1
2
3
4
5
6
7
8
9
10
11
12
13
14
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
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
|
// repair_database.cpp
/**
* Copyright (C) 2014 MongoDB Inc.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License, version 3,
* as published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* As a special exception, the copyright holders give permission to link the
* code of portions of this program with the OpenSSL library under certain
* conditions as described in each individual source file and distribute
* linked combinations including the program with the OpenSSL library. You
* must comply with the GNU Affero General Public License in all respects for
* all of the code used other than as permitted herein. If you modify file(s)
* with this exception, you may extend this exception to your version of the
* file(s), but you are not obligated to do so. If you do not wish to do so,
* delete this exception statement from your version. If you delete this
* exception statement from all source files in the program, then also delete
* it in the license file.
*/
#define MONGO_LOG_DEFAULT_COMPONENT ::mongo::logger::LogComponent::kStorage
#include "mongo/platform/basic.h"
#include "mongo/db/storage/mmap_v1/mmap_v1_engine.h"
#include <boost/filesystem/operations.hpp>
#include "mongo/db/background.h"
#include "mongo/db/catalog/collection.h"
#include "mongo/db/catalog/database.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/catalog/index_create.h"
#include "mongo/db/client.h"
#include "mongo/db/db_raii.h"
#include "mongo/db/index/index_descriptor.h"
#include "mongo/db/storage/mmap_v1/dur.h"
#include "mongo/db/storage/mmap_v1/file_allocator.h"
#include "mongo/db/storage/mmap_v1/mmap.h"
#include "mongo/db/storage/mmap_v1/mmap_v1_database_catalog_entry.h"
#include "mongo/db/storage/mmap_v1/mmap_v1_options.h"
#include "mongo/util/file.h"
#include "mongo/util/log.h"
#include "mongo/util/scopeguard.h"
namespace mongo {
using std::unique_ptr;
using std::endl;
using std::map;
using std::string;
using std::stringstream;
using std::vector;
typedef boost::filesystem::path Path;
// inheritable class to implement an operation that may be applied to all
// files in a database using _applyOpToDataFiles()
class FileOp {
public:
virtual ~FileOp() {}
// Return true if file exists and operation successful
virtual bool apply(const boost::filesystem::path& p) = 0;
virtual const char* op() const = 0;
};
void _applyOpToDataFiles(const string& database,
FileOp& fo,
bool afterAllocator = false,
const string& path = storageGlobalParams.dbpath);
void _deleteDataFiles(const std::string& database) {
if (storageGlobalParams.directoryperdb) {
FileAllocator::get()->waitUntilFinished();
MONGO_ASSERT_ON_EXCEPTION_WITH_MSG(
boost::filesystem::remove_all(boost::filesystem::path(storageGlobalParams.dbpath) /
database),
"delete data files with a directoryperdb");
return;
}
class : public FileOp {
virtual bool apply(const boost::filesystem::path& p) {
return boost::filesystem::remove(p);
}
virtual const char* op() const {
return "remove";
}
} deleter;
_applyOpToDataFiles(database, deleter, true);
}
void boostRenameWrapper(const Path& from, const Path& to) {
try {
boost::filesystem::rename(from, to);
} catch (const boost::filesystem::filesystem_error&) {
// boost rename doesn't work across partitions
boost::filesystem::copy_file(from, to);
boost::filesystem::remove(from);
}
}
// back up original database files to 'temp' dir
void _renameForBackup(const std::string& database, const Path& reservedPath) {
Path newPath(reservedPath);
if (storageGlobalParams.directoryperdb)
newPath /= database;
class Renamer : public FileOp {
public:
Renamer(const Path& newPath) : newPath_(newPath) {}
private:
const boost::filesystem::path& newPath_;
virtual bool apply(const Path& p) {
if (!boost::filesystem::exists(p))
return false;
boostRenameWrapper(p, newPath_ / (p.leaf().string() + ".bak"));
return true;
}
virtual const char* op() const {
return "renaming";
}
} renamer(newPath);
_applyOpToDataFiles(database, renamer, true);
}
intmax_t dbSize(const string& database) {
class SizeAccumulator : public FileOp {
public:
