leveldb源码解析三——memtable、WAL和迭代器
leveldb在写入的时候，先将KV插入内存中的memtable，然后写入WAL
Memtable
class MemTable {
...
// key比较器
KeyComparator comparator_;
// 引用个数
int refs_;
// 内存分配
Arena arena_;
// 跳表，存储KV
Table table_;
};
添加KV:
void MemTable::Add(SequenceNumber s, ValueType type, const Slice& key,
const Slice& value) {
// Format of an entry is concatenation of:
// key_size : varint32 of internal_key.size()
// key bytes : char[internal_key.size()]
// value_size : varint32 of value.size()
// value bytes : char[value.size()]
size_t key_size = key.size();
size_t val_size = value.size();
size_t internal_key_size = key_size + 8;
const size_t encoded_len = VarintLength(internal_key_size) +
internal_key_size + VarintLength(val_size) +
val_size;
char* buf = arena_.Allocate(encoded_len);
char* p = EncodeVarint32(buf, internal_key_size);
std::memcpy(p, key.data(), key_size);
p += key_size;
EncodeFixed64(p, (s << 8) | type);
p += 8;
p = EncodeVarint32(p, val_size);
std::memcpy(p, value.data(), val_size);
assert(p + val_size == buf + encoded_len);
table_.Insert(buf);
}
查找KV:
bool MemTable::Get(const LookupKey& key, std::string* value, Status* s) {
Slice memkey = key.memtable_key();
// 通过跳表中的迭代器查找
Table::Iterator iter(&table_);
iter.Seek(memkey.data());
if (iter.Valid()) {
// entry format is:
// klength varint32
// userkey char[klength]
// tag uint64
// vlength varint32
// value char[vlength]
// Check that it belongs to same user key. We do not check the
// sequence number since the Seek() call above should have skipped
// all entries with overly large sequence numbers.
const char* entry = iter.key();
uint32_t key_length;
const char* key_ptr = GetVarint32Ptr(entry, entry + 5, &key_length);
if (comparator_.comparator.user_comparator()->Compare(
Slice(key_ptr, key_length - 8), key.user_key()) == 0) {
// Correct user key
const uint64_t tag = DecodeFixed64(key_ptr + key_length - 8);
switch (static_cast<ValueType>(tag & 0xff)) {
case kTypeValue: {
Slice v = GetLengthPrefixedSlice(key_ptr + key_length);
value->assign(v.data(), v.size());
return true;
}
case kTypeDeletion:
*s = Status::NotFound(Slice());
return true;
}
}
}
return false;
}
WAL
wal日志按照block为单位进行存储的，每个block大小为32k，而每个block是由一系列record组成的。每个record的大小无法确定，有可能会跨越block，所以每个record包括类型，说明这个record是否跨越block。
Record写入：
Status Writer::AddRecord(const Slice& slice) {
const char* ptr = slice.data();
size_t left = slice.size();
// 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
Status s;
bool begin = true;
do {
const int leftover = kBlockSize - block_offset_;
assert(leftover >= 0);
if (leftover < kHeaderSize) {
// 不够header，则padding，然后创建新的block
// Switch to a new block
if (leftover > 0) {
// Fill the trailer (literal below relies on kHeaderSize being 7)
static_assert(kHeaderSize == 7, "");
dest_->Append(Slice("\x00\x00\x00\x00\x00\x00", leftover));
}
block_offset_ = 0;
}
// Invariant: we never leave < kHeaderSize bytes in a block.
