db_impl.h 6.4 KB
Newer Older
J
jorlow@chromium.org 已提交
1 2 3 4 5 6 7
// 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.

#ifndef STORAGE_LEVELDB_DB_DB_IMPL_H_
#define STORAGE_LEVELDB_DB_DB_IMPL_H_

8
#include <deque>
J
jorlow@chromium.org 已提交
9 10 11 12
#include <set>
#include "db/dbformat.h"
#include "db/log_writer.h"
#include "db/snapshot.h"
13 14
#include "leveldb/db.h"
#include "leveldb/env.h"
J
jorlow@chromium.org 已提交
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
#include "port/port.h"

namespace leveldb {

class MemTable;
class TableCache;
class Version;
class VersionEdit;
class VersionSet;

class DBImpl : public DB {
 public:
  DBImpl(const Options& options, const std::string& dbname);
  virtual ~DBImpl();

  // Implementations of the DB interface
  virtual Status Put(const WriteOptions&, const Slice& key, const Slice& value);
  virtual Status Delete(const WriteOptions&, const Slice& key);
  virtual Status Write(const WriteOptions& options, WriteBatch* updates);
  virtual Status Get(const ReadOptions& options,
                     const Slice& key,
                     std::string* value);
  virtual Iterator* NewIterator(const ReadOptions&);
  virtual const Snapshot* GetSnapshot();
  virtual void ReleaseSnapshot(const Snapshot* snapshot);
40
  virtual bool GetProperty(const Slice& property, std::string* value);
J
jorlow@chromium.org 已提交
41
  virtual void GetApproximateSizes(const Range* range, int n, uint64_t* sizes);
G
Gabor Cselle 已提交
42
  virtual void CompactRange(const Slice* begin, const Slice* end);
J
jorlow@chromium.org 已提交
43 44 45

  // Extra methods (for testing) that are not in the public DB interface

G
Gabor Cselle 已提交
46 47
  // Compact any files in the named level that overlap [*begin,*end]
  void TEST_CompactRange(int level, const Slice* begin, const Slice* end);
J
jorlow@chromium.org 已提交
48 49 50 51 52 53 54 55 56

  // Force current memtable contents to be compacted.
  Status TEST_CompactMemTable();

  // Return an internal iterator over the current state of the database.
  // The keys of this iterator are internal keys (see format.h).
  // The returned iterator should be deleted when no longer needed.
  Iterator* TEST_NewInternalIterator();

57 58
  // Return the maximum overlapping data (in bytes) at next level for any
  // file at a level >= 1.
J
jorlow@chromium.org 已提交
59
  int64_t TEST_MaxNextLevelOverlappingBytes();
60

J
jorlow@chromium.org 已提交
61 62
 private:
  friend class DB;
63 64
  struct CompactionState;
  struct Writer;
J
jorlow@chromium.org 已提交
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88

  Iterator* NewInternalIterator(const ReadOptions&,
                                SequenceNumber* latest_snapshot);

  Status NewDB();

  // Recover the descriptor from persistent storage.  May do a significant
  // amount of work to recover recently logged updates.  Any changes to
  // be made to the descriptor are added to *edit.
  Status Recover(VersionEdit* edit);

  void MaybeIgnoreError(Status* s) const;

  // Delete any unneeded files and stale in-memory entries.
  void DeleteObsoleteFiles();

  // Compact the in-memory write buffer to disk.  Switches to a new
  // log-file/memtable and writes a new descriptor iff successful.
  Status CompactMemTable();

  Status RecoverLogFile(uint64_t log_number,
                        VersionEdit* edit,
                        SequenceNumber* max_sequence);

89
  Status WriteLevel0Table(MemTable* mem, VersionEdit* edit, Version* base);
J
jorlow@chromium.org 已提交
90

91
  Status MakeRoomForWrite(bool force /* compact even if there is room? */);
92
  WriteBatch* BuildBatchGroup(Writer** last_writer);
J
jorlow@chromium.org 已提交
93 94 95 96 97 98 99 100 101 102 103 104 105 106 107

  void MaybeScheduleCompaction();
  static void BGWork(void* db);
  void BackgroundCall();
  void BackgroundCompaction();
  void CleanupCompaction(CompactionState* compact);
  Status DoCompactionWork(CompactionState* compact);

