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
|
/**
* Copyright (C) 2020-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/db/repl/insert_group.h"
namespace mongo {
class CollatorInterface;
class OpCounters;
namespace repl {
/**
* Caches per-collection properties which are relevant for oplog application, so that they don't
* have to be retrieved repeatedly for each op.
*/
class CachedCollectionProperties {
public:
struct CollectionProperties {
bool isCapped = false;
bool isClustered = false;
const CollatorInterface* collator = nullptr;
};
CollectionProperties getCollectionProperties(OperationContext* opCtx,
const StringMapHashedKey& ns);
private:
StringMap<CollectionProperties> _cache;
};
/**
* This class contains some static methods common to ordinary oplog application and oplog
* application as part of tenant migration.
*/
class OplogApplierUtils {
public:
/**
* Sort operations by their namespaces. There are some special rules for sorting:
*
* 1. Prepared transaction commands (prepare/commit/abort) act as delimiters, which
* creates partitions in the ops vector so that any reordering can only happen
* within each partition but not across partitions.
* 2. Within each partition, DDL ops (with the '$cmd' collection name) are ordered
* before all other CRUD ops since DDL ops need to be run before CRUD ops.
*
* As an example, the right side below is a possible output after calling this:
* [ [
* insert(ns2, docA), create(ns1), -
* create(ns1), insert(ns1, docB), |
* insert(ns1, docB), sort insert(ns1, docC), |- sorted
* update(ns2, docA), ==========> insert(ns2, docA), |
* insert(ns1, docC), update(ns2, docA), -
* prepare(sess1, txn1), prepare(sess1, txn1),
* insert(ns1, docD), insert(ns1, docD), -
* delete(ns2, docA), update(ns1, docB), |- sorted
* update(ns1, docB), delete(ns2, docA), -
* ] ]
*/
static void stableSortByNamespace(std::vector<ApplierOperation>* ops);
/**
* Updates a CRUD op's hash and isForCappedCollection field if necessary.
*/
static void processCrudOp(
OperationContext* opCtx,
OplogEntry* op,
uint32_t* hash,
const CachedCollectionProperties::CollectionProperties& collProperties);
/**
* Adds a single oplog entry to the appropriate writer vector. Returns the index of the
* writer vector the entry was written to.
*/
static uint32_t addToWriterVector(OperationContext* opCtx,
OplogEntry* op,
std::vector<std::vector<const OplogEntry*>>* writerVectors,
CachedCollectionProperties* collPropertiesCache,
boost::optional<uint32_t> forceWriterId = boost::none);
/**
* Same as above, except that the type of ops in the writer vectors are different.
*/
static uint32_t addToWriterVector(OperationContext* opCtx,
OplogEntry* op,
std::vector<std::vector<ApplierOperation>>* writerVectors,
CachedCollectionProperties* collPropertiesCache,
boost::optional<uint32_t> forceWriterId = boost::none);
/**
* Adds a set of derivedOps to writerVectors. For ops derived from prepared transactions, the
* addDerivedPrepares() variant should be used.
*
* If `serial` is true, assign all derived operations to the writer vector corresponding to the
* hash of the first operation in `derivedOps`.
*/
static void addDerivedOps(OperationContext* opCtx,
std::vector<OplogEntry>* derivedOps,
std::vector<std::vector<ApplierOperation>>* writerVectors,
CachedCollectionProperties* collPropertiesCache,
bool serial);
/**
* Adds a set of derived prepared transaction operations to writerVectors.
*
* The prepareOp and derivedOps are inputs that we use to generate ApplierOperation's to be
* added to the writerVectors. The derivedOps contains all the CRUD ops inside the applyOps
* part of the prepareOp. When this function finishes the writerVectors may look like this:
*
* ========================== for non-empty prepared transaction ==========================
* writer vector 1: [ ]
* writer vector 2: [
* ApplierOperation{
* op: prepareOp,
* instruction: applySplitPrepareOp,
* subSession: <split_session_id_1>,
* splitPrepareOps: [ crud_op_1, crud_op_3 ],
* },
* ]
* writer vector 3: [
* // This op should already exist in the writerVector prior to this function call.
* ApplierOperation{ op: <config.transaction_update_op>, instruction: applyOplogEntry },
* ]
* writer vector 4: [
* ApplierOperation{
* op: prepareOp,
* instruction: applySplitPrepareOp,
* subSession: <split_session_id_2>,
* splitPrepareOps: [ crud_op_2, crud_op_4 ],
* },
* ]
* ============================ for empty prepared transaction ============================
* writer vector 1: [ ]
* writer vector 2: [
* // This op should already exist in the writerVector prior to this function call.
* ApplierOperation{ op: <config.transaction_update_op>, instruction: applyOplogEntry },
* ]
* writer vector 3: [ ]
* writer vector 4: [
* // The splitPrepareOps list is made empty, which we can still correctly apply.
* ApplierOperation{
* op: <original_prepare_op>,
* instruction: applySplitPrepareOp,
* subSession: <split_session_id_1>,
* splitPrepareOps: [ ],
* },
* ]
*/
static void addDerivedPrepares(OperationContext* opCtx,
OplogEntry* prepareOp,
std::vector<OplogEntry>* derivedOps,
std::vector<std::vector<ApplierOperation>>* writerVectors,
CachedCollectionProperties* collPropertiesCache);
/**
* Adds commit or abort transaction operations to the writerVectors.
*
* The commitOrAbortOp is the input that we use to generate ApplierOperation's to be added
* to those writerVectors that previously got assigned the prepare ops. When this function
* finishes the writerVectors may look like this:
*
* writer vector 1: [ ]
* writer vector 2: [
* ApplierOperation{
* op: commitOrAbortOp,
* instruction: applySplitCommitOrAbortOp,
* subSession: <split_session_id_1>,
* splitPrepareOps: [ ],
* },
* ]
* writer vector 3: [
* // This op should already exist in the writerVector prior to this function call.
* ApplierOperation{ op: <config.transaction_update_op>, instruction: applyOplogEntry },
* ]
* writer vector 4: [
* ApplierOperation{
* op: commitOrAbortOp,
* instruction: applySplitCommitOrAbortOp,
* subSession: <split_session_id_2>,
* splitPrepareOps: [ ],
* },
* ]
*/
static void addDerivedCommitsOrAborts(OperationContext* opCtx,
OplogEntry* commitOrAbortOp,
std::vector<std::vector<ApplierOperation>>* writerVectors,
CachedCollectionProperties* collPropertiesCache);
/**
* Returns the namespace string for this oplogEntry; if it has a UUID it looks up the
* corresponding namespace and returns it, otherwise it returns the oplog entry 'nss'. If
* there is a UUID and no namespace with that ID is found, throws NamespaceNotFound.
*/
static NamespaceString parseUUIDOrNs(OperationContext* opCtx, const OplogEntry& oplogEntry);
/**
* If the oplog entry has a UUID, returns the UUID with the database from 'nss'. Otherwise
* returns 'nss'
*/
static NamespaceStringOrUUID getNsOrUUID(const NamespaceString& nss, const OplogEntry& op);
/**
* The logic for oplog entry application which is shared between standard and tenant oplog
* application.
*/
static Status applyOplogEntryOrGroupedInsertsCommon(
OperationContext* opCtx,
const