summaryrefslogtreecommitdiff
path: root/src/mongo/db/dbcommands_admin.cpp
blob: ffcc3f261fe1ac4029bbbc319fca1ccd439cee07 (plain)
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
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
// dbcommands_admin.cpp

/**
 *
 *    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/>.
 */

/**
   this file has dbcommands that are for dba type administration
   mostly around dbs and collections
   NOT system stuff
*/


#include "pch.h"
#include "jsobj.h"
#include "pdfile.h"
#include "namespace-inl.h"
#include "commands.h"
#include "cmdline.h"
#include "btree.h"
#include "curop-inl.h"
#include "../util/background.h"
#include "../util/logfile.h"
#include "../util/alignedbuilder.h"
#include "../util/paths.h"
#include "../scripting/engine.h"
#include "../util/timer.h"

namespace mongo {

    class CleanCmd : public Command {
    public:
        CleanCmd() : Command( "clean" ) {}

        virtual bool slaveOk() const { return true; }
        virtual LockType locktype() const { return WRITE; }

        virtual void help(stringstream& h) const { h << "internal"; }

        bool run(const string& dbname, BSONObj& cmdObj, int, string& errmsg, BSONObjBuilder& result, bool fromRepl ) {
            string dropns = dbname + "." + cmdObj.firstElement().valuestrsafe();

            if ( !cmdLine.quiet )
                tlog() << "CMD: clean " << dropns << endl;

            NamespaceDetails *d = nsdetails(dropns.c_str());

            if ( ! d ) {
                errmsg = "ns not found";
                return 0;
            }

            for ( int i = 0; i < Buckets; i++ )
                d->deletedList[i].Null();

            result.append("ns", dropns.c_str());
            return 1;
        }

    } cleanCmd;

    namespace dur {
        boost::filesystem::path getJournalDir();
    }
 
    class JournalLatencyTestCmd : public Command {
    public:
        JournalLatencyTestCmd() : Command( "journalLatencyTest" ) {}

        virtual bool slaveOk() const { return true; }
        virtual LockType locktype() const { return NONE; }
        virtual bool adminOnly() const { return true; }
        virtual void help(stringstream& h) const { h << "test how long to write and fsync to a test file in the journal/ directory"; }

        bool run(const string& dbname, BSONObj& cmdObj, int, string& errmsg, BSONObjBuilder& result, bool fromRepl ) {
            boost::filesystem::path p = dur::getJournalDir();
            p /= "journalLatencyTest";
        
            // remove file if already present
            try { 
                remove(p);
            }
            catch(...) { }

            BSONObjBuilder bb[2];
            for( int pass = 0; pass < 2; pass++ ) {
                LogFile f(p.string());
                AlignedBuilder b(1024 * 1024);
                {
                    Timer t;
                    for( int i = 0 ; i < 100; i++ ) { 
                        f.synchronousAppend(b.buf(), 8192);
                    }
                    bb[pass].append("8KB", t.millis() / 100.0);
                }
                {
                    const int N = 50;
                    Timer t2;
                    long long x = 0;
                    for( int i = 0 ; i < N; i++ ) { 
                        Timer t;
                        f.synchronousAppend(b.buf(), 8192);
                        x += t.micros();
                        sleepmillis(4);
                    }
                    long long y = t2.micros() - 4*N*1000;
                    // not really trusting the timer granularity on all platforms so whichever is higher of x and y
                    bb[pass].append("8KBWithPauses", max(x,y) / (N*1000.0));
                }
                {
                    Timer t;
                    for( int i = 0 ; i < 20; i++ ) { 
                        f.synchronousAppend(b.buf(), 1024 * 1024);
                    }
                    bb[pass].append("1MB", t.millis() / 20.0);
                }
                // second time around, we are prealloced.
            }
            result.append("timeMillis", bb[0].obj());
            result.append("timeMillisWithPrealloc", bb[1].obj());

            try { 
                remove(p);
            }
            catch(...) { }

            try {
                result.append("onSamePartition", onSamePartition(dur::getJournalDir().string(), dbpath));
            }
            catch(...) { }

            return 1;
        }
    } journalLatencyTestCmd;

    class ValidateCmd : public Command {
    public:
        ValidateCmd() : Command( "validate" ) {}

        virtual bool slaveOk() const {
            return true;
        }

        virtual void help(stringstream& h) const { h << "Validate contents of a namespace by scanning its data structures for correctness.  Slow.\n"
                                                        "Add full:true option to do a more thorough check"; }

