rocksdb/db/transaction_log_impl.cc
Mayank Agarwal f837f5b1c9 Making the transaction log iterator more robust
Summary:
strict essentially means that we MUST find the startsequence. Thus we should return if starteSequence is not found in the first file in case strict is set. This will take care of ending the iterator in case of permanent gaps due to corruptions in the log files
Also created NextImpl function that will have internal variable to distinguish whether Next is being called from StartSequence or by application.
Set NotFoudn::gaps status to give an indication of gaps happeneing.
Polished the inline documentation at various places

Test Plan:
* db_repl_stress test
* db_test relating to transaction log iterator
* fbcode/wormhole/rocksdb/rocks_log_iterator
* sigma production machine sigmafio032.prn1

Reviewers: dhruba

Reviewed By: dhruba

CC: leveldb

Differential Revision: https://reviews.facebook.net/D13689
2013-11-04 20:49:03 -08:00

259 lines
8.5 KiB
C++

// Copyright (c) 2013, Facebook, Inc. All rights reserved.
// This source code is licensed under the BSD-style license found in the
// LICENSE file in the root directory of this source tree. An additional grant
// of patent rights can be found in the PATENTS file in the same directory.
//
#include "db/transaction_log_impl.h"
#include "db/write_batch_internal.h"
namespace rocksdb {
TransactionLogIteratorImpl::TransactionLogIteratorImpl(
const std::string& dir,
const Options* options,
const EnvOptions& soptions,
const SequenceNumber seq,
std::unique_ptr<VectorLogPtr> files,
DBImpl const * const dbimpl) :
dir_(dir),
options_(options),
soptions_(soptions),
startingSequenceNumber_(seq),
files_(std::move(files)),
started_(false),
isValid_(false),
currentFileIndex_(0),
currentBatchSeq_(0),
currentLastSeq_(0),
dbimpl_(dbimpl) {
assert(files_ != nullptr);
assert(dbimpl_ != nullptr);
reporter_.env = options_->env;
reporter_.info_log = options_->info_log.get();
SeekToStartSequence(); // Seek till starting sequence
}
Status TransactionLogIteratorImpl::OpenLogFile(
const LogFile* logFile,
unique_ptr<SequentialFile>* file) {
Env* env = options_->env;
if (logFile->Type() == kArchivedLogFile) {
std::string fname = ArchivedLogFileName(dir_, logFile->LogNumber());
return env->NewSequentialFile(fname, file, soptions_);
} else {
std::string fname = LogFileName(dir_, logFile->LogNumber());
Status status = env->NewSequentialFile(fname, file, soptions_);
if (!status.ok()) {
// If cannot open file in DB directory.
// Try the archive dir, as it could have moved in the meanwhile.
fname = ArchivedLogFileName(dir_, logFile->LogNumber());
status = env->NewSequentialFile(fname, file, soptions_);
if (!status.ok()) {
return Status::IOError("Requested file not present in the dir");
}
}
return status;
}
}
BatchResult TransactionLogIteratorImpl::GetBatch() {
assert(isValid_); // cannot call in a non valid state.
BatchResult result;
result.sequence = currentBatchSeq_;
result.writeBatchPtr = std::move(currentBatch_);
return result;
}
Status TransactionLogIteratorImpl::status() {
return currentStatus_;
}
bool TransactionLogIteratorImpl::Valid() {
return started_ && isValid_;
}
bool TransactionLogIteratorImpl::RestrictedRead(
Slice* record,
std::string* scratch) {
// Don't read if no more complete entries to read from logs
if (currentLastSeq_ >= dbimpl_->GetLatestSequenceNumber()) {
return false;
}
return currentLogReader_->ReadRecord(record, scratch);
}
void TransactionLogIteratorImpl::SeekToStartSequence(
uint64_t startFileIndex,
bool strict) {
std::string scratch;
Slice record;
started_ = false;
isValid_ = false;
if (files_->size() <= startFileIndex) {
return;
}
Status s = OpenLogReader(files_->at(startFileIndex).get());
if (!s.ok()) {
currentStatus_ = s;
return;
}
while (RestrictedRead(&record, &scratch)) {
if (record.size() < 12) {
reporter_.Corruption(
record.size(), Status::Corruption("very small log record"));
continue;
}
UpdateCurrentWriteBatch(record);
if (currentLastSeq_ >= startingSequenceNumber_) {
if (strict && currentBatchSeq_ != startingSequenceNumber_) {
currentStatus_ = Status::Corruption("Gap in sequence number. Could not "
"seek to required sequence number");
reporter_.Info(currentStatus_.ToString().c_str());
return;
} else if (strict) {
