db.h 53.7 KB
Newer Older
1
// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
S
Siying Dong 已提交
2 3 4
//  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).
J
jorlow@chromium.org 已提交
5 6 7 8
// 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.

9 10
#ifndef STORAGE_ROCKSDB_INCLUDE_DB_H_
#define STORAGE_ROCKSDB_INCLUDE_DB_H_
J
jorlow@chromium.org 已提交
11 12 13

#include <stdint.h>
#include <stdio.h>
14
#include <map>
15
#include <memory>
16
#include <string>
17
#include <unordered_map>
18
#include <vector>
19
#include "rocksdb/iterator.h"
20
#include "rocksdb/listener.h"
21 22
#include "rocksdb/metadata.h"
#include "rocksdb/options.h"
A
agiardullo 已提交
23
#include "rocksdb/snapshot.h"
24
#include "rocksdb/sst_file_writer.h"
Y
Yueh-Hsuan Chiang 已提交
25
#include "rocksdb/thread_status.h"
26 27 28
#include "rocksdb/transaction_log.h"
#include "rocksdb/types.h"
#include "rocksdb/version.h"
J
jorlow@chromium.org 已提交
29

30 31 32 33 34
#ifdef _WIN32
// Windows API macro interference
#undef DeleteFile
#endif

35 36 37 38 39
#if defined(__GNUC__) || defined(__clang__)
#define ROCKSDB_DEPRECATED_FUNC __attribute__((__deprecated__))
#elif _WIN32
#define ROCKSDB_DEPRECATED_FUNC __declspec(deprecated)
#endif
40

41
namespace rocksdb {
J
jorlow@chromium.org 已提交
42

43 44 45 46 47 48 49
struct Options;
struct DBOptions;
struct ColumnFamilyOptions;
struct ReadOptions;
struct WriteOptions;
struct FlushOptions;
struct CompactionOptions;
50
struct CompactRangeOptions;
51
struct TableProperties;
52
struct ExternalSstFileInfo;
53 54 55
class WriteBatch;
class Env;
class EventListener;
56
enum EntryType;
57

58 59
using std::unique_ptr;

60
extern const std::string kDefaultColumnFamilyName;
61
struct ColumnFamilyDescriptor {
62
  std::string name;
63
  ColumnFamilyOptions options;
I
Igor Canadi 已提交
64
  ColumnFamilyDescriptor()
65
      : name(kDefaultColumnFamilyName), options(ColumnFamilyOptions()) {}
66 67 68
  ColumnFamilyDescriptor(const std::string& _name,
                         const ColumnFamilyOptions& _options)
      : name(_name), options(_options) {}
69 70
};

71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
class ColumnFamilyHandle {
 public:
  virtual ~ColumnFamilyHandle() {}
  // Returns the name of the column family associated with the current handle.
  virtual const std::string& GetName() const = 0;
  // Returns the ID of the column family associated with the current handle.
  virtual uint32_t GetID() const = 0;
  // Fills "*desc" with the up-to-date descriptor of the column family
  // associated with this handle. Since it fills "*desc" with the up-to-date
  // information, this call might internally lock and release DB mutex to
  // access the up-to-date CF options.  In addition, all the pointer-typed
  // options cannot be referenced any longer than the original options exist.
  //
  // Note that this function is not supported in RocksDBLite.
  virtual Status GetDescriptor(ColumnFamilyDescriptor* desc) = 0;
86 87 88
  // Returns the comparator of the column family associated with the
  // current handle.
  virtual const Comparator* GetComparator() const = 0;
89 90
};

91 92
static const int kMajorVersion = __ROCKSDB_MAJOR__;
static const int kMinorVersion = __ROCKSDB_MINOR__;
93

94
// A range of keys
J
jorlow@chromium.org 已提交
95
struct Range {
96 97
  Slice start;
  Slice limit;
J
jorlow@chromium.org 已提交
98

99
  Range() { }
J
jorlow@chromium.org 已提交
100 101 102
  Range(const Slice& s, const Slice& l) : start(s), limit(l) { }
};

103 104 105 106 107 108 109 110
struct RangePtr {
  const Slice* start;
  const Slice* limit;

  RangePtr() : start(nullptr), limit(nullptr) { }
  RangePtr(const Slice* s, const Slice* l) : start(s), limit(l) { }
};

111 112 113 114 115 116
// A collections of table properties objects, where
//  key: is the table's file name.
//  value: the table properties object of the given table.
typedef std::unordered_map<std::string, std::shared_ptr<const TableProperties>>
    TablePropertiesCollection;

J
jorlow@chromium.org 已提交
117
// A DB is a persistent ordered map from keys to values.
118 119
// A DB is safe for concurrent access from multiple threads without
// any external synchronization.
J
jorlow@chromium.org 已提交
120 121 122 123 124
class DB {
 public:
  // Open the database with the specified "name".
  // Stores a pointer to a heap-allocated database in *dbptr and returns
  // OK on success.
A
Abhishek Kona 已提交
125
  // Stores nullptr in *dbptr and returns a non-OK status on error.
J
jorlow@chromium.org 已提交
126 127 128 129 130
  // Caller should delete *dbptr when it is no longer needed.
  static Status Open(const Options& options,
                     const std::string& name,
                     DB** dbptr);

H
heyongqiang 已提交
131 132 133 134
  // Open the database for read only. All DB interfaces
  // that modify data, like put/delete, will return error.
  // If the db is opened in read only mode, then no compactions
  // will happen.
135 136 137
  //
  // Not supported in ROCKSDB_LITE, in which case the function will
  // return Status::NotSupported.
H
heyongqiang 已提交
138 139
  static Status OpenForReadOnly(const Options& options,
      const std::string& name, DB** dbptr,
140
      bool error_if_log_file_exist = false);
H
heyongqiang 已提交
141

142 143 144 145 146
  // Open the database for read only with column families. When opening DB with
  // read only, you can specify only a subset of column families in the
  // database that should be opened. However, you always need to specify default
  // column family. The default column family name is 'default' and it's stored
  // in rocksdb::kDefaultColumnFamilyName
147 148 149
  //
  // Not supported in ROCKSDB_LITE, in which case the function will
  // return Status::NotSupported.
150 151 152 153 154 155
  static Status OpenForReadOnly(
      const DBOptions& db_options, const std::string& name,
      const std::vector<ColumnFamilyDescriptor>& column_families,
      std::vector<ColumnFamilyHandle*>* handles, DB** dbptr,
      bool error_if_log_file_exist = false);

156 157
  // Open DB with column families.
  // db_options specify database specific options
J
Jonah Cohen 已提交
158
  // column_families is the vector of all column families in the database,
159 160 161 162 163 164
  // containing column family name and options. You need to open ALL column
  // families in the database. To get the list of column families, you can use
  // ListColumnFamilies(). Also, you can open only a subset of column families
  // for read-only access.
  // The default column family name is 'default' and it's stored
  // in rocksdb::kDefaultColumnFamilyName.
165 166
  // If everything is OK, handles will on return be the same size
  // as column_families --- handles[i] will be a handle that you
167 168 169
  // will use to operate on column family column_family[i].
  // Before delete DB, you have to close All column families by calling
  // DestroyColumnFamilyHandle() with all the handles.
170 171
  static Status Open(const DBOptions& db_options, const std::string& name,
                     const std::vector<ColumnFamilyDescriptor>& column_families,
172
                     std::vector<ColumnFamilyHandle*>* handles, DB** dbptr);
173