        virtual LockType locktype() const { return READ; }
        //{ validate: "collectionnamewithoutthedbpart" [, scandata: <bool>] [, full: <bool> } */

        bool run(const string& dbname , BSONObj& cmdObj, int, string& errmsg, BSONObjBuilder& result, bool fromRepl ) {
            string ns = dbname + "." + cmdObj.firstElement().valuestrsafe();
            NamespaceDetails * d = nsdetails( ns.c_str() );
            if ( !cmdLine.quiet )
                tlog() << "CMD: validate " << ns << endl;

            if ( ! d ) {
                errmsg = "ns not found";
                return 0;
            }

            result.append( "ns", ns );
            validateNS( ns.c_str() , d, cmdObj, result);
            return 1;
        }

    private:
        void validateNS(const char *ns, NamespaceDetails *d, const BSONObj& cmdObj, BSONObjBuilder& result) {
            const bool full = cmdObj["full"].trueValue();
            const bool scanData = full || cmdObj["scandata"].trueValue();

            bool valid = true;
            BSONArrayBuilder errors; // explanation(s) for why valid = false
            if ( d->capped ){
                result.append("capped", d->capped);
                result.append("max", d->max);
            }

            result.append("firstExtent", str::stream() << d->firstExtent.toString() << " ns:" << d->firstExtent.ext()->nsDiagnostic.toString());
            result.append( "lastExtent", str::stream() <<  d->lastExtent.toString() << " ns:" <<  d->lastExtent.ext()->nsDiagnostic.toString());
            
            BSONArrayBuilder extentData;

            try {
                d->firstExtent.ext()->assertOk();
                d->lastExtent.ext()->assertOk();

                DiskLoc el = d->firstExtent;
                int ne = 0;
                while( !el.isNull() ) {
                    Extent *e = el.ext();
                    e->assertOk();
                    el = e->xnext;
                    ne++;
                    if ( full )
                        extentData << e->dump();
                    
                    killCurrentOp.checkForInterrupt();
                }
                result.append("extentCount", ne);
            }
            catch (...) {
                valid=false;
                errors << "extent asserted";
            }

            if ( full )
                result.appendArray( "extents" , extentData.arr() );

            
            result.appendNumber("datasize", d->stats.datasize);
            result.appendNumber("nrecords", d->stats.nrecords);
            result.appendNumber("lastExtentSize", d->lastExtentSize);
            result.appendNumber("padding", d->paddingFactor);
            

            try {

                try {
                    result.append("firstExtentDetails", d->firstExtent.ext()->dump());

                    valid = valid && d->firstExtent.ext()->validates() && 
                        d->firstExtent.ext()->xprev.isNull();
                }
                catch (...) {
                    errors << "exception firstextent";
                    valid = false;
                }

                set<DiskLoc> recs;
                if( scanData ) {
                    shared_ptr<Cursor> c = theDataFileMgr.findAll(ns);
                    int n = 0;
                    int nInvalid = 0;
                    long long len = 0;
                    long long nlen = 0;
                    int outOfOrder = 0;
                    DiskLoc cl_last;
                    while ( c->ok() ) {
                        n++;

                        DiskLoc cl = c->currLoc();
                        if ( n < 1000000 )
                            recs.insert(cl);
                        if ( d->capped ) {
                            if ( cl < cl_last )
                                outOfOrder++;
                            cl_last = cl;
                        }

                        Record *r = c->_current();
                        len += r->lengthWithHeaders;
                        nlen += r->netLength();

                        if (full){
                            BSONObj obj(r);
                            if (!obj.isValid() || !obj.valid()){ // both fast and deep checks
                                valid = false;
                                if (nInvalid == 0) // only log once;
                                    errors << "invalid bson object detected (see logs for more info)";

                                nInvalid++;
                                if (strcmp("_id", obj.firstElementFieldName()) == 0){
                                    try {
                                        obj.firstElement().validate(); // throws on error
                                        log() << "Invalid bson detected in " << ns << " with _id: " << obj.firstElement().toString(false) << endl;
                                    }
                                    catch(...){
                                        log() << "Invalid bson detected in " << ns << " with corrupt _id" << endl;
                                    }
                                }
                                else {
                                    log() << "Invalid bson detected in " << ns << " and couldn't find _id" << endl;
                                }
                            }
                        }

