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

165 lines
5.0 KiB

// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "db/log_writer.h"
#include <stdint.h>
#include "file/writable_file_writer.h"
#include "rocksdb/env.h"
#include "util/coding.h"
#include "util/crc32c.h"
namespace ROCKSDB_NAMESPACE {
namespace log {
Writer::Writer(std::unique_ptr<WritableFileWriter>&& dest, uint64_t log_number,
Optimize for serial commits in 2PC Summary: Throughput: 46k tps in our sysbench settings (filling the details later) The idea is to have the simplest change that gives us a reasonable boost in 2PC throughput. Major design changes: 1. The WAL file internal buffer is not flushed after each write. Instead it is flushed before critical operations (WAL copy via fs) or when FlushWAL is called by MySQL. Flushing the WAL buffer is also protected via mutex_. 2. Use two sequence numbers: last seq, and last seq for write. Last seq is the last visible sequence number for reads. Last seq for write is the next sequence number that should be used to write to WAL/memtable. This allows to have a memtable write be in parallel to WAL writes. 3. BatchGroup is not used for writes. This means that we can have parallel writers which changes a major assumption in the code base. To accommodate for that i) allow only 1 WriteImpl that intends to write to memtable via mem_mutex_--which is fine since in 2PC almost all of the memtable writes come via group commit phase which is serial anyway, ii) make all the parts in the code base that assumed to be the only writer (via EnterUnbatched) to also acquire mem_mutex_, iii) stat updates are protected via a stat_mutex_. Note: the first commit has the approach figured out but is not clean. Submitting the PR anyway to get the early feedback on the approach. If we are ok with the approach I will go ahead with this updates: 0) Rebase with Yi's pipelining changes 1) Currently batching is disabled by default to make sure that it will be consistent with all unit tests. Will make this optional via a config. 2) A couple of unit tests are disabled. They need to be updated with the serial commit of 2PC taken into account. 3) Replacing BatchGroup with mem_mutex_ got a bit ugly as it requires releasing mutex_ beforehand (the same way EnterUnbatched does). This needs to be cleaned up. Closes https://github.com/facebook/rocksdb/pull/2345 Differential Revision: D5210732 Pulled By: maysamyabandeh fbshipit-source-id: 78653bd95a35cd1e831e555e0e57bdfd695355a4
8 years ago
bool recycle_log_files, bool manual_flush)
: dest_(std::move(dest)),
block_offset_(0),
log_number_(log_number),
Optimize for serial commits in 2PC Summary: Throughput: 46k tps in our sysbench settings (filling the details later) The idea is to have the simplest change that gives us a reasonable boost in 2PC throughput. Major design changes: 1. The WAL file internal buffer is not flushed after each write. Instead it is flushed before critical operations (WAL copy via fs) or when FlushWAL is called by MySQL. Flushing the WAL buffer is also protected via mutex_. 2. Use two sequence numbers: last seq, and last seq for write. Last seq is the last visible sequence number for reads. Last seq for write is the next sequence number that should be used to write to WAL/memtable. This allows to have a memtable write be in parallel to WAL writes. 3. BatchGroup is not used for writes. This means that we can have parallel writers which changes a major assumption in the code base. To accommodate for that i) allow only 1 WriteImpl that intends to write to memtable via mem_mutex_--which is fine since in 2PC almost all of the memtable writes come via group commit phase which is serial anyway, ii) make all the parts in the code base that assumed to be the only writer (via EnterUnbatched) to also acquire mem_mutex_, iii) stat updates are protected via a stat_mutex_. Note: the first commit has the approach figured out but is not clean. Submitting the PR anyway to get the early feedback on the approach. If we are ok with the approach I will go ahead with this updates: 0) Rebase with Yi's pipelining changes 1) Currently batching is disabled by default to make sure that it will be consistent with all unit tests. Will make this optional via a config. 