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
|
/**
* Copyright (C) 2008 10gen 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.
*/
#include "mongo/db/repl/finding_start_cursor.h"
#include "mongo/client/dbclientinterface.h"
#include "mongo/db/cursor.h"
#include "mongo/db/matcher.h"
#include "mongo/db/query_plan.h"
#include "mongo/db/queryutil.h"
namespace mongo {
// Configurable for testing.
int FindingStartCursor::_initialTimeout = 5;
// -------------------------------------
FindingStartCursor *FindingStartCursor::make( const QueryPlan &qp ) {
auto_ptr<FindingStartCursor> ret( new FindingStartCursor( qp ) );
ret->init();
return ret.release();
}
FindingStartCursor::FindingStartCursor( const QueryPlan &qp ) :
_qp( qp ),
_findingStart( true ),
_findingStartMode() {
}
void FindingStartCursor::next() {
if ( !_findingStartCursor || !_findingStartCursor->ok() ) {
_findingStart = false;
_c = _qp.newCursor(); // on error, start from beginning
destroyClientCursor();
return;
}
switch( _findingStartMode ) {
// Initial mode: scan backwards from end of collection
case Initial: {
if ( !_matcher->matchesCurrent( _findingStartCursor->c() ) ) {
_findingStart = false; // found first record out of query range, so scan normally
_c = _qp.newCursor( _findingStartCursor->currLoc() );
destroyClientCursor();
return;
}
_findingStartCursor->advance();
RARELY {
if ( _findingStartTimer.seconds() >= _initialTimeout ) {
// If we've scanned enough, switch to find extent mode.
createClientCursor( extentFirstLoc( _findingStartCursor->currLoc() ) );
_findingStartMode = FindExtent;
return;
}
}
return;
}
// FindExtent mode: moving backwards through extents, check first
// document of each extent.
case FindExtent: {
if ( !_matcher->matchesCurrent( _findingStartCursor->c() ) ) {
_findingStartMode = InExtent;
return;
}
DiskLoc prev = prevExtentFirstLoc( _findingStartCursor->currLoc() );
if ( prev.isNull() ) { // hit beginning, so start scanning from here
createClientCursor();
_findingStartMode = InExtent;
return;
}
// There might be a more efficient implementation than creating new cursor & client cursor each time,
// not worrying about that for now
createClientCursor( prev );
return;
}
// InExtent mode: once an extent is chosen, find starting doc in the extent.
case InExtent: {
if ( _matcher->matchesCurrent( _findingStartCursor->c() ) ) {
_findingStart = false; // found first record in query range, so scan normally
_c = _qp.newCursor( _findingStartCursor->currLoc() );
destroyClientCursor();
return;
}
_findingStartCursor->advance();
return;
}
default: {
massert( 14038, "invalid _findingStartMode", false );
}
}
}
DiskLoc FindingStartCursor::extentFirstLoc( const DiskLoc &rec ) {
Extent *e = rec.rec()->myExtent( rec );
if ( !_qp.nsd()->capLooped() || ( e->myLoc != _qp.nsd()->capExtent() ) )
return e->firstRecord;
// Likely we are on the fresh side of capExtent, so return first fresh record.
// If we are on the stale side of capExtent, then the collection is small and it
// doesn't matter if we start the extent scan with capFirstNewRecord.
return _qp.nsd()->capFirstNewRecord();
}
DiskLoc FindingStartCursor::prevExtentFirstLoc( const DiskLoc& rec ) const {
Extent *e = rec.rec()->myExtent( rec );
if ( _qp.nsd()->capLooped() ) {
while( true ) {
// Advance e to preceding extent (looping to lastExtent if necessary).
if ( e->xprev.isNull() ) {
e = _qp.nsd()->lastExtent().ext();
}
else {
e = e->xprev.ext();
}
if ( e->myLoc == _qp.nsd()->capExtent() ) {
// Reached the extent containing the oldest data in the collection.
return DiskLoc();
}
if ( !e->firstRecord.isNull() ) {
// Return the first record of the first non empty extent encountered.
return e->firstRecord;
}
}
}
else {
while( true ) {
if ( e->xprev.isNull() ) {
// Reached the beginning of the collection.
return DiskLoc();
}
e = e->xprev.ext();
if ( !e->firstRecord.isNull() ) {
// Return the first record of the first non empty extent encountered.
return e->firstRecord;
}
}
}
}
void FindingStartCursor::createClientCursor( const DiskLoc &startLoc ) {
shared_ptr<Cursor> c = _qp.newCursor( startLoc );
_findingStartCursor.reset( new ClientCursor(QueryOption_NoCursorTimeout, c, _qp.ns()) );
}
bool FindingStartCursor::firstDocMatchesOrEmpty() const {
shared_ptr<Cursor> c = _qp.newCursor();
return !c->ok() || _matcher->matchesCurrent( c.get() );
}
void FindingStartCursor::init() {
BSONElement tsElt = _qp.originalQuery()[ "ts" ];
massert( 13044, "no ts field in query", !tsElt.eoo() );
BSONObjBuilder b;
b.append( tsElt );
BSONObj tsQuery = b.obj();
_matcher.reset(new CoveredIndexMatcher(tsQuery, _qp.indexKey()));
if ( firstDocMatchesOrEmpty() ) {
_c = _qp.newCursor();
_findingStart = false;
return;
}
// Use a ClientCursor here so we can release db mutex while scanning
// oplog (can take quite a while with large oplogs).
shared_ptr<Cursor> c = _qp.newReverseCursor();
_findingStartCursor.reset( new ClientCursor(QueryOption_NoCursorTimeout, c, _qp.ns(), BSONObj()) );
_findingStartTimer.reset();
_findingStartMode = Initial;
}
shared_ptr<Cursor> FindingStartCursor::getCursor( const char *ns, const BSONObj &query, const BSONObj &order ) {
NamespaceDetails *d = nsdetails(ns);
if ( !d ) {
return shared_ptr<Cursor>( new BasicCursor( DiskLoc() ) );
}
FieldRangeSetPair frsp( ns, query );
scoped_ptr<QueryPlan> oplogPlan( QueryPlan::make( d, -1, frsp, 0, query, order ) );
scoped_ptr<FindingStartCursor> finder( FindingStartCursor::make( *oplogPlan ) );
ElapsedTracker yieldCondition( 256, 20 );
while( !finder->done() ) {
if ( yieldCondition.intervalHasElapsed() ) {
if ( finder->prepareToYield() ) {
ClientCursor::staticYield( 0, ns, 0 );
finder->recoverFromYield();
}
}
finder->next();
}
shared_ptr<Cursor> ret = finder->cursor();
shared_ptr<CoveredIndexMatcher> matcher( new CoveredIndexMatcher( query, BSONObj() ) );
ret->setMatcher( matcher );
return ret;
}
}
|