                        c->advance();
                    }
                    if ( d->capped && !d->capLooped() ) {
                        result.append("cappedOutOfOrder", outOfOrder);
                        if ( outOfOrder > 1 ) {
                            valid = false;
                            errors << "too many out of order records";
                        }
                    }
                    result.append("objectsFound", n);

                    if (full) {
                        result.append("invalidObjects", nInvalid);
                    }

                    result.appendNumber("bytesWithHeaders", len);
                    result.appendNumber("bytesWithoutHeaders", nlen);
                }

                BSONArrayBuilder deletedListArray;
                for ( int i = 0; i < Buckets; i++ ) {
                    deletedListArray << d->deletedList[i].isNull();
                }

                int ndel = 0;
                long long delSize = 0;
                int incorrect = 0;
                for ( int i = 0; i < Buckets; i++ ) {
                    DiskLoc loc = d->deletedList[i];
                    try {
                        int k = 0;
                        while ( !loc.isNull() ) {
                            if ( recs.count(loc) )
                                incorrect++;
                            ndel++;

                            if ( loc.questionable() ) {
                                if( d->capped && !loc.isValid() && i == 1 ) {
                                    /* the constructor for NamespaceDetails intentionally sets deletedList[1] to invalid
                                       see comments in namespace.h
                                    */
                                    break;
                                }

                                if ( loc.a() <= 0 || strstr(ns, "hudsonSmall") == 0 ) {
                                    string err (str::stream() << "bad deleted loc: " << loc.toString() << " bucket:" << i << " k:" << k);
                                    errors << err;

                                    valid = false;
                                    break;
                                }
                            }

                            DeletedRecord *d = loc.drec();
                            delSize += d->lengthWithHeaders;
                            loc = d->nextDeleted;
                            k++;
                            killCurrentOp.checkForInterrupt();
                        }
                    }
                    catch (...) {
                        errors << ("exception in deleted chain for bucket " + BSONObjBuilder::numStr(i));
                        valid = false;
                    }
                }
                result.appendNumber("deletedCount", ndel);
                result.appendNumber("deletedSize", delSize);

                if ( incorrect ) {
                    errors << (BSONObjBuilder::numStr(incorrect) + " records from datafile are in deleted list");
                    valid = false;
                }

                int idxn = 0;
                try  {
                    result.append("nIndexes", d->nIndexes);
                    BSONObjBuilder indexes; // not using subObjStart to be exception safe
                    NamespaceDetails::IndexIterator i = d->ii();
                    while( i.more() ) {
                        IndexDetails& id = i.next();
                        long long keys = id.idxInterface().fullValidate(id.head, id.keyPattern());
                        indexes.appendNumber(id.indexNamespace(), keys);
                    }
                    result.append("keysPerIndex", indexes.done());
                }
                catch (...) {
                    errors << ("exception during index validate idxn " + BSONObjBuilder::numStr(idxn));
                    valid=false;
                }

            }
            catch (AssertionException) {
                errors << "exception during validate";
                valid = false;
            }

            result.appendBool("valid", valid);
            result.append("errors", errors.arr());

            if ( !full ){
                result.append("warning", "Some checks omitted for speed. use {full:true} option to do more thorough scan.");
            }
            
            if ( !valid ) {
                result.append("advice", "ns corrupt, requires repair");
            }

        }
    } validateCmd;

    bool lockedForWriting = false; // read from db/instance.cpp
    static bool unlockRequested = false;
    static mongo::mutex fsyncLockMutex("fsyncLock");
    static boost::condition fsyncLockCondition;
    static OID fsyncLockID; // identifies the current lock job

    /*
        class UnlockCommand : public Command {
        public:
            UnlockCommand() : Command( "unlock" ) { }
            virtual bool readOnly() { return true; }
            virtual bool slaveOk() const { return true; }
            virtual bool adminOnly() const { return true; }
            virtual bool run(const char *ns, BSONObj& cmdObj, string& errmsg, BSONObjBuilder& result, bool fromRepl) {
                if( lockedForWriting ) {
                    log() << "command: unlock requested" << endl;
                    errmsg = "unlock requested";
                    unlockRequested = true;
                }
                else {
                    errmsg = "not locked, so cannot unlock";
                    return 0;
                }
                return 1;
            }

