reads_++;
// Since async request was submitted in last call directly by calling
// PrefetchAsync, it skips num_file_reads_ check as this call is to poll the
// data submitted in previous call.
if (explicit_prefetch_submitted_) {
return true;
}
if (num_file_reads_ <= num_file_reads_for_auto_readahead_) {
UpdateReadPattern(offset, n, false /*decrease_readaheadsize*/);
return false;
}
return true;
}
// Helper functions.
bool IsDataBlockInBuffer(uint64_t offset, size_t length, uint32_t index) {
return (offset >= bufs_[index].offset_ &&
offset + length <=
bufs_[index].offset_ + bufs_[index].buffer_.CurrentSize());
}
bool IsOffsetInBuffer(uint64_t offset, uint32_t index) {
return (offset >= bufs_[index].offset_ &&
offset < bufs_[index].offset_ + bufs_[index].buffer_.CurrentSize());
}
bool DoesBufferContainData(uint32_t index) {
return bufs_[index].buffer_.CurrentSize() > 0;
}
bool IsBufferOutdated(uint64_t offset, uint32_t index) {
return (
!bufs_[index].async_read_in_progress_ && DoesBufferContainData(index) &&
offset >= bufs_[index].offset_ + bufs_[index].buffer_.CurrentSize());
}
bool IsBufferOutdatedWithAsyncProgress(uint64_t offset, uint32_t index) {
return (bufs_[index].async_read_in_progress_ &&
bufs_[index].io_handle_ != nullptr &&
offset >= bufs_[index].offset_ + bufs_[index].async_req_len_);
}
bool IsOffsetInBufferWithAsyncProgress(uint64_t offset, uint32_t index) {
return (bufs_[index].async_read_in_progress_ &&
offset >= bufs_[index].offset_ &&
offset < bufs_[index].offset_ + bufs_[index].async_req_len_);
}
bool IsSecondBuffEligibleForPrefetching() {
uint32_t second = curr_ ^ 1;
if (bufs_[second].async_read_in_progress_) {
return false;
}
assert(!bufs_[curr_].async_read_in_progress_);
if (DoesBufferContainData(curr_) && DoesBufferContainData(second) &&
(bufs_[curr_].offset_ + bufs_[curr_].buffer_.CurrentSize() ==
bufs_[second].offset_)) {
return false;
}
bufs_[second].buffer_.Clear();
return true;
}
void DestroyAndClearIOHandle(uint32_t index) {
if (bufs_[index].io_handle_ != nullptr && bufs_[index].del_fn_ != nullptr) {
bufs_[index].del_fn_(bufs_[index].io_handle_);
bufs_[index].io_handle_ = nullptr;
bufs_[index].del_fn_ = nullptr;
}
bufs_[index].async_read_in_progress_ = false;
}
Status HandleOverlappingData(const IOOptions& opts,
RandomAccessFileReader* reader, uint64_t offset,
size_t length, size_t readahead_size,
Env::IOPriority rate_limiter_priority,
bool& copy_to_third_buffer, uint64_t& tmp_offset,
size_t& tmp_length);
bool TryReadFromCacheUntracked(const IOOptions& opts,
RandomAccessFileReader* reader,
uint64_t offset, size_t n, Slice* result,
Status* s,
Env::IOPriority rate_limiter_priority,
bool for_compaction = false);
bool TryReadFromCacheAsyncUntracked(const IOOptions& opts,
RandomAccessFileReader* reader,
uint64_t offset, size_t n, Slice* result,
Status* status,
Env::IOPriority rate_limiter_priority);
std::vector<BufferInfo> bufs_;
// curr_ represents the index for bufs_ indicating which buffer is being
// consumed currently.
uint32_t curr_;
size_t readahead_size_;
size_t initial_auto_readahead_size_;
// FilePrefetchBuffer object won't be created from Iterator flow if
// max_readahead_size_ = 0.
size_t max_readahead_size_;
// The minimum `offset` ever passed to TryReadFromCache().
size_t min_offset_read_;
// if false, TryReadFromCache() always return false, and we only take stats
// for track_min_offset_ if track_min_offset_ = true
bool enable_;
// If true, track minimum `offset` ever passed to TryReadFromCache(), which
// can be fetched from min_offset_read().
bool track_min_offset_;
// implicit_auto_readahead is enabled by rocksdb internally after 2
// sequential IOs.
bool implicit_auto_readahead_;
uint64_t prev_offset_;
size_t prev_len_;
// num_file_reads_ and num_file_reads_for_auto_readahead_ is only used when
// implicit_auto_readahead_ is set.
uint64_t num_file_reads_for_auto_readahead_;
uint64_t num_file_reads_;
// If explicit_prefetch_submitted_ is set then it indicates RocksDB called
// PrefetchAsync to submit request. It needs to call TryReadFromCacheAsync to
// poll the submitted request without checking if data is sequential and
// num_file_reads_.
bool explicit_prefetch_submitted_;
FileSystem* fs_;
SystemClock* clock_;
Statistics* stats_;
FilePrefetchBufferUsage usage_;
};
} // namespace ROCKSDB_NAMESPACE