assert(kBlockSize - block_offset_ - kHeaderSize >= 0);
const size_t avail = kBlockSize - block_offset_ - kHeaderSize;
const size_t fragment_length = (left < avail) ? left : avail;
// 根据record大小和block剩余空间大小，确定本record类型
RecordType type;
const bool end = (left == fragment_length);
if (begin && end) {
type = kFullType;
} else if (begin) {
type = kFirstType;
} else if (end) {
type = kLastType;
} else {
type = kMiddleType;
}
// 编码并写入磁盘
s = EmitPhysicalRecord(type, ptr, fragment_length);
ptr += fragment_length;
left -= fragment_length;
begin = false;
} while (s.ok() && left > 0);
return s;
}
Record读取时写入的反过程，把数据从磁盘上读出，并解析：
unsigned int Reader::ReadPhysicalRecord(Slice* result) {
while (true) {
if (buffer_.size() < kHeaderSize) {
if (!eof_) {
// Last read was a full read, so this is a trailer to skip
buffer_.clear();
Status status = file_->Read(kBlockSize, &buffer_, backing_store_);
end_of_buffer_offset_ += buffer_.size();
if (!status.ok()) {
buffer_.clear();
ReportDrop(kBlockSize, status);
eof_ = true;
return kEof;
} else if (buffer_.size() < kBlockSize) {
eof_ = true;
}
continue;
} else {
// Note that if buffer_ is non-empty, we have a truncated header at the
// end of the file, which can be caused by the writer crashing in the
// middle of writing the header. Instead of considering this an error,
// just report EOF.
buffer_.clear();
return kEof;
}
}
// Parse the header
const char* header = buffer_.data();
const uint32_t a = static_cast<uint32_t>(header[4]) & 0xff;
const uint32_t b = static_cast<uint32_t>(header[5]) & 0xff;
const unsigned int type = header[6];
const uint32_t length = a | (b << 8);
if (kHeaderSize + length > buffer_.size()) {
size_t drop_size = buffer_.size();
buffer_.clear();
if (!eof_) {
ReportCorruption(drop_size, "bad record length");
return kBadRecord;
}
// If the end of the file has been reached without reading |length| bytes
// of payload, assume the writer died in the middle of writing the record.
// Don't report a corruption.
return kEof;
}
if (type == kZeroType && length == 0) {
// Skip zero length record without reporting any drops since
// such records are produced by the mmap based writing code in
// env_posix.cc that preallocates file regions.
buffer_.clear();
return kBadRecord;
}
// Check crc
if (checksum_) {
uint32_t expected_crc = crc32c::Unmask(DecodeFixed32(header));
uint32_t actual_crc = crc32c::Value(header + 6, 1 + length);
if (actual_crc != expected_crc) {
// Drop the rest of the buffer since "length" itself may have
// been corrupted and if we trust it, we could find some
// fragment of a real log record that just happens to look
// like a valid log record.
size_t drop_size = buffer_.size();
buffer_.clear();
ReportCorruption(drop_size, "checksum mismatch");
return kBadRecord;
}
}
buffer_.remove_prefix(kHeaderSize + length);
// Skip physical record that started before initial_offset_
if (end_of_buffer_offset_ - buffer_.size() - kHeaderSize - length <
initial_offset_) {
result->clear();
return kBadRecord;
}
*result = Slice(header + kHeaderSize, length);
return type;
}
}
key的表现形式
用户传进来的key、WAL、memtable中的key和sstable中的key在编码方式上有所区别。
用户传进来的key是Slice。
WAL中的key编码方式：checksum (4 bytes), length (2 bytes), type (1 byte)，key_data。
memtable中key的编码方式：internal_key_size,key_data,seq+type(1 byte),value_size,value_data。memtable中把value也编码到key中
sstable中key的编码方式：shared_len,no_shared_len,value_len,no_shared_key_data,value_data。
另外为了方便编码，leveldb内部自定义了LookupKey、InternalKey、ParsedInternalKey，以上几折的关系为：
LookUpKey是user_key+seq的封装，用来从memtable中查询数据
InternalKey是user_key+seq+type的紧凑形式，ParsedInternalKey是InternalKey解析到内存中的形式
迭代器
无论是memtable、sstable、block，还是整个db，当scan或者查找某个key的时候，都是通过迭代器查找
memtable中的迭代器
memtable中的迭代器是内部skiplist迭代器的封装
class MemTable {
...