2) A couple of unit tests are disabled. They need to be updated with the serial commit of 2PC taken into account. 3) Replacing BatchGroup with mem_mutex_ got a bit ugly as it requires releasing mutex_ beforehand (the same way EnterUnbatched does). This needs to be cleaned up. Closes https://github.com/facebook/rocksdb/pull/2345 Differential Revision: D5210732 Pulled By: maysamyabandeh fbshipit-source-id: 78653bd95a35cd1e831e555e0e57bdfd695355a4
8 years ago
recycle_log_files_(recycle_log_files),
manual_flush_(manual_flush) {
for (int i = 0; i <= kMaxRecordType; i++) {
char t = static_cast<char>(i);
type_crc_[i] = crc32c::Value(&t, 1);
}
}
Writer::~Writer() {
if (dest_) {
WriteBuffer().PermitUncheckedError();
}
}
Optimize for serial commits in 2PC Summary: Throughput: 46k tps in our sysbench settings (filling the details later) The idea is to have the simplest change that gives us a reasonable boost in 2PC throughput. Major design changes: 1. The WAL file internal buffer is not flushed after each write. Instead it is flushed before critical operations (WAL copy via fs) or when FlushWAL is called by MySQL. Flushing the WAL buffer is also protected via mutex_. 2. Use two sequence numbers: last seq, and last seq for write. Last seq is the last visible sequence number for reads. Last seq for write is the next sequence number that should be used to write to WAL/memtable. This allows to have a memtable write be in parallel to WAL writes. 3. BatchGroup is not used for writes. This means that we can have parallel writers which changes a major assumption in the code base. To accommodate for that i) allow only 1 WriteImpl that intends to write to memtable via mem_mutex_--which is fine since in 2PC almost all of the memtable writes come via group commit phase which is serial anyway, ii) make all the parts in the code base that assumed to be the only writer (via EnterUnbatched) to also acquire mem_mutex_, iii) stat updates are protected via a stat_mutex_. Note: the first commit has the approach figured out but is not clean. Submitting the PR anyway to get the early feedback on the approach. If we are ok with the approach I will go ahead with this updates: 0) Rebase with Yi's pipelining changes 1) Currently batching is disabled by default to make sure that it will be consistent with all unit tests. Will make this optional via a config. 2) A couple of unit tests are disabled. They need to be updated with the serial commit of 2PC taken into account. 3) Replacing BatchGroup with mem_mutex_ got a bit ugly as it requires releasing mutex_ beforehand (the same way EnterUnbatched does). This needs to be cleaned up. Closes https://github.com/facebook/rocksdb/pull/2345 Differential Revision: D5210732 Pulled By: maysamyabandeh fbshipit-source-id: 78653bd95a35cd1e831e555e0e57bdfd695355a4
8 years ago
IOStatus Writer::WriteBuffer() { return dest_->Flush(); }
IOStatus Writer::Close() {
IOStatus s;
if (dest_) {
s = dest_->Close();
dest_.reset();
}
return s;
}
IOStatus Writer::AddRecord(const Slice& slice) {
const char* ptr = slice.data();
size_t left = slice.size();
// Header size varies depending on whether we are recycling or not.
const int header_size =
recycle_log_files_ ? kRecyclableHeaderSize : kHeaderSize;
// Fragment the record if necessary and emit it. Note that if slice
// is empty, we still want to iterate once to emit a single
// zero-length record
IOStatus s;
bool begin = true;
do {
const int64_t leftover = kBlockSize - block_offset_;
assert(leftover >= 0);
if (leftover < header_size) {
// Switch to a new block
if (leftover > 0) {
// Fill the trailer (literal below relies on kHeaderSize and
// kRecyclableHeaderSize being <= 11)
assert(header_size <= 11);
s = dest_->Append(Slice("\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00",
static_cast<size_t>(leftover)));
if (!s.ok()) {
break;
}
}
block_offset_ = 0;
}
// Invariant: we never leave < header_size bytes in a block.
