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
|
/**
* Copyright (C) 2019-present MongoDB, Inc.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the Server Side Public License, version 1,
* as published by MongoDB, Inc.
*
* 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
* Server Side Public License for more details.
*
* You should have received a copy of the Server Side Public License
* along with this program. If not, see
* <http://www.mongodb.com/licensing/server-side-public-license>.
*
* 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 Server Side 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.
*/
#pragma once
#include "mongo/platform/basic.h"
#include <fmt/format.h>
#include "mongo/db/db_raii.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/pipeline/document_source.h"
#include "mongo/db/read_concern.h"
#include "mongo/db/storage/recovery_unit.h"
namespace mongo {
using namespace fmt::literals;
/**
* Manipulates the state of the OperationContext so that while this object is in scope, reads and
* writes will use a local read concern and see the latest version of the data. It will also reset
* ignore_prepared on the recovery unit so that any reads or writes will block on a conflict with a
* prepared transaction. Resets the OperationContext back to its original state upon destruction.
*/
class DocumentSourceWriteBlock {
OperationContext* _opCtx;
repl::ReadConcernArgs _originalArgs;
RecoveryUnit::ReadSource _originalSource;
EnforcePrepareConflictsBlock _enforcePrepareConflictsBlock;
public:
DocumentSourceWriteBlock(OperationContext* opCtx)
: _opCtx(opCtx), _enforcePrepareConflictsBlock(opCtx) {
_originalArgs = repl::ReadConcernArgs::get(_opCtx);
_originalSource = _opCtx->recoveryUnit()->getTimestampReadSource();
repl::ReadConcernArgs::get(_opCtx) = repl::ReadConcernArgs();
_opCtx->recoveryUnit()->setTimestampReadSource(RecoveryUnit::kUnset);
}
~DocumentSourceWriteBlock() {
repl::ReadConcernArgs::get(_opCtx) = _originalArgs;
_opCtx->recoveryUnit()->setTimestampReadSource(_originalSource);
}
};
/**
* This is a base abstract class for all stages performing a write operation into an output
* collection. The writes are organized in batches in which elements are objects of the templated
* type 'B'. A subclass must override two methods to be able to write into the output collection:
*
* 1. 'makeBatchObject()' - to create an object of type 'B' from the given 'Document', which is,
* essentially, a result of the input source's 'getNext()' .
* 2. 'spill()' - to write the batch into the output collection.
*
* Two other virtual methods exist which a subclass may override: 'initialize()' and 'finalize()',
* which are called before the first element is read from the input source, and after the last one
* has been read, respectively.
*/
template <typename B>
class DocumentSourceWriter : public DocumentSource {
public:
using BatchObject = B;
using BatchedObjects = std::vector<BatchObject>;
DocumentSourceWriter(NamespaceString outputNs,
const boost::intrusive_ptr<ExpressionContext>& expCtx)
: DocumentSource(expCtx),
_outputNs(std::move(outputNs)),
_writeConcern(expCtx->opCtx->getWriteConcern()) {}
DepsTracker::State getDependencies(DepsTracker* deps) const override {
deps->needWholeDocument = true;
return DepsTracker::State::EXHAUSTIVE_ALL;
}
GetModPathsReturn getModifiedPaths() const override {
// For purposes of tracking which fields come from where, the writer stage does not modify
// any fields by default.
return {GetModPathsReturn::Type::kFiniteSet, std::set<std::string>{}, {}};
}
boost::optional<DistributedPlanLogic> distributedPlanLogic() override {
return DistributedPlanLogic{nullptr, this, boost::none};
}
bool canRunInParallelBeforeWriteStage(
const std::set<std::string>& nameOfShardKeyFieldsUponEntryToStage) const override {
return true;
}
const NamespaceString& getOutputNs() const {
return _outputNs;
}
GetNextResult getNext() final override;
protected:
/**
* Prepares the stage to be able to write incoming batches.
*/
virtual void initialize() {}
/**
* Finalize the output collection, called when there are no more documents to write.
*/
virtual void finalize() {}
/**
* Writes the documents in 'batch' to the output namespace.
*/
virtual void spill(BatchedObjects&& batch) = 0;
/**
* Creates a batch object from the given document and returns it to the caller along with the
* object size.
*/
virtual std::pair<B, int> makeBatchObject(Document&& doc) const = 0;
/**
* A subclass may override this method to enable a fail point right after a next input element
* has been retrieved, but not processed yet.
*/
virtual void waitWhileFailPointEnabled() {}
// The namespace where the output will be written to.
const NamespaceString _outputNs;
// Stash the writeConcern of the original command as the operation context may change by the
// time we start to spill writes. This is because certain aggregations (e.g. $exchange)
// establish cursors with batchSize 0 then run subsequent getMore's which use a new operation
// context. The getMore's will not have an attached writeConcern however we still want to
// respect the writeConcern of the original command.
WriteConcernOptions _writeConcern;
private:
bool _initialized{false};
bool _done{false};
};
template <typename B>
DocumentSource::GetNextResult DocumentSourceWriter<B>::getNext() {
pExpCtx->checkForInterrupt();
if (_done) {
return GetNextResult::makeEOF();
}
if (!_initialized) {
// Explain should never try to actually execute any writes. We only ever expect
// getNext() to be called for the 'executionStats' and 'allPlansExecution' explain
// modes. This assertion should not be triggered for 'queryPlanner' explain, which
// is perfectly legal.
uassert(51029,
"explain of {} is not allowed with verbosity {}"_format(
getSourceName(), ExplainOptions::verbosityString(*pExpCtx->explain)),
!pExpCtx->explain);
initialize();
_initialized = true;
}
BatchedObjects batch;
int bufferedBytes = 0;
auto nextInput = pSource->getNext();
for (; nextInput.isAdvanced(); nextInput = pSource->getNext()) {
waitWhileFailPointEnabled();
auto doc = nextInput.releaseDocument();
auto[obj, objSize] = makeBatchObject(std::move(doc));
bufferedBytes += objSize;
if (!batch.empty() &&
(bufferedBytes > BSONObjMaxUserSize || batch.size() >= write_ops::kMaxWriteBatchSize)) {
spill(std::move(batch));
batch.clear();
bufferedBytes = objSize;
}
batch.push_back(obj);
}
if (!batch.empty()) {
spill(std::move(batch));
batch.clear();
}
switch (nextInput.getStatus()) {
case GetNextResult::ReturnStatus::kAdvanced: {
MONGO_UNREACHABLE; // We consumed all advances above.
}
case GetNextResult::ReturnStatus::kPauseExecution: {
return nextInput; // Propagate the pause.
}
case GetNextResult::ReturnStatus::kEOF: {
_done = true;
finalize();
return nextInput;
}
}
MONGO_UNREACHABLE;
}
} // namespace mongo
|