SizeAccumulator() : totalSize_(0) {}
intmax_t size() const {
return totalSize_;
}
private:
virtual bool apply(const boost::filesystem::path& p) {
if (!boost::filesystem::exists(p))
return false;
totalSize_ += boost::filesystem::file_size(p);
return true;
}
virtual const char* op() const {
return "checking size";
}
intmax_t totalSize_;
};
SizeAccumulator sa;
_applyOpToDataFiles(database, sa);
return sa.size();
}
// move temp files to standard data dir
void _replaceWithRecovered(const string& database, const char* reservedPathString) {
Path newPath(storageGlobalParams.dbpath);
if (storageGlobalParams.directoryperdb)
newPath /= database;
class Replacer : public FileOp {
public:
Replacer(const Path& newPath) : newPath_(newPath) {}
private:
const boost::filesystem::path& newPath_;
virtual bool apply(const Path& p) {
if (!boost::filesystem::exists(p))
return false;
boostRenameWrapper(p, newPath_ / p.leaf());
return true;
}
virtual const char* op() const {
return "renaming";
}
} replacer(newPath);
_applyOpToDataFiles(database, replacer, true, reservedPathString);
}
// generate a directory name for storing temp data files
Path uniqueReservedPath(const char* prefix) {
Path repairPath = Path(storageGlobalParams.repairpath);
Path reservedPath;
int i = 0;
bool exists = false;
do {
stringstream ss;
ss << prefix << "_repairDatabase_" << i++;
reservedPath = repairPath / ss.str();
MONGO_ASSERT_ON_EXCEPTION(exists = boost::filesystem::exists(reservedPath));
} while (exists);
return reservedPath;
}
void _applyOpToDataFiles(const string& database,
FileOp& fo,
bool afterAllocator,
const string& path) {
if (afterAllocator)
FileAllocator::get()->waitUntilFinished();
string c = database;
c += '.';
boost::filesystem::path p(path);
if (storageGlobalParams.directoryperdb)
p /= database;
boost::filesystem::path q;
q = p / (c + "ns");
bool ok = false;
MONGO_ASSERT_ON_EXCEPTION(ok = fo.apply(q));
if (ok) {
LOG(2) << fo.op() << " file " << q.string() << endl;
}
int i = 0;
int extra = 10; // should not be necessary, this is defensive in case there are missing files
while (1) {
verify(i <= DiskLoc::MaxFiles);
stringstream ss;
ss << c << i;
q = p / ss.str();
MONGO_ASSERT_ON_EXCEPTION(ok = fo.apply(q));
if (ok) {
if (extra != 10) {
LOG(1) << fo.op() << " file " << q.string() << endl;
log() << " _applyOpToDataFiles() warning: extra == " << extra << endl;
}
} else if (--extra <= 0)
break;
i++;
}
}
class RepairFileDeleter {
public:
RepairFileDeleter(OperationContext* txn,
const string& dbName,
const string& pathString,
const Path& path)
: _txn(txn), _dbName(dbName), _pathString(pathString), _path(path), _success(false) {}
~RepairFileDeleter() {
if (_success)
return;
log() << "cleaning up failed repair "
<< "db: " << _dbName << " path: " << _pathString;
try {
getDur().syncDataAndTruncateJournal(_txn);
// need both in case journaling is disabled
MongoFile::flushAll(_txn, true);
MONGO_ASSERT_ON_EXCEPTION(boost::filesystem::remove_all(_path));
} catch (DBException& e) {
error() << "RepairFileDeleter failed to cleanup: " << redact(e);
error() << "aborting";
fassertFailed(17402);
}
}
void success() {
_success = true;
}
private:
OperationContext* _txn;
string _dbName;
string _pathString;
Path _path;
bool _success;
};
Status MMAPV1Engine::repairDatabase(OperationContext* txn,
const std::string& dbName,
bool preserveClonedFilesOnFailure,
bool backupOriginalFiles) {
unique_ptr<RepairFileDeleter> repairFileDeleter;
// Must be done before and after repair
getDur().syncDataAndTruncateJournal(txn);
intmax_t totalSize = dbSize(dbName);
intmax_t freeSize = File::freeSpace(storageGlobalParams.repairpath);
if (freeSize > -1 && freeSize < totalSize) {
return Status(ErrorCodes::OutOfDiskSpace,
str::stream() << "Cannot repair database " << dbName << " having size: "
<< totalSize
<< " (bytes) because free disk space is: "
<< freeSize
<< " (bytes)");
}
txn->checkForInterrupt();
Path reservedPath = uniqueReservedPath(
(preserveClonedFilesOnFailure || backupOriginalFiles) ? "backup" : "_tmp");
bool created = false;
MONGO_ASSERT_ON_EXCEPTION(created = boost::filesystem::create_directory(reservedPath));
invariant(created);
string reservedPathString = reservedPath.string();