174
  // Close the DB by releasing resources, closing files etc. This should be
175
  // called before calling the destructor so that the caller can get back a
176
  // status in case there are any errors. This will not fsync the WAL files.
S
Siying Dong 已提交
177 178
  // If syncing is required, the caller must first call SyncWAL(), or Write()
  // using an empty write batch with WriteOptions.sync=true.
179 180 181 182
  // Regardless of the return status, the DB must be freed. If the return
  // status is NotSupported(), then the DB implementation does cleanup in the
  // destructor
  virtual Status Close() { return Status::NotSupported(); }
183

184 185 186 187
  // ListColumnFamilies will open the DB specified by argument name
  // and return the list of all column families in that DB
  // through column_families argument. The ordering of
  // column families in column_families is unspecified.
188 189 190
  static Status ListColumnFamilies(const DBOptions& db_options,
                                   const std::string& name,
                                   std::vector<std::string>* column_families);
191

J
jorlow@chromium.org 已提交
192 193 194
  DB() { }
  virtual ~DB();

195 196 197
  // Create a column_family and return the handle of column family
  // through the argument handle.
  virtual Status CreateColumnFamily(const ColumnFamilyOptions& options,
198
                                    const std::string& column_family_name,
199
                                    ColumnFamilyHandle** handle);
200

Y
Yi Wu 已提交
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219
  // Bulk create column families with the same column family options.
  // Return the handles of the column families through the argument handles.
  // In case of error, the request may succeed partially, and handles will
  // contain column family handles that it managed to create, and have size
  // equal to the number of created column families.
  virtual Status CreateColumnFamilies(
      const ColumnFamilyOptions& options,
      const std::vector<std::string>& column_family_names,
      std::vector<ColumnFamilyHandle*>* handles);

  // Bulk create column families.
  // Return the handles of the column families through the argument handles.
  // In case of error, the request may succeed partially, and handles will
  // contain column family handles that it managed to create, and have size
  // equal to the number of created column families.
  virtual Status CreateColumnFamilies(
      const std::vector<ColumnFamilyDescriptor>& column_families,
      std::vector<ColumnFamilyHandle*>* handles);

220 221 222 223
  // Drop a column family specified by column_family handle. This call
  // only records a drop record in the manifest and prevents the column
  // family from flushing and compacting.
  virtual Status DropColumnFamily(ColumnFamilyHandle* column_family);
Y
Yi Wu 已提交
224 225 226 227 228 229 230 231

  // Bulk drop column families. This call only records drop records in the
  // manifest and prevents the column families from flushing and compacting.
  // In case of error, the request may succeed partially. User may call
  // ListColumnFamilies to check the result.
  virtual Status DropColumnFamilies(
      const std::vector<ColumnFamilyHandle*>& column_families);

232 233 234 235 236
  // Close a column family specified by column_family handle and destroy
  // the column family handle specified to avoid double deletion. This call
  // deletes the column family handle by default. Use this method to
  // close column family instead of deleting column family handle directly
  virtual Status DestroyColumnFamilyHandle(ColumnFamilyHandle* column_family);
237

238
  // Set the database entry for "key" to "value".
239
  // If "key" already exists, it will be overwritten.
240
  // Returns OK on success, and a non-OK status on error.
241
  // Note: consider setting options.sync = true.
J
jorlow@chromium.org 已提交
242
  virtual Status Put(const WriteOptions& options,
243
                     ColumnFamilyHandle* column_family, const Slice& key,
J
jorlow@chromium.org 已提交
244
                     const Slice& value) = 0;
245 246
  virtual Status Put(const WriteOptions& options, const Slice& key,
                     const Slice& value) {
247
    return Put(options, DefaultColumnFamily(), key, value);
248
  }
J
jorlow@chromium.org 已提交
249 250 251 252

  // Remove the database entry (if any) for "key".  Returns OK on
  // success, and a non-OK status on error.  It is not an error if "key"
  // did not exist in the database.
253
  // Note: consider setting options.sync = true.
254
  virtual Status Delete(const WriteOptions& options,
255
                        ColumnFamilyHandle* column_family,
256
                        const Slice& key) = 0;
257
  virtual Status Delete(const WriteOptions& options, const Slice& key) {
258
    return Delete(options, DefaultColumnFamily(), key);
259
  }
J
jorlow@chromium.org 已提交
260

A
Andres Noetzli 已提交
261 262 263
  // Remove the database entry for "key". Requires that the key exists
  // and was not overwritten. Returns OK on success, and a non-OK status
  // on error.  It is not an error if "key" did not exist in the database.
A
agiardullo 已提交
264 265 266 267 268 269 270 271 272 273 274 275
  //
  // If a key is overwritten (by calling Put() multiple times), then the result
  // of calling SingleDelete() on this key is undefined.  SingleDelete() only
  // behaves correctly if there has been only one Put() for this key since the
  // previous call to SingleDelete() for this key.
  //
  // This feature is currently an experimental performance optimization
  // for a very specific workload.  It is up to the caller to ensure that
  // SingleDelete is only used for a key that is not deleted using Delete() or
  // written using Merge().  Mixing SingleDelete operations with Deletes and
  // Merges can result in undefined behavior.
  //
A
Andres Noetzli 已提交
276 277 278 279 280 281 282 283
  // Note: consider setting options.sync = true.
  virtual Status SingleDelete(const WriteOptions& options,
                              ColumnFamilyHandle* column_family,
                              const Slice& key) = 0;
  virtual Status SingleDelete(const WriteOptions& options, const Slice& key) {
    return SingleDelete(options, DefaultColumnFamily(), key);
  }

A
Andrew Kryczka 已提交
284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
  // Removes the database entries in the range ["begin_key", "end_key"), i.e.,
  // including "begin_key" and excluding "end_key". Returns OK on success, and
  // a non-OK status on error. It is not an error if no keys exist in the range
  // ["begin_key", "end_key").
  //
  // This feature is currently an experimental performance optimization for
  // deleting very large ranges of contiguous keys. Invoking it many times or on
  // small ranges may severely degrade read performance; in particular, the
  // resulting performance can be worse than calling Delete() for each key in
  // the range. Note also the degraded read performance affects keys outside the
  // deleted ranges, and affects database operations involving scans, like flush
  // and compaction.
  //
  // Consider setting ReadOptions::ignore_range_deletions = true to speed
  // up reads for key(s) that are known to be unaffected by range deletions.
  virtual Status DeleteRange(const WriteOptions& options,
                             ColumnFamilyHandle* column_family,
                             const Slice& begin_key, const Slice& end_key);

303 304 305 306 307
  // Merge the database entry for "key" with "value".  Returns OK on success,
  // and a non-OK status on error. The semantics of this operation is
  // determined by the user provided merge_operator when opening DB.
  // Note: consider setting options.sync = true.
  virtual Status Merge(const WriteOptions& options,
308 309
                       ColumnFamilyHandle* column_family, const Slice& key,
                       const Slice& value) = 0;
310 311
  virtual Status Merge(const WriteOptions& options, const Slice& key,
                       const Slice& value) {
312
    return Merge(options, DefaultColumnFamily(), key, value);
313
  }
314

J
jorlow@chromium.org 已提交
315
  // Apply the specified updates to the database.
316 317
  // If `updates` contains no update, WAL will still be synced if
  // options.sync=true.
J
jorlow@chromium.org 已提交
318
  // Returns OK on success, non-OK on failure.
319
  // Note: consider setting options.sync = true.
J
jorlow@chromium.org 已提交
320 321 322 323 324 325 326 327 328
  virtual Status Write(const WriteOptions& options, WriteBatch* updates) = 0;

  // If the database contains an entry for "key" store the
  // corresponding value in *value and return OK.
  //
  // If there is no entry for "key" leave *value unchanged and return
  // a status for which Status::IsNotFound() returns true.
  //
  // May return some other Status on an error.
329 330 331
  virtual inline Status Get(const ReadOptions& options,
                            ColumnFamilyHandle* column_family, const Slice& key,
                            std::string* value) {
M
Maysam Yabandeh 已提交
332 333 334 335
    assert(value != nullptr);
    PinnableSlice pinnable_val(value);
    assert(!pinnable_val.IsPinned());
    auto s = Get(options, column_family, key, &pinnable_val);
336
    if (s.ok() && pinnable_val.IsPinned()) {
M
Maysam Yabandeh 已提交
337 338 339 340
      value->assign(pinnable_val.data(), pinnable_val.size());
    }  // else value is already assigned
    return s;
  }
J
jorlow@chromium.org 已提交
341
  virtual Status Get(const ReadOptions& options,
342
                     ColumnFamilyHandle* column_family, const Slice& key,
M
Maysam Yabandeh 已提交
343
                     PinnableSlice* value) = 0;
344
  virtual Status Get(const ReadOptions& options, const Slice& key, std::string* value) {
345
    return Get(options, DefaultColumnFamily(), key, value);
346
  }
J
jorlow@chromium.org 已提交
347

348 349 350 351 352 353 354 355 356 357
  // If keys[i] does not exist in the database, then the i'th returned
  // status will be one for which Status::IsNotFound() is true, and
  // (*values)[i] will be set to some arbitrary value (often ""). Otherwise,
  // the i'th returned status will have Status::ok() true, and (*values)[i]
  // will store the value associated with keys[i].
  //
  // (*values) will always be resized to be the same size as (keys).
  // Similarly, the number of returned statuses will be the number of keys.
  // Note: keys will not be "de-duplicated". Duplicate keys will return
  // duplicate values in order.
358 359
  virtual std::vector<Status> MultiGet(
      const ReadOptions& options,
360
      const std::vector<ColumnFamilyHandle*>& column_family,
361
      const std::vector<Slice>& keys, std::vector<std::string>* values) = 0;
362 363 364
  virtual std::vector<Status> MultiGet(const ReadOptions& options,
                                       const std::vector<Slice>& keys,
                                       std::vector<std::string>* values) {
365 366
    return MultiGet(options, std::vector<ColumnFamilyHandle*>(
                                 keys.size(), DefaultColumnFamily()),
367 368
                    keys, values);
  }
369

370
  // If the key definitely does not exist in the database, then this method
371 372 373 374 375 376
  // returns false, else true. If the caller wants to obtain value when the key
  // is found in memory, a bool for 'value_found' must be passed. 'value_found'
  // will be true on return if value has been set properly.
  // This check is potentially lighter-weight than invoking DB::Get(). One way
  // to make this lighter weight is to avoid doing any IOs.
  // Default implementation here returns true and sets 'value_found' to false
A
Alex Loukissas 已提交
377 378 379 380
  virtual bool KeyMayExist(const ReadOptions& /*options*/,
                           ColumnFamilyHandle* /*column_family*/,
                           const Slice& /*key*/, std::string* /*value*/,
                           bool* value_found = nullptr) {
381 382 383
    if (value_found != nullptr) {
      *value_found = false;
    }
384 385
    return true;
  }
386 387
  virtual bool KeyMayExist(const ReadOptions& options, const Slice& key,
                           std::string* value, bool* value_found = nullptr) {
388
    return KeyMayExist(options, DefaultColumnFamily(), key, value, value_found);
389
  }
390

J
jorlow@chromium.org 已提交
391 392 393 394 395 396
  // Return a heap-allocated iterator over the contents of the database.
  // The result of NewIterator() is initially invalid (caller must
  // call one of the Seek methods on the iterator before using it).
  //
  // Caller should delete the iterator when it is no longer needed.
  // The returned iterator should be deleted before this db is deleted.
397
  virtual Iterator* NewIterator(const ReadOptions& options,
398
                                ColumnFamilyHandle* column_family) = 0;
399
  virtual Iterator* NewIterator(const ReadOptions& options) {
400
    return NewIterator(options, DefaultColumnFamily());
401 402 403 404 405 406
  }
  // Returns iterators from a consistent database state across multiple
  // column families. Iterators are heap allocated and need to be deleted
  // before the db is deleted
  virtual Status NewIterators(
      const ReadOptions& options,
I
Igor Canadi 已提交
407
      const std::vector<ColumnFamilyHandle*>& column_families,
408
      std::vector<Iterator*>* iterators) = 0;
J
jorlow@chromium.org 已提交
409 410 411 412 413

  // Return a handle to the current DB state.  Iterators created with
  // this handle will all observe a stable snapshot of the current DB
  // state.  The caller must call ReleaseSnapshot(result) when the
  // snapshot is no longer needed.
414 415 416
  //
  // nullptr will be returned if the DB fails to take a snapshot or does
  // not support snapshot.
J
jorlow@chromium.org 已提交
417 418 419 420 421 422
  virtual const Snapshot* GetSnapshot() = 0;

  // Release a previously acquired snapshot.  The caller must not
  // use "snapshot" after this call.
  virtual void ReleaseSnapshot(const Snapshot* snapshot) = 0;

423
#ifndef ROCKSDB_LITE
424
  // Contains all valid property arguments for GetProperty().
425 426 427
  //
  // NOTE: Property names cannot end in numbers since those are interpreted as
  //       arguments, e.g., see kNumFilesAtLevelPrefix.
428
  struct Properties {
429 430 431
    //  "rocksdb.num-files-at-level<N>" - returns string containing the number
    //      of files at level <N>, where <N> is an ASCII representation of a
    //      level number (e.g., "0").
432
    static const std::string kNumFilesAtLevelPrefix;
433

434 435 436 437 438 439 440
    //  "rocksdb.compression-ratio-at-level<N>" - returns string containing the
    //      compression ratio of data at level <N>, where <N> is an ASCII
    //      representation of a level number (e.g., "0"). Here, compression
    //      ratio is defined as uncompressed data size / compressed file size.
    //      Returns "-1.0" if no open files at level <N>.
    static const std::string kCompressionRatioAtLevelPrefix;

441 442
    //  "rocksdb.stats" - returns a multi-line string containing the data
    //      described by kCFStats followed by the data described by kDBStats.
443
    static const std::string kStats;
444 445 446

    //  "rocksdb.sstables" - returns a multi-line string summarizing current
    //      SST files.
447
    static const std::string kSSTables;
448

449 450 451 452 453 454 455 456 457
    //  "rocksdb.cfstats" - Both of "rocksdb.cfstats-no-file-histogram" and
    //      "rocksdb.cf-file-histogram" together. See below for description
    //      of the two.
    static const std::string kCFStats;

    //  "rocksdb.cfstats-no-file-histogram" - returns a multi-line string with
    //      general columm family stats per-level over db's lifetime ("L<n>"),
    //      aggregated over db's lifetime ("Sum"), and aggregated over the
    //      interval since the last retrieval ("Int").
458 459 460
    //  It could also be used to return the stats in the format of the map.
    //  In this case there will a pair of string to array of double for
    //  each level as well as for "Sum". "Int" stats will not be affected
Y
yizhu.sun 已提交
461
    //  when this form of stats are retrieved.
462 463 464 465 466
    static const std::string kCFStatsNoFileHistogram;

    //  "rocksdb.cf-file-histogram" - print out how many file reads to every
    //      level, as well as the histogram of latency of single requests.
    static const std::string kCFFileHistogram;
467 468 469 470

    //  "rocksdb.dbstats" - returns a multi-line string with general database
    //      stats, both cumulative (over the db's lifetime) and interval (since
    //      the last retrieval of kDBStats).
471
    static const std::string kDBStats;
472 473 474

    //  "rocksdb.levelstats" - returns multi-line string containing the number
    //      of files per level and total size of each level (MB).
475
    static const std::string kLevelStats;
476 477 478

    //  "rocksdb.num-immutable-mem-table" - returns number of immutable
    //      memtables that have not yet been flushed.
479
    static const std::string kNumImmutableMemTable;
480 481 482

    //  "rocksdb.num-immutable-mem-table-flushed" - returns number of immutable
    //      memtables that have already been flushed.
483
    static const std::string kNumImmutableMemTableFlushed;
484 485 486

    //  "rocksdb.mem-table-flush-pending" - returns 1 if a memtable flush is
    //      pending; otherwise, returns 0.
487
    static const std::string kMemTableFlushPending;
488 489 490

    //  "rocksdb.num-running-flushes" - returns the number of currently running
    //      flushes.
491
    static const std::string kNumRunningFlushes;
492 493 494

    //  "rocksdb.compaction-pending" - returns 1 if at least one compaction is
    //      pending; otherwise, returns 0.
495
    static const std::string kCompactionPending;
496 497 498

    //  "rocksdb.num-running-compactions" - returns the number of currently
    //      running compactions.
499
    static const std::string kNumRunningCompactions;
500 501 502

    //  "rocksdb.background-errors" - returns accumulated number of background
    //      errors.
503
    static const std::string kBackgroundErrors;
504 505 506

    //  "rocksdb.cur-size-active-mem-table" - returns approximate size of active
    //      memtable (bytes).
507
    static const std::string kCurSizeActiveMemTable;
508 509 510

    //  "rocksdb.cur-size-all-mem-tables" - returns approximate size of active
    //      and unflushed immutable memtables (bytes).
511
    static const std::string kCurSizeAllMemTables;
512 513 514

    //  "rocksdb.size-all-mem-tables" - returns approximate size of active,
    //      unflushed immutable, and pinned immutable memtables (bytes).
515
    static const std::string kSizeAllMemTables;
516 517 518

    //  "rocksdb.num-entries-active-mem-table" - returns total number of entries
    //      in the active memtable.
519
    static const std::string kNumEntriesActiveMemTable;
520 521 522

    //  "rocksdb.num-entries-imm-mem-tables" - returns total number of entries
    //      in the unflushed immutable memtables.
523
    static const std::string kNumEntriesImmMemTables;
524 525 526

    //  "rocksdb.num-deletes-active-mem-table" - returns total number of delete
    //      entries in the active memtable.
527
    static const std::string kNumDeletesActiveMemTable;
528 529 530

    //  "rocksdb.num-deletes-imm-mem-tables" - returns total number of delete
    //      entries in the unflushed immutable memtables.
531
    static const std::string kNumDeletesImmMemTables;
532 533

    //  "rocksdb.estimate-num-keys" - returns estimated number of total keys in
O
oranagra 已提交
534
    //      the active and unflushed immutable memtables and storage.
535
    static const std::string kEstimateNumKeys;
536 537 538 539

    //  "rocksdb.estimate-table-readers-mem" - returns estimated memory used for
    //      reading SST tables, excluding memory used in block cache (e.g.,
    //      filter and index blocks).
540
    static const std::string kEstimateTableReadersMem;
541 542 543

    //  "rocksdb.is-file-deletions-enabled" - returns 0 if deletion of obsolete
    //      files is enabled; otherwise, returns a non-zero number.
544
    static const std::string kIsFileDeletionsEnabled;
545 546 547

    //  "rocksdb.num-snapshots" - returns number of unreleased snapshots of the
    //      database.
548
    static const std::string kNumSnapshots;
549 550 551

    //  "rocksdb.oldest-snapshot-time" - returns number representing unix
    //      timestamp of oldest unreleased snapshot.
552
    static const std::string kOldestSnapshotTime;
553 554 555 556 557

    //  "rocksdb.num-live-versions" - returns number of live versions. `Version`
    //      is an internal data structure. See version_set.h for details. More
    //      live versions often mean more SST files are held from being deleted,
    //      by iterators or unfinished compactions.
558
    static const std::string kNumLiveVersions;
559

Y
yizhu.sun 已提交
560
    //  "rocksdb.current-super-version-number" - returns number of current LSM
561 562 563 564
    //  version. It is a uint64_t integer number, incremented after there is
    //  any change to the LSM tree. The number is not preserved after restarting
    //  the DB. After DB restart, it will start from 0 again.
    static const std::string kCurrentSuperVersionNumber;
565

566 567
    //  "rocksdb.estimate-live-data-size" - returns an estimate of the amount of
    //      live data in bytes.
A
Andres Notzli 已提交
568
    static const std::string kEstimateLiveDataSize;
569

Y
yizhu.sun 已提交
570
    //  "rocksdb.min-log-number-to-keep" - return the minimum log number of the
571 572 573
    //      log files that should be kept.
    static const std::string kMinLogNumberToKeep;

574 575 576
    //  "rocksdb.total-sst-files-size" - returns total size (bytes) of all SST
    //      files.
    //  WARNING: may slow down online queries if there are too many files.
577
    static const std::string kTotalSstFilesSize;
578

579 580 581 582
    //  "rocksdb.live-sst-files-size" - returns total size (bytes) of all SST
    //      files belong to the latest LSM tree.
    static const std::string kLiveSstFilesSize;

583 584 585 586 587 588 589 590
    //  "rocksdb.base-level" - returns number of level to which L0 data will be
    //      compacted.
    static const std::string kBaseLevel;

    //  "rocksdb.estimate-pending-compaction-bytes" - returns estimated total
    //      number of bytes compaction needs to rewrite to get all levels down
    //      to under target size. Not valid for other compactions than level-
    //      based.
591
    static const std::string kEstimatePendingCompactionBytes;
592 593 594

    //  "rocksdb.aggregated-table-properties" - returns a string representation
    //      of the aggregated table properties of the target column family.
595
    static const std::string kAggregatedTableProperties;
596 597 598 599

    //  "rocksdb.aggregated-table-properties-at-level<N>", same as the previous
    //      one but only returns the aggregated table properties of the
    //      specified level "N" at the target column family.
600
    static const std::string kAggregatedTablePropertiesAtLevel;
601 602 603 604 605 606 607

    //  "rocksdb.actual-delayed-write-rate" - returns the current actual delayed
    //      write rate. 0 means no delay.
    static const std::string kActualDelayedWriteRate;

    //  "rocksdb.is-write-stopped" - Return 1 if write has been stopped.
    static const std::string kIsWriteStopped;
Y
Yi Wu 已提交
608 609 610 611 612 613

    //  "rocksdb.estimate-oldest-key-time" - returns an estimation of
    //      oldest key timestamp in the DB. Currently only available for
    //      FIFO compaction with
    //      compaction_options_fifo.allow_compaction = false.
    static const std::string kEstimateOldestKeyTime;
614 615 616
  };
#endif /* ROCKSDB_LITE */

617 618 619 620
  // DB implementations can export properties about their state via this method.
  // If "property" is a valid property understood by this DB implementation (see
  // Properties struct above for valid options), fills "*value" with its current
  // value and returns true.  Otherwise, returns false.
621
  virtual bool GetProperty(ColumnFamilyHandle* column_family,
622
                           const Slice& property, std::string* value) = 0;
623
  virtual bool GetProperty(const Slice& property, std::string* value) {
624
    return GetProperty(DefaultColumnFamily(), property, value);
625
  }
626 627
  virtual bool GetMapProperty(ColumnFamilyHandle* column_family,
                              const Slice& property,
628
                              std::map<std::string, std::string>* value) = 0;
629
  virtual bool GetMapProperty(const Slice& property,
630
                              std::map<std::string, std::string>* value) {
631 632
    return GetMapProperty(DefaultColumnFamily(), property, value);
  }
J
jorlow@chromium.org 已提交
633

634
  // Similar to GetProperty(), but only works for a subset of properties whose
635 636 637 638 639 640 641 642
  // return value is an integer. Return the value by integer. Supported
  // properties:
  //  "rocksdb.num-immutable-mem-table"
  //  "rocksdb.mem-table-flush-pending"
  //  "rocksdb.compaction-pending"
  //  "rocksdb.background-errors"
  //  "rocksdb.cur-size-active-mem-table"
  //  "rocksdb.cur-size-all-mem-tables"
643
  //  "rocksdb.size-all-mem-tables"
644 645
  //  "rocksdb.num-entries-active-mem-table"
  //  "rocksdb.num-entries-imm-mem-tables"
646 647
  //  "rocksdb.num-deletes-active-mem-table"
  //  "rocksdb.num-deletes-imm-mem-tables"
648 649 650 651 652
  //  "rocksdb.estimate-num-keys"
  //  "rocksdb.estimate-table-readers-mem"
  //  "rocksdb.is-file-deletions-enabled"
  //  "rocksdb.num-snapshots"
  //  "rocksdb.oldest-snapshot-time"
653
  //  "rocksdb.num-live-versions"
654
  //  "rocksdb.current-super-version-number"
A
Andres Notzli 已提交
655
  //  "rocksdb.estimate-live-data-size"
656
  //  "rocksdb.min-log-number-to-keep"
657
  //  "rocksdb.total-sst-files-size"
658
  //  "rocksdb.live-sst-files-size"
659 660
  //  "rocksdb.base-level"
  //  "rocksdb.estimate-pending-compaction-bytes"
661 662
  //  "rocksdb.num-running-compactions"
  //  "rocksdb.num-running-flushes"
663 664
  //  "rocksdb.actual-delayed-write-rate"
  //  "rocksdb.is-write-stopped"
Y
Yi Wu 已提交
665
  //  "rocksdb.estimate-oldest-key-time"
666 667 668 669 670 671
  virtual bool GetIntProperty(ColumnFamilyHandle* column_family,
                              const Slice& property, uint64_t* value) = 0;
  virtual bool GetIntProperty(const Slice& property, uint64_t* value) {
    return GetIntProperty(DefaultColumnFamily(), property, value);
  }

S
Siying Dong 已提交
672 673 674 675 676 677 678
  // Reset internal stats for DB and all column families.
  // Note this doesn't reset options.statistics as it is not owned by
  // DB.
  virtual Status ResetStats() {
    return Status::NotSupported("Not implemented");
  }

679 680 681 682 683
  // Same as GetIntProperty(), but this one returns the aggregated int
  // property from all column families.
  virtual bool GetAggregatedIntProperty(const Slice& property,
                                        uint64_t* value) = 0;

684
  // Flags for DB::GetSizeApproximation that specify whether memtable
685 686 687 688 689 690 691
  // stats should be included, or file stats approximation or both
  enum SizeApproximationFlags : uint8_t {
    NONE = 0,
    INCLUDE_MEMTABLES = 1,
    INCLUDE_FILES = 1 << 1
  };

J
jorlow@chromium.org 已提交
692 693 694 695 696 697 698
  // For each i in [0,n-1], store in "sizes[i]", the approximate
  // file system space used by keys in "[range[i].start .. range[i].limit)".
  //
  // Note that the returned sizes measure file system space usage, so
  // if the user data compresses by a factor of ten, the returned
  // sizes will be one-tenth the size of the corresponding user data size.
  //
699 700 701 702
  // If include_flags defines whether the returned size should include
  // the recently written data in the mem-tables (if
  // the mem-table type supports it), data serialized to disk, or both.
  // include_flags should be of type DB::SizeApproximationFlags
703
  virtual void GetApproximateSizes(ColumnFamilyHandle* column_family,
704
                                   const Range* range, int n, uint64_t* sizes,
705 706
                                   uint8_t include_flags
                                   = INCLUDE_FILES) = 0;
707
  virtual void GetApproximateSizes(const Range* range, int n, uint64_t* sizes,
708 709
                                   uint8_t include_flags
                                   = INCLUDE_FILES) {
710
    GetApproximateSizes(DefaultColumnFamily(), range, n, sizes,
711 712 713
                        include_flags);
  }

714 715 716 717 718 719 720 721 722 723 724 725
  // The method is similar to GetApproximateSizes, except it
  // returns approximate number of records in memtables.
  virtual void GetApproximateMemTableStats(ColumnFamilyHandle* column_family,
                                           const Range& range,
                                           uint64_t* const count,
                                           uint64_t* const size) = 0;
  virtual void GetApproximateMemTableStats(const Range& range,
                                           uint64_t* const count,
                                           uint64_t* const size) {
    GetApproximateMemTableStats(DefaultColumnFamily(), range, count, size);
  }

726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744
  // Deprecated versions of GetApproximateSizes
  ROCKSDB_DEPRECATED_FUNC virtual void GetApproximateSizes(
      const Range* range, int n, uint64_t* sizes,
      bool include_memtable) {
    uint8_t include_flags = SizeApproximationFlags::INCLUDE_FILES;
    if (include_memtable) {
      include_flags |= SizeApproximationFlags::INCLUDE_MEMTABLES;
    }
    GetApproximateSizes(DefaultColumnFamily(), range, n, sizes, include_flags);
  }
  ROCKSDB_DEPRECATED_FUNC virtual void GetApproximateSizes(
      ColumnFamilyHandle* column_family,
      const Range* range, int n, uint64_t* sizes,
      bool include_memtable) {
    uint8_t include_flags = SizeApproximationFlags::INCLUDE_FILES;
    if (include_memtable) {
      include_flags |= SizeApproximationFlags::INCLUDE_MEMTABLES;
    }
    GetApproximateSizes(column_family, range, n, sizes, include_flags);
745
  }
J
jorlow@chromium.org 已提交
746

G
Gabor Cselle 已提交
747
  // Compact the underlying storage for the key range [*begin,*end].
748
  // The actual compaction interval might be superset of [*begin, *end].
G
Gabor Cselle 已提交
749 750 751 752 753
  // In particular, deleted and overwritten versions are discarded,
  // and the data is rearranged to reduce the cost of operations
  // needed to access the data.  This operation should typically only
  // be invoked by users who understand the underlying implementation.
  //
A
Abhishek Kona 已提交
754 755
  // begin==nullptr is treated as a key before all keys in the database.
  // end==nullptr is treated as a key after all keys in the database.
G
Gabor Cselle 已提交
756
  // Therefore the following call will compact the entire database:
757
  //    db->CompactRange(options, nullptr, nullptr);
758
  // Note that after the entire database is compacted, all data are pushed
759 760 761 762 763 764 765 766 767 768 769
  // down to the last level containing any data. If the total data size after
  // compaction is reduced, that level might not be appropriate for hosting all
  // the files. In this case, client could set options.change_level to true, to
  // move the files back to the minimum level capable of holding the data set
  // or a given level (specified by non-negative options.target_level).
  virtual Status CompactRange(const CompactRangeOptions& options,
                              ColumnFamilyHandle* column_family,
                              const Slice* begin, const Slice* end) = 0;
  virtual Status CompactRange(const CompactRangeOptions& options,
                              const Slice* begin, const Slice* end) {
    return CompactRange(options, DefaultColumnFamily(), begin, end);
770
  }
771

772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
  ROCKSDB_DEPRECATED_FUNC virtual Status CompactRange(
      ColumnFamilyHandle* column_family, const Slice* begin, const Slice* end,
      bool change_level = false, int target_level = -1,
      uint32_t target_path_id = 0) {
    CompactRangeOptions options;
    options.change_level = change_level;
    options.target_level = target_level;
    options.target_path_id = target_path_id;
    return CompactRange(options, column_family, begin, end);
  }

  ROCKSDB_DEPRECATED_FUNC virtual Status CompactRange(
      const Slice* begin, const Slice* end, bool change_level = false,
      int target_level = -1, uint32_t target_path_id = 0) {
    CompactRangeOptions options;
    options.change_level = change_level;
    options.target_level = target_level;
    options.target_path_id = target_path_id;
    return CompactRange(options, DefaultColumnFamily(), begin, end);
  }

S
sdong 已提交
793 794
  virtual Status SetOptions(
      ColumnFamilyHandle* /*column_family*/,
A
Alex Loukissas 已提交
795
      const std::unordered_map<std::string, std::string>& /*new_options*/) {
796
    return Status::NotSupported("Not implemented");
797
  }
798
  virtual Status SetOptions(
799 800 801
      const std::unordered_map<std::string, std::string>& new_options) {
    return SetOptions(DefaultColumnFamily(), new_options);
  }
J
jorlow@chromium.org 已提交
802

803 804 805
  virtual Status SetDBOptions(
      const std::unordered_map<std::string, std::string>& new_options) = 0;

A
Andres Notzli 已提交
806 807 808 809
  // CompactFiles() inputs a list of files specified by file numbers and
  // compacts them to the specified level. Note that the behavior is different
  // from CompactRange() in that CompactFiles() performs the compaction job
  // using the CURRENT thread.
810 811 812 813 814 815 816
  //
  // @see GetDataBaseMetaData
  // @see GetColumnFamilyMetaData
  virtual Status CompactFiles(
      const CompactionOptions& compact_options,
      ColumnFamilyHandle* column_family,
      const std::vector<std::string>& input_file_names,
817 818
      const int output_level, const int output_path_id = -1,
      std::vector<std::string>* const output_file_names = nullptr) = 0;
819 820 821 822

  virtual Status CompactFiles(
      const CompactionOptions& compact_options,
      const std::vector<std::string>& input_file_names,
823 824
      const int output_level, const int output_path_id = -1,
      std::vector<std::string>* const output_file_names = nullptr) {
825
    return CompactFiles(compact_options, DefaultColumnFamily(),
826 827
                        input_file_names, output_level, output_path_id,
                        output_file_names);
828
  }
829 830 831

  // This function will wait until all currently running background processes
  // finish. After it returns, no background process will be run until
S
Sagar Vemuri 已提交
832
  // ContinueBackgroundWork is called
833 834 835
  virtual Status PauseBackgroundWork() = 0;
  virtual Status ContinueBackgroundWork() = 0;

836
  // This function will enable automatic compactions for the given column
837 838 839 840 841 842 843 844
  // families if they were previously disabled. The function will first set the
  // disable_auto_compactions option for each column family to 'false', after
  // which it will schedule a flush/compaction.
  //
  // NOTE: Setting disable_auto_compactions to 'false' through SetOptions() API
  // does NOT schedule a flush/compaction afterwards, and only changes the
  // parameter itself within the column family option.
  //
845 846 847
  virtual Status EnableAutoCompaction(
      const std::vector<ColumnFamilyHandle*>& column_family_handles) = 0;

848
  // Number of levels used for this DB.
849
  virtual int NumberLevels(ColumnFamilyHandle* column_family) = 0;
850
  virtual int NumberLevels() { return NumberLevels(DefaultColumnFamily()); }
851 852 853

  // Maximum level to which a new compacted memtable is pushed if it
  // does not create overlap.
854
  virtual int MaxMemCompactionLevel(ColumnFamilyHandle* column_family) = 0;
855
  virtual int MaxMemCompactionLevel() {
856
    return MaxMemCompactionLevel(DefaultColumnFamily());
857
  }
858 859

  // Number of files in level-0 that would stop writes.
860
  virtual int Level0StopWriteTrigger(ColumnFamilyHandle* column_family) = 0;
861
  virtual int Level0StopWriteTrigger() {
862
    return Level0StopWriteTrigger(DefaultColumnFamily());
863
  }
864

I
Igor Canadi 已提交
865 866 867 868
  // Get DB name -- the exact same name that was provided as an argument to
  // DB::Open()
  virtual const std::string& GetName() const = 0;

869 870 871
  // Get Env object from the DB
  virtual Env* GetEnv() const = 0;

872 873 874 875
  // Get DB Options that we use.  During the process of opening the
  // column family, the options provided when calling DB::Open() or
  // DB::CreateColumnFamily() will have been "sanitized" and transformed
  // in an implementation-defined manner.
876 877
  virtual Options GetOptions(ColumnFamilyHandle* column_family) const = 0;
  virtual Options GetOptions() const {
878
    return GetOptions(DefaultColumnFamily());
879
  }
I
Igor Canadi 已提交
880

881
  virtual DBOptions GetDBOptions() const = 0;
882

H
heyongqiang 已提交
883
  // Flush all mem-table data.
884
  virtual Status Flush(const FlushOptions& options,
885
                       ColumnFamilyHandle* column_family) = 0;
886
  virtual Status Flush(const FlushOptions& options) {
887
    return Flush(options, DefaultColumnFamily());
888
  }
H
heyongqiang 已提交
889

890 891
  // Flush the WAL memory buffer to the file. If sync is true, it calls SyncWAL
  // afterwards.
A
Andrew Kryczka 已提交
892
  virtual Status FlushWAL(bool /*sync*/) {
893 894
    return Status::NotSupported("FlushWAL not implemented");
  }
895 896 897 898 899 900
  // Sync the wal. Note that Write() followed by SyncWAL() is not exactly the
  // same as Write() with sync=true: in the latter case the changes won't be
  // visible until the sync is done.
  // Currently only works if allow_mmap_writes = false in Options.
  virtual Status SyncWAL() = 0;

I
Igor Canadi 已提交
901 902 903
  // The sequence number of the most recent transaction.
  virtual SequenceNumber GetLatestSequenceNumber() const = 0;

904 905 906 907 908 909 910 911
  // Instructs DB to preserve deletes with sequence numbers >= passed seqnum.
  // Has no effect if DBOptions.preserve_deletes is set to false.
  // This function assumes that user calls this function with monotonically
  // increasing seqnums (otherwise we can't guarantee that a particular delete
  // hasn't been already processed); returns true if the value was successfully
  // updated, false if user attempted to call if with seqnum <= current value.
  virtual bool SetPreserveDeletesSequenceNumber(SequenceNumber seqnum) = 0;

I
Igor Canadi 已提交
912 913
#ifndef ROCKSDB_LITE

914 915 916 917 918
  // Prevent file deletions. Compactions will continue to occur,
  // but no obsolete files will be deleted. Calling this multiple
  // times have the same effect as calling it once.
  virtual Status DisableFileDeletions() = 0;

919
  // Allow compactions to delete obsolete files.
920 921 922 923 924 925 926 927 928
  // If force == true, the call to EnableFileDeletions() will guarantee that
  // file deletions are enabled after the call, even if DisableFileDeletions()
  // was called multiple times before.
  // If force == false, EnableFileDeletions will only enable file deletion
  // after it's been called at least as many times as DisableFileDeletions(),
  // enabling the two methods to be called by two threads concurrently without
  // synchronization -- i.e., file deletions will be enabled only after both
  // threads call EnableFileDeletions()
  virtual Status EnableFileDeletions(bool force = true) = 0;
929

930 931
  // GetLiveFiles followed by GetSortedWalFiles can generate a lossless backup

932
  // Retrieve the list of all files in the database. The files are
933 934 935 936
  // relative to the dbname and are not absolute paths. The valid size of the
  // manifest file is returned in manifest_file_size. The manifest file is an
  // ever growing file, but only the portion specified by manifest_file_size is
  // valid for this snapshot.
937 938 939
  // Setting flush_memtable to true does Flush before recording the live files.
  // Setting flush_memtable to false is useful when we don't want to wait for
  // flush which may have to wait for compaction to complete taking an
940 941 942 943 944 945
  // indeterminate time.
  //
  // In case you have multiple column families, even if flush_memtable is true,
  // you still need to call GetSortedWalFiles after GetLiveFiles to compensate
  // for new data that arrived to already-flushed column families while other
  // column families were flushing
946
  virtual Status GetLiveFiles(std::vector<std::string>&,
947 948
                              uint64_t* manifest_file_size,
                              bool flush_memtable = true) = 0;
949

950 951 952
  // Retrieve the sorted list of all wal files with earliest file first
  virtual Status GetSortedWalFiles(VectorLogPtr& files) = 0;

953
  // Note: this API is not yet consistent with WritePrepared transactions.
954 955 956
  // Sets iter to an iterator that is positioned at a write-batch containing
  // seq_number. If the sequence number is non existent, it returns an iterator
  // at the first available seq_no after the requested seq_no
957
  // Returns Status::OK if iterator is valid
958 959 960 961
  // Must set WAL_ttl_seconds or WAL_size_limit_MB to large values to
  // use this api, else the WAL files will get
  // cleared aggressively and the iterator might keep getting invalid before
  // an update is read.
962 963 964 965
  virtual Status GetUpdatesSince(
      SequenceNumber seq_number, unique_ptr<TransactionLogIterator>* iter,
      const TransactionLogIterator::ReadOptions&
          read_options = TransactionLogIterator::ReadOptions()) = 0;
966

D
Dmitri Smirnov 已提交
967 968
// Windows API macro interference
#undef DeleteFile
969 970 971 972
  // Delete the file name from the db directory and update the internal state to
  // reflect that. Supports deletion of sst and log files only. 'name' must be
  // path relative to the db directory. eg. 000001.sst, /archive/000003.log
  virtual Status DeleteFile(std::string name) = 0;
973 974 975

  // Returns a list of all table files with their level, start key
  // and end key
A
Alex Loukissas 已提交
976
  virtual void GetLiveFilesMetaData(
S
sdong 已提交
977
      std::vector<LiveFileMetaData>* /*metadata*/) {}
978

979 980 981 982 983 984
  // Obtains the meta data of the specified column family of the DB.
  // Status::NotFound() will be returned if the current DB does not have
  // any column family match the specified name.
  //
  // If cf_name is not specified, then the metadata of the default
  // column family will be returned.
S
sdong 已提交
985 986
  virtual void GetColumnFamilyMetaData(ColumnFamilyHandle* /*column_family*/,
                                       ColumnFamilyMetaData* /*metadata*/) {}
987 988

  // Get the metadata of the default column family.
989
  void GetColumnFamilyMetaData(
990 991 992
      ColumnFamilyMetaData* metadata) {
    GetColumnFamilyMetaData(DefaultColumnFamily(), metadata);
  }
993

994
  // IngestExternalFile() will load a list of external SST files (1) into the DB
995 996 997 998 999 1000 1001
  // Two primary modes are supported:
  // - Duplicate keys in the new files will overwrite exiting keys (default)
  // - Duplicate keys will be skipped (set ingest_behind=true)
  // In the first mode we will try to find the lowest possible level that
  // the file can fit in, and ingest the file into this level (2). A file that
  // have a key range that overlap with the memtable key range will require us
  // to Flush the memtable first before ingesting the file.
1002
  // In the second mode we will always ingest in the bottom most level (see
1003
  // docs to IngestExternalFileOptions::ingest_behind).
1004
  //
1005 1006
  // (1) External SST files can be created using SstFileWriter
  // (2) We will try to ingest the files to the lowest possible level
Y
yizhu.sun 已提交
1007
  //     even if the file compression doesn't match the level compression
1008 1009 1010
  // (3) If IngestExternalFileOptions->ingest_behind is set to true,
  //     we always ingest at the bottommost level, which should be reserved
  //     for this purpose (see DBOPtions::allow_ingest_behind flag).
1011 1012 1013 1014 1015 1016 1017 1018 1019
  virtual Status IngestExternalFile(
      ColumnFamilyHandle* column_family,
      const std::vector<std::string>& external_files,
      const IngestExternalFileOptions& options) = 0;

  virtual Status IngestExternalFile(
      const std::vector<std::string>& external_files,
      const IngestExternalFileOptions& options) {
    return IngestExternalFile(DefaultColumnFamily(), external_files, options);
1020
  }
1021

A
Aaron G 已提交
1022 1023
  virtual Status VerifyChecksum() = 0;

1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
  // AddFile() is deprecated, please use IngestExternalFile()
  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      ColumnFamilyHandle* column_family,
      const std::vector<std::string>& file_path_list, bool move_file = false,
      bool skip_snapshot_check = false) {
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(column_family, file_path_list, ifo);
  }

  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      const std::vector<std::string>& file_path_list, bool move_file = false,
      bool skip_snapshot_check = false) {
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(DefaultColumnFamily(), file_path_list, ifo);
  }

  // AddFile() is deprecated, please use IngestExternalFile()
  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      ColumnFamilyHandle* column_family, const std::string& file_path,
      bool move_file = false, bool skip_snapshot_check = false) {
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(column_family, {file_path}, ifo);
  }

  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      const std::string& file_path, bool move_file = false,
      bool skip_snapshot_check = false) {
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(DefaultColumnFamily(), {file_path}, ifo);
  }

  // Load table file with information "file_info" into "column_family"
  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      ColumnFamilyHandle* column_family,
      const std::vector<ExternalSstFileInfo>& file_info_list,
      bool move_file = false, bool skip_snapshot_check = false) {
    std::vector<std::string> external_files;
    for (const ExternalSstFileInfo& file_info : file_info_list) {
      external_files.push_back(file_info.file_path);
    }
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(column_family, external_files, ifo);
  }

  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      const std::vector<ExternalSstFileInfo>& file_info_list,
      bool move_file = false, bool skip_snapshot_check = false) {
    std::vector<std::string> external_files;
    for (const ExternalSstFileInfo& file_info : file_info_list) {
      external_files.push_back(file_info.file_path);
    }
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(DefaultColumnFamily(), external_files, ifo);
  }

  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      ColumnFamilyHandle* column_family, const ExternalSstFileInfo* file_info,
      bool move_file = false, bool skip_snapshot_check = false) {
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(column_family, {file_info->file_path}, ifo);
  }

  ROCKSDB_DEPRECATED_FUNC virtual Status AddFile(
      const ExternalSstFileInfo* file_info, bool move_file = false,
      bool skip_snapshot_check = false) {
    IngestExternalFileOptions ifo;
    ifo.move_files = move_file;
    ifo.snapshot_consistency = !skip_snapshot_check;
    ifo.allow_global_seqno = false;
    ifo.allow_blocking_flush = false;
    return IngestExternalFile(DefaultColumnFamily(), {file_info->file_path},
                              ifo);
  }

I
Igor Canadi 已提交
1126 1127
#endif  // ROCKSDB_LITE

1128 1129 1130
  // Sets the globally unique ID created at database creation time by invoking
  // Env::GenerateUniqueId(), in identity. Returns Status::OK if identity could
  // be set properly
1131
  virtual Status GetDbIdentity(std::string& identity) const = 0;
1132

1133 1134 1135
  // Returns default column family handle
  virtual ColumnFamilyHandle* DefaultColumnFamily() const = 0;

I
Igor Canadi 已提交
1136
#ifndef ROCKSDB_LITE
I
Igor Canadi 已提交
1137 1138
  virtual Status GetPropertiesOfAllTables(ColumnFamilyHandle* column_family,
                                          TablePropertiesCollection* props) = 0;
1139
  virtual Status GetPropertiesOfAllTables(TablePropertiesCollection* props) {
I
Igor Canadi 已提交
1140 1141
    return GetPropertiesOfAllTables(DefaultColumnFamily(), props);
  }
1142
  virtual Status GetPropertiesOfTablesInRange(
1143
      ColumnFamilyHandle* column_family, const Range* range, std::size_t n,
1144
      TablePropertiesCollection* props) = 0;
S
Siying Dong 已提交
1145

A
Andrew Kryczka 已提交
1146 1147 1148
  virtual Status SuggestCompactRange(ColumnFamilyHandle* /*column_family*/,
                                     const Slice* /*begin*/,
                                     const Slice* /*end*/) {
S
Siying Dong 已提交
1149 1150 1151
    return Status::NotSupported("SuggestCompactRange() is not implemented.");
  }

A
Andrew Kryczka 已提交
1152 1153
  virtual Status PromoteL0(ColumnFamilyHandle* /*column_family*/,
                           int /*target_level*/) {
S
Siying Dong 已提交
1154 1155 1156
    return Status::NotSupported("PromoteL0() is not implemented.");
  }

I
Igor Canadi 已提交
1157
#endif  // ROCKSDB_LITE
1158

A
agiardullo 已提交
1159 1160 1161
  // Needed for StackableDB
  virtual DB* GetRootDB() { return this; }

J
jorlow@chromium.org 已提交
1162 1163 1164 1165 1166 1167 1168 1169
 private:
  // No copying allowed
  DB(const DB&);
  void operator=(const DB&);
};

// Destroy the contents of the specified database.
// Be very careful using this method.
1170 1171 1172
Status DestroyDB(const std::string& name, const Options& options,
                 const std::vector<ColumnFamilyDescriptor>& column_families =
                   std::vector<ColumnFamilyDescriptor>());
J
jorlow@chromium.org 已提交
1173

I
Igor Canadi 已提交
1174
#ifndef ROCKSDB_LITE
J
jorlow@chromium.org 已提交
1175 1176 1177 1178
// If a DB cannot be opened, you may attempt to call this method to
// resurrect as much of the contents of the database as possible.
// Some data may be lost, so be careful when calling this function
// on a database that contains important information.
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194
//
// With this API, we will warn and skip data associated with column families not
// specified in column_families.
//
// @param column_families Descriptors for known column families
Status RepairDB(const std::string& dbname, const DBOptions& db_options,
                const std::vector<ColumnFamilyDescriptor>& column_families);

// @param unknown_cf_opts Options for column families encountered during the
//                        repair that were not specified in column_families.
Status RepairDB(const std::string& dbname, const DBOptions& db_options,
                const std::vector<ColumnFamilyDescriptor>& column_families,
                const ColumnFamilyOptions& unknown_cf_opts);

// @param options These options will be used for the database and for ALL column
//                families encountered during the repair
J
jorlow@chromium.org 已提交
1195
Status RepairDB(const std::string& dbname, const Options& options);
1196

I
Igor Canadi 已提交
1197
#endif
J
jorlow@chromium.org 已提交
1198

1199 1200
}  // namespace rocksdb

1201
#endif  // STORAGE_ROCKSDB_INCLUDE_DB_H_