  Status OpenCompactionOutputFile(CompactionState* compact);
  Status FinishCompactionOutputFile(CompactionState* compact, Iterator* input);
  Status InstallCompactionResults(CompactionState* compact);

  // Constant after construction
  Env* const env_;
  const InternalKeyComparator internal_comparator_;
S
Sanjay Ghemawat 已提交
108
  const InternalFilterPolicy internal_filter_policy_;
J
jorlow@chromium.org 已提交
109 110
  const Options options_;  // options_.comparator == &internal_comparator_
  bool owns_info_log_;
111
  bool owns_cache_;
J
jorlow@chromium.org 已提交
112 113 114 115 116 117 118 119 120 121 122
  const std::string dbname_;

  // table_cache_ provides its own synchronization
  TableCache* table_cache_;

  // Lock over the persistent DB state.  Non-NULL iff successfully acquired.
  FileLock* db_lock_;

  // State below is protected by mutex_
  port::Mutex mutex_;
  port::AtomicPointer shutting_down_;
H
hans@chromium.org 已提交
123
  port::CondVar bg_cv_;          // Signalled when background work finishes
J
jorlow@chromium.org 已提交
124
  MemTable* mem_;
125 126
  MemTable* imm_;                // Memtable being compacted
  port::AtomicPointer has_imm_;  // So bg thread can detect non-NULL imm_
J
jorlow@chromium.org 已提交
127
  WritableFile* logfile_;
128
  uint64_t logfile_number_;
J
jorlow@chromium.org 已提交
129
  log::Writer* log_;
130 131 132 133 134

  // Queue of writers.
  std::deque<Writer*> writers_;
  WriteBatch* tmp_batch_;

J
jorlow@chromium.org 已提交
135 136 137 138 139 140 141 142 143
  SnapshotList snapshots_;

  // Set of table files to protect from deletion because they are
  // part of ongoing compactions.
  std::set<uint64_t> pending_outputs_;

  // Has a background compaction been scheduled or is running?
  bool bg_compaction_scheduled_;

H
hans@chromium.org 已提交
144 145 146
  // Information for a manual compaction
  struct ManualCompaction {
    int level;
G
Gabor Cselle 已提交
147 148 149 150
    bool done;
    const InternalKey* begin;   // NULL means beginning of key range
    const InternalKey* end;     // NULL means end of key range
    InternalKey tmp_storage;    // Used to keep track of compaction progress
H
hans@chromium.org 已提交
151 152
  };
  ManualCompaction* manual_compaction_;
J
jorlow@chromium.org 已提交
153 154 155 156 157 158

  VersionSet* versions_;

  // Have we encountered a background error in paranoid mode?
  Status bg_error_;

159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175
  // Per level compaction stats.  stats_[level] stores the stats for
  // compactions that produced data for the specified "level".
  struct CompactionStats {
    int64_t micros;
    int64_t bytes_read;
    int64_t bytes_written;

    CompactionStats() : micros(0), bytes_read(0), bytes_written(0) { }

    void Add(const CompactionStats& c) {
      this->micros += c.micros;
      this->bytes_read += c.bytes_read;
      this->bytes_written += c.bytes_written;
    }
  };
  CompactionStats stats_[config::kNumLevels];

J
jorlow@chromium.org 已提交
176 177 178 179 180 181 182 183 184 185 186 187 188
  // No copying allowed
  DBImpl(const DBImpl&);
  void operator=(const DBImpl&);

  const Comparator* user_comparator() const {
    return internal_comparator_.user_comparator();
  }
};

// Sanitize db options.  The caller should delete result.info_log if
// it is not equal to src.info_log.
extern Options SanitizeOptions(const std::string& db,
                               const InternalKeyComparator* icmp,
S
Sanjay Ghemawat 已提交
189
                               const InternalFilterPolicy* ipolicy,
J
jorlow@chromium.org 已提交
190 191
                               const Options& src);

H
Hans Wennborg 已提交
192
}  // namespace leveldb
J
jorlow@chromium.org 已提交
193 194

#endif  // STORAGE_LEVELDB_DB_DB_IMPL_H_