        } unlockCommand;
    */
    /* see unlockFsync() for unlocking:
       db.$cmd.sys.unlock.findOne()
    */
    class FSyncCommand : public Command {
        static const char* url() { return  "http://www.mongodb.org/display/DOCS/fsync+Command"; }
        class LockDBJob : public BackgroundJob {
        protected:
            virtual string name() const { return "lockdbjob"; }
            void run() {
                Client::initThread("fsyncjob");
                Client& c = cc();
                {
                    scoped_lock lk(fsyncLockMutex);
                    while (lockedForWriting){ // there is a small window for two LockDBJob's to be active. This prevents it.
                        fsyncLockCondition.wait(lk.boost());
                    }
                    lockedForWriting = true;
                    fsyncLockID.init();
                }
                readlock lk("");
                MemoryMappedFile::flushAll(true);
                log() << "db is now locked for snapshotting, no writes allowed. db.fsyncUnlock() to unlock" << endl;
                log() << "    For more info see " << FSyncCommand::url() << endl;
                _ready = true;
                {
                    scoped_lock lk(fsyncLockMutex);
                    while( !unlockRequested ) {
                        fsyncLockCondition.wait(lk.boost());
                    }
                    unlockRequested = false;
                    lockedForWriting = false;
                    fsyncLockCondition.notify_all();
                }
                c.shutdown();
            }
        public:
            bool& _ready;
            LockDBJob(bool& ready) : BackgroundJob( true /* delete self */ ), _ready(ready) {
                _ready = false;
            }
        };
    public:
        FSyncCommand() : Command( "fsync" ) {}
        virtual LockType locktype() const { return WRITE; }
        virtual bool slaveOk() const { return true; }
        virtual bool adminOnly() const { return true; }
        /*virtual bool localHostOnlyIfNoAuth(const BSONObj& cmdObj) {
            string x = cmdObj["exec"].valuestrsafe();
            return !x.empty();
        }*/
        virtual void help(stringstream& h) const { h << url(); }
        virtual bool run(const string& dbname, BSONObj& cmdObj, int, string& errmsg, BSONObjBuilder& result, bool fromRepl) {
            bool sync = !cmdObj["async"].trueValue(); // async means do an fsync, but return immediately
            bool lock = cmdObj["lock"].trueValue();
            log() << "CMD fsync:  sync:" << sync << " lock:" << lock << endl;

            if( lock ) {
                // fsync and lock variation 

                uassert(12034, "fsync: can't lock while an unlock is pending", !unlockRequested);
                uassert(12032, "fsync: sync option must be true when using lock", sync);
                /* With releaseEarly(), we must be extremely careful we don't do anything
                   where we would have assumed we were locked.  profiling is one of those things.
                   Perhaps at profile time we could check if we released early -- however,
                   we need to be careful to keep that code very fast it's a very common code path when on.
                */
                uassert(12033, "fsync: profiling must be off to enter locked mode", cc().database()->profile == 0);

                // todo future: Perhaps we could do this in the background thread.  As is now, writes may interleave between 
                //              the releaseEarly below and the acquisition of the readlock in the background thread. 
                //              However the real problem is that it seems complex to unlock here and then have a window for 
                //              writes before the bg job -- can be done correctly but harder to reason about correctness.
                //              If this command ran within a read lock in the first place, would it work, and then that 
                //              would be quite easy?
                //              Or, could we downgrade the write lock to a read lock, wait for ready, then release?
                getDur().syncDataAndTruncateJournal();

                bool ready = false;
                LockDBJob *l = new LockDBJob(ready);

                d.dbMutex.releaseEarly();
                
                // There is a narrow window for another lock request to come in
                // here before the LockDBJob grabs the readlock. LockDBJob will
                // ensure that the requests are serialized and never running
                // concurrently

                l->go();
                // don't return until background thread has acquired the read lock
                while( !ready ) {
                    sleepmillis(10);
                }
                result.append("info", "now locked against writes, use db.fsyncUnlock() to unlock");
                result.append("seeAlso", url());
            }
            else {
                // the simple fsync command case

                if (sync)
                    getDur().commitNow();
                result.append( "numFiles" , MemoryMappedFile::flushAll( sync ) );
            }
            return 1;
        }

    } fsyncCmd;

    // Note that this will only unlock the current lock.  If another thread
    // relocks before we return we still consider the unlocking successful.
    // This is imporant because if two scripts are trying to fsync-lock, each
    // one must be assured that between the fsync return and the call to unlock
    // that the database is fully locked
    void unlockFsyncAndWait(){
        scoped_lock lk(fsyncLockMutex);
        if (lockedForWriting) { // could have handled another unlock before we grabbed the lock
            OID curOp = fsyncLockID;
            unlockRequested = true;
            fsyncLockCondition.notify_all();
            while (lockedForWriting && fsyncLockID == curOp){
                fsyncLockCondition.wait( lk.boost() );
            }
        }
    }
}