assert(static_cast<int64_t>(kBlockSize - block_offset_) >= header_size);
const size_t avail = kBlockSize - block_offset_ - header_size;
const size_t fragment_length = (left < avail) ? left : avail;
RecordType type;
const bool end = (left == fragment_length);
if (begin && end) {
type = recycle_log_files_ ? kRecyclableFullType : kFullType;
} else if (begin) {
type = recycle_log_files_ ? kRecyclableFirstType : kFirstType;
} else if (end) {
type = recycle_log_files_ ? kRecyclableLastType : kLastType;
} else {
type = recycle_log_files_ ? kRecyclableMiddleType : kMiddleType;
}
s = EmitPhysicalRecord(type, ptr, fragment_length);
ptr += fragment_length;
left -= fragment_length;
begin = false;
} while (s.ok() && left > 0);
if (s.ok()) {
if (!manual_flush_) {
s = dest_->Flush();
}
}
return s;
}
bool Writer::TEST_BufferIsEmpty() { return dest_->TEST_BufferIsEmpty(); }
IOStatus Writer::EmitPhysicalRecord(RecordType t, const char* ptr, size_t n) {
assert(n <= 0xffff); // Must fit in two bytes
size_t header_size;
char buf[kRecyclableHeaderSize];
// Format the header
buf[4] = static_cast<char>(n & 0xff);
buf[5] = static_cast<char>(n >> 8);
buf[6] = static_cast<char>(t);
uint32_t crc = type_crc_[t];
if (t < kRecyclableFullType) {
// Legacy record format
assert(block_offset_ + kHeaderSize + n <= kBlockSize);
header_size = kHeaderSize;
} else {
// Recyclable record format
assert(block_offset_ + kRecyclableHeaderSize + n <= kBlockSize);
header_size = kRecyclableHeaderSize;
// Only encode low 32-bits of the 64-bit log number. This means
// we will fail to detect an old record if we recycled a log from
// ~4 billion logs ago, but that is effectively impossible, and
// even if it were we'dbe far more likely to see a false positive
// on the 32-bit CRC.
EncodeFixed32(buf + 7, static_cast<uint32_t>(log_number_));
crc = crc32c::Extend(crc, buf + 7, 4);
}
// Compute the crc of the record type and the payload.
crc = crc32c::Extend(crc, ptr, n);
crc = crc32c::Mask(crc); // Adjust for storage
Fail recovery when MANIFEST record checksum mismatch (#6996) Summary: https://github.com/facebook/rocksdb/issues/5411 refactored `VersionSet::Recover` but introduced a bug, explained as follows. Before, once a checksum mismatch happens, `reporter` will set `s` to be non-ok. Therefore, Recover will stop processing the MANIFEST any further. ``` // Correct // Inside Recover LogReporter reporter; reporter.status = &s; log::Reader reader(..., reporter); while (reader.ReadRecord() && s.ok()) { ... } ``` The bug is that, the local variable `s` in `ReadAndRecover` won't be updated by `reporter` while reading the MANIFEST. It is possible that the reader sees a checksum mismatch in a record, but `ReadRecord` retries internally read and finds the next valid record. The mismatched record will be ignored and no error is reported. ``` // Incorrect // Inside Recover LogReporter reporter; reporter.status = &s; log::Reader reader(..., reporter); s = ReadAndRecover(reader, ...); // Inside ReadAndRecover Status s; // Shadows the s in Recover. while (reader.ReadRecord() && s.ok()) { ... } ``` `LogReporter` can use a separate `log_read_status` to track the errors while reading the MANIFEST. RocksDB can process more MANIFEST entries only if `log_read_status.ok()`. Test plan (devserver): make check Pull Request resolved: https://github.com/facebook/rocksdb/pull/6996 Reviewed By: ajkr Differential Revision: D22105746 Pulled By: riversand963 fbshipit-source-id: b22f717a423457a41ca152a242abbb64cf91fc38
5 years ago
TEST_SYNC_POINT_CALLBACK("LogWriter::EmitPhysicalRecord:BeforeEncodeChecksum",
&crc);
EncodeFixed32(buf, crc);
// Write the header and the payload
IOStatus s = dest_->Append(Slice(buf, header_size));
if (s.ok()) {
s = dest_->Append(Slice(ptr, n));
}
block_offset_ += header_size + n;
return s;
}
} // namespace log
} // namespace ROCKSDB_NAMESPACE