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
|
/**
* 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.
*/
#pragma once
#include "mongo/db/catalog/collection.h"
#include "mongo/db/exec/plan_stage.h"
#include "mongo/db/exec/working_set.h"
#include "mongo/db/jsobj.h"
#include "mongo/db/query/canonical_query.h"
#include "mongo/db/query/plan_ranker.h"
#include "mongo/db/query/plan_yield_policy.h"
#include "mongo/db/query/query_solution.h"
#include "mongo/db/record_id.h"
namespace mongo {
/**
* This stage outputs its mainChild, and possibly it's backup child
* and also updates the cache.
*
* Preconditions: Valid RecordId.
*
* Owns the query solutions and PlanStage roots for all candidate plans.
*/
class MultiPlanStage final : public PlanStage {
public:
/**
* Callers use this to specify how the MultiPlanStage should interact with the plan cache.
*/
enum class CachingMode {
// Always write a cache entry for the winning plan to the plan cache, overwriting any
// previously existing cache entry for the query shape.
AlwaysCache,
// Write a cache entry for the query shape *unless* we encounter one of the following edge
// cases:
// - Two or more plans tied for the win.
// - The winning plan returned zero query results during the plan ranking trial period.
SometimesCache,
// Do not write to the plan cache.
NeverCache,
};
/**
* Takes no ownership.
*
* If 'shouldCache' is true, writes a cache entry for the winning plan to the plan cache
* when possible. If 'shouldCache' is false, the plan cache will never be written.
*/
MultiPlanStage(OperationContext* opCtx,
const Collection* collection,
CanonicalQuery* cq,
CachingMode cachingMode = CachingMode::AlwaysCache);
bool isEOF() final;
StageState doWork(WorkingSetID* out) final;
StageType stageType() const final {
return STAGE_MULTI_PLAN;
}
std::unique_ptr<PlanStageStats> getStats() final;
const SpecificStats* getSpecificStats() const final;
/**
* Takes ownership of PlanStage. Does not take ownership of WorkingSet.
*/
void addPlan(std::unique_ptr<QuerySolution> solution, PlanStage* root, WorkingSet* sharedWs);
/**
* Runs all plans added by addPlan, ranks them, and picks a best.
* All further calls to work(...) will return results from the best plan.
*
* If 'yieldPolicy' is non-NULL, then all locks may be yielded in between round-robin
* works of the candidate plans. By default, 'yieldPolicy' is NULL and no yielding will
* take place.
*
* Returns a non-OK status if query planning fails. In particular, this function returns
* ErrorCodes::QueryPlanKilled if the query plan was killed during a yield.
*/
Status pickBestPlan(PlanYieldPolicy* yieldPolicy);
/**
* Returns the number of times that we are willing to work a plan during a trial period.
*
* Calculated based on a fixed query knob and the size of the collection.
*/
static size_t getTrialPeriodWorks(OperationContext* opCtx, const Collection* collection);
/**
* Returns the max number of documents which we should allow any plan to return during the
* trial period. As soon as any plan hits this number of documents, the trial period ends.
*/
static size_t getTrialPeriodNumToReturn(const CanonicalQuery& query);
/** Return true if a best plan has been chosen */
bool bestPlanChosen() const;
/** Return the index of the best plan chosen, for testing */
int bestPlanIdx() const;
/**
* Returns the QuerySolution for the best plan, or NULL if no best plan
*
* The MultiPlanStage retains ownership of the winning QuerySolution and returns an
* unowned pointer.
*/
QuerySolution* bestSolution();
/**
* Returns true if a backup plan was picked.
* This is the case when the best plan has a blocking stage.
* Exposed for testing.
*/
bool hasBackupPlan() const;
//
// Used by explain.
//
static const char* kStageType;
private:
//
// Have all our candidate plans do something.
// If all our candidate plans fail, *objOut will contain
// information on the failure.
//
/**
* Calls work on each child plan in a round-robin fashion. We stop when any plan hits EOF
* or returns 'numResults' results.
*
* Returns true if we need to keep working the plans and false otherwise.
*/
bool workAllPlans(size_t numResults, PlanYieldPolicy* yieldPolicy);
/**
* Checks whether we need to perform either a timing-based yield or a yield for a document
* fetch. If so, then uses 'yieldPolicy' to actually perform the yield.
*
* Returns a non-OK status if killed during a yield or if the query has exceeded its time limit.
*/
Status tryYield(PlanYieldPolicy* yieldPolicy);
static const int kNoSuchPlan = -1;
// Not owned here. Must be non-null.
const Collection* _collection;
// Describes the cases in which we should write an entry for the winning plan to the plan cache.
const CachingMode _cachingMode;
// The query that we're trying to figure out the best solution to.
// not owned here
CanonicalQuery* _query;
// Candidate plans. Each candidate includes a child PlanStage tree and QuerySolution. Ownership
// of all QuerySolutions is retained here, and will *not* be tranferred to the PlanExecutor that
// wraps this stage. Ownership of the PlanStages will be in PlanStage::_children which maps
// one-to-one with _candidates.
std::vector<CandidatePlan> _candidates;
// index into _candidates, of the winner of the plan competition
// uses -1 / kNoSuchPlan when best plan is not (yet) known
int _bestPlanIdx;
// index into _candidates, of the backup plan for sort
// uses -1 / kNoSuchPlan when best plan is not (yet) known
int _backupPlanIdx;
// Set if this MultiPlanStage cannot continue, and the query must fail. This can happen in
// two ways. The first is that all candidate plans fail. Note that one plan can fail
// during normal execution of the plan competition. Here is an example:
//
// Plan 1: collection scan with sort. Sort runs out of memory.
// Plan 2: ixscan that provides sort. Won't run out of memory.
//
// We want to choose plan 2 even if plan 1 fails.
//
// The second way for failure to occur is that the execution of this query is killed during
// a yield, by some concurrent event such as a collection drop.
bool _failure;
// If everything fails during the plan competition, we can't pick one.
size_t _failureCount;
// if pickBestPlan fails, this is set to the wsid of the statusMember
// returned by ::work()
WorkingSetID _statusMemberId;
// When a stage requests a yield for document fetch, it gives us back a RecordFetcher*
// to use to pull the record into memory. We take ownership of the RecordFetcher here,
// deleting it after we've had a chance to do the fetch. For timing-based yields, we
// just pass a NULL fetcher.
std::unique_ptr<RecordFetcher> _fetcher;
// Stats
MultiPlanStats _specificStats;
};
} // namespace mongo
|