Iterator* NewIterator();
};
Iterator* MemTable::NewIterator() { return new MemTableIterator(&table_); }
class MemTableIterator : public Iterator {
public:
explicit MemTableIterator(MemTable::Table* table) : iter_(table) {}
MemTableIterator(const MemTableIterator&) = delete;
MemTableIterator& operator=(const MemTableIterator&) = delete;
~MemTableIterator() override = default;
bool Valid() const override { return iter_.Valid(); }
void Seek(const Slice& k) override { iter_.Seek(EncodeKey(&tmp_, k)); }
void SeekToFirst() override { iter_.SeekToFirst(); }
void SeekToLast() override { iter_.SeekToLast(); }
void Next() override { iter_.Next(); }
void Prev() override { iter_.Prev(); }
Slice key() const override { return GetLengthPrefixedSlice(iter_.key()); }
Slice value() const override {
Slice key_slice = GetLengthPrefixedSlice(iter_.key());
return GetLengthPrefixedSlice(key_slice.data() + key_slice.size());
}
Status status() const override { return Status::OK(); }
private:
MemTable::Table::Iterator iter_;
std::string tmp_; // For passing to EncodeKey
};
block和table中的迭代器
block中的迭代器就是根据block的重启点和block中record有序的特性，解析出一个个KV对
class Block {
...
Iterator* NewIterator(const Comparator* comparator);
...
}
Iterator* Block::NewIterator(const Comparator* comparator) {
if (size_ < sizeof(uint32_t)) {
return NewErrorIterator(Status::Corruption("bad block contents"));
}
const uint32_t num_restarts = NumRestarts();
if (num_restarts == 0) {
return NewEmptyIterator();
} else {
return new Iter(comparator, data_, restart_offset_, num_restarts);
}
}
void Seek(const Slice& target) override {
// Binary search in restart array to find the last restart point
// with a key < target
uint32_t left = 0;
uint32_t right = num_restarts_ - 1;
int current_key_compare = 0;
if (Valid()) {
// If we're already scanning, use the current position as a starting
// point. This is beneficial if the key we're seeking to is ahead of the
// current position.
// 如果当前以及开始遍历了，那么先用当前的遍历点作为参考点
current_key_compare = Compare(key_, target);
if (current_key_compare < 0) {
// key_ is smaller than target
left = restart_index_;
} else if (current_key_compare > 0) {
right = restart_index_;
} else {
// We're seeking to the key we're already at.
return;
}
}
// record有序，二分查找
while (left < right) {
uint32_t mid = (left + right + 1) / 2;
uint32_t region_offset = GetRestartPoint(mid);
uint32_t shared, non_shared, value_length;
const char* key_ptr =
DecodeEntry(data_ + region_offset, data_ + restarts_, &shared,
&non_shared, &value_length);
// 这是起始点出的KV对，所以shared的长度为0
if (key_ptr == nullptr || (shared != 0)) {
CorruptionError();
return;
}
Slice mid_key(key_ptr, non_shared);
if (Compare(mid_key, target) < 0) {
// Key at "mid" is smaller than "target". Therefore all
// blocks before "mid" are uninteresting.
left = mid;
} else {
// Key at "mid" is >= "target". Therefore all blocks at or
// after "mid" are uninteresting.
right = mid - 1;
}
}
// We might be able to use our current position within the restart block.
// This is true if we determined the key we desire is in the current block
// and is after than the current key.
assert(current_key_compare == 0 || Valid());
bool skip_seek = left == restart_index_ && current_key_compare < 0;
if (!skip_seek) {
SeekToRestartPoint(left);
}
// Linear search (within restart block) for first key >= target
while (true) {
if (!ParseNextKey()) {
return;
}
if (Compare(key_, target) >= 0) {
return;
}
}
}
table中的迭代器是两层迭代器，第一层用来迭代table中的block，第二层用来迭代block中的record，最终遍历查询所需的KV对
Iterator* Table::NewIterator(const ReadOptions& options) const {
return NewTwoLevelIterator(
rep_->index_block->NewIterator(rep_->options.comparator),
&Table::BlockReader, const_cast<Table*>(this), options);
}
void TwoLevelIterator::Seek(const Slice& target) {
// 第一层index_iter迭代block，找到对应target所在的block
index_iter_.Seek(target);
InitDataBlock();
// 第二层再在block中迭代，找到对应的record
if (data_iter_.iter() != nullptr) data_iter_.Seek(target);
SkipEmptyDataBlocksForward();
}