reporter_.Info("Could seek required sequence number. Iterator will "
"continue.");
}
isValid_ = true;
started_ = true; // set started_ as we could seek till starting sequence
return;
} else {
isValid_ = false;
}
}
// Could not find start sequence in first file. Normally this must be the
// only file. Otherwise log the error and let the iterator return next entry
// If strict is set, we want to seek exactly till the start sequence and it
// should have been present in the file we scanned above
if (strict) {
currentStatus_ = Status::Corruption("Gap in sequence number. Could not "
"seek to required sequence number");
reporter_.Info(currentStatus_.ToString().c_str());
} else if (files_->size() != 1) {
currentStatus_ = Status::Corruption("Start sequence was not found, "
"skipping to the next available");
reporter_.Info(currentStatus_.ToString().c_str());
// Let NextImpl find the next available entry. started_ remains false
// because we don't want to check for gaps while moving to start sequence
NextImpl(true);
}
}
void TransactionLogIteratorImpl::Next() {
return NextImpl(false);
}
void TransactionLogIteratorImpl::NextImpl(bool internal) {
std::string scratch;
Slice record;
isValid_ = false;
if (!internal && !started_) {
// Runs every time until we can seek to the start sequence
return SeekToStartSequence();
}
while(true) {
assert(currentLogReader_);
if (currentLogReader_->IsEOF()) {
currentLogReader_->UnmarkEOF();
}
while (RestrictedRead(&record, &scratch)) {
if (record.size() < 12) {
reporter_.Corruption(
record.size(), Status::Corruption("very small log record"));
continue;
} else {
// started_ should be true if called by application
assert(internal || started_);
// started_ should be false if called internally
assert(!internal || !started_);
UpdateCurrentWriteBatch(record);
if (internal && !started_) {
started_ = true;
}
return;
}
}
// Open the next file
if (currentFileIndex_ < files_->size() - 1) {
++currentFileIndex_;
Status status =OpenLogReader(files_->at(currentFileIndex_).get());
if (!status.ok()) {
isValid_ = false;
currentStatus_ = status;
return;
}
} else {
isValid_ = false;
if (currentLastSeq_ == dbimpl_->GetLatestSequenceNumber()) {
currentStatus_ = Status::OK();
} else {
currentStatus_ = Status::IOError("NO MORE DATA LEFT");
}
return;
}
}
}
bool TransactionLogIteratorImpl::IsBatchExpected(
const WriteBatch* batch,
const SequenceNumber expectedSeq) {
assert(batch);
SequenceNumber batchSeq = WriteBatchInternal::Sequence(batch);
if (batchSeq != expectedSeq) {
char buf[200];
snprintf(buf, sizeof(buf),
"Discontinuity in log records. Got seq=%lu, Expected seq=%lu, "
"Last flushed seq=%lu.Log iterator will reseek the correct batch.",
batchSeq, expectedSeq, dbimpl_->GetLatestSequenceNumber());
reporter_.Info(buf);
return false;
}
return true;
}
void TransactionLogIteratorImpl::UpdateCurrentWriteBatch(const Slice& record) {
WriteBatch* batch = new WriteBatch();
WriteBatchInternal::SetContents(batch, record);
SequenceNumber expectedSeq = currentLastSeq_ + 1;
// If the iterator has started, then confirm that we get continuous batches
if (started_ && !IsBatchExpected(batch, expectedSeq)) {
// Seek to the batch having expected sequence number
if (expectedSeq < files_->at(currentFileIndex_)->StartSequence()) {
// Expected batch must lie in the previous log file
currentFileIndex_--;
currentFileIndex_ = (currentFileIndex_ >= 0) ? currentFileIndex_ : 0;
}
startingSequenceNumber_ = expectedSeq;
// currentStatus_ will be set to Ok if reseek succeeds
currentStatus_ = Status::NotFound("Gap in sequence numbers");
return SeekToStartSequence(currentFileIndex_, true);
}
currentBatchSeq_ = WriteBatchInternal::Sequence(batch);
currentLastSeq_ = currentBatchSeq_ + WriteBatchInternal::Count(batch) - 1;
// currentBatchSeq_ can only change here
assert(currentLastSeq_ <= dbimpl_->GetLatestSequenceNumber());
currentBatch_.reset(batch);
isValid_ = true;
currentStatus_ = Status::OK();
}
Status TransactionLogIteratorImpl::OpenLogReader(const LogFile* logFile) {
unique_ptr<SequentialFile> file;
Status status = OpenLogFile(logFile, &file);
if (!status.ok()) {
return status;
}
assert(file);
currentLogReader_.reset(
new log::Reader(std::move(file), &reporter_, true, 0)
);
return Status::OK();
}
} // namespace rocksdb