if (!preserveClonedFilesOnFailure)
repairFileDeleter.reset(
new RepairFileDeleter(txn, dbName, reservedPathString, reservedPath));
{
Database* originalDatabase = dbHolder().openDb(txn, dbName);
if (originalDatabase == NULL) {
return Status(ErrorCodes::NamespaceNotFound, "database does not exist to repair");
}
unique_ptr<MMAPV1DatabaseCatalogEntry> dbEntry;
unique_ptr<Database> tempDatabase;
// Must call this before MMAPV1DatabaseCatalogEntry's destructor closes the DB files
ON_BLOCK_EXIT([&dbEntry, &txn] {
getDur().syncDataAndTruncateJournal(txn);
dbEntry->close(txn);
});
{
dbEntry.reset(new MMAPV1DatabaseCatalogEntry(
txn,
dbName,
reservedPathString,
storageGlobalParams.directoryperdb,
true,
_extentManagerFactory->create(
dbName, reservedPathString, storageGlobalParams.directoryperdb)));
tempDatabase.reset(new Database(txn, dbName, dbEntry.get()));
}
map<string, CollectionOptions> namespacesToCopy;
{
string ns = dbName + ".system.namespaces";
OldClientContext ctx(txn, ns);
Collection* coll = originalDatabase->getCollection(ns);
if (coll) {
auto cursor = coll->getCursor(txn);
while (auto record = cursor->next()) {
BSONObj obj = record->data.releaseToBson();
string ns = obj["name"].String();
NamespaceString nss(ns);
if (nss.isSystem()) {
if (nss.isSystemDotIndexes())
continue;
if (nss.coll() == "system.namespaces")
continue;
}
if (!nss.isNormal())
continue;
CollectionOptions options;
if (obj["options"].isABSONObj()) {
Status status = options.parse(obj["options"].Obj());
if (!status.isOK())
return status;
}
namespacesToCopy[ns] = options;
}
}
}
for (map<string, CollectionOptions>::const_iterator i = namespacesToCopy.begin();
i != namespacesToCopy.end();
++i) {
string ns = i->first;
CollectionOptions options = i->second;
Collection* tempCollection = NULL;
{
WriteUnitOfWork wunit(txn);
tempCollection = tempDatabase->createCollection(txn, ns, options, false);
wunit.commit();
}
OldClientContext readContext(txn, ns, originalDatabase);
Collection* originalCollection = originalDatabase->getCollection(ns);
invariant(originalCollection);
// data
// TODO SERVER-14812 add a mode that drops duplicates rather than failing
MultiIndexBlock indexer(txn, tempCollection);
{
vector<BSONObj> indexes;
IndexCatalog::IndexIterator ii =
originalCollection->getIndexCatalog()->getIndexIterator(txn, false);
while (ii.more()) {
IndexDescriptor* desc = ii.next();
indexes.push_back(desc->infoObj());
}
Status status = indexer.init(indexes).getStatus();
if (!status.isOK()) {
return status;
}
}
std::vector<MultiIndexBlock*> indexers{&indexer};
auto cursor = originalCollection->getCursor(txn);
while (auto record = cursor->next()) {
BSONObj doc = record->data.releaseToBson();
WriteUnitOfWork wunit(txn);
Status status = tempCollection->insertDocument(txn, doc, indexers, false);
if (!status.isOK())
return status;
wunit.commit();
txn->checkForInterrupt();
}
Status status = indexer.doneInserting();
if (!status.isOK())
return status;
{
WriteUnitOfWork wunit(txn);
indexer.commit();
wunit.commit();
}
}
getDur().syncDataAndTruncateJournal(txn);
// need both in case journaling is disabled
MongoFile::flushAll(txn, true);
txn->checkForInterrupt();
}
// at this point if we abort, we don't want to delete new files
// as they might be the only copies
if (repairFileDeleter.get())
repairFileDeleter->success();
// Close the database so we can rename/delete the original data files
dbHolder().close(txn, dbName);
if (backupOriginalFiles) {
_renameForBackup(dbName, reservedPath);
} else {
// first make new directory before deleting data
Path newDir = Path(storageGlobalParams.dbpath) / dbName;
MONGO_ASSERT_ON_EXCEPTION(boost::filesystem::create_directory(newDir));
// this deletes old files
_deleteDataFiles(dbName);
if (!boost::filesystem::exists(newDir)) {
// we deleted because of directoryperdb
// re-create
MONGO_ASSERT_ON_EXCEPTION(boost::filesystem::create_directory(newDir));
}
}
_replaceWithRecovered(dbName, reservedPathString.c_str());
if (!backupOriginalFiles) {
MONGO_ASSERT_ON_EXCEPTION(boost::filesystem::remove_all(reservedPath));
}
// Reopen the database so it's discoverable
dbHolder().openDb(txn, dbName);
return Status::OK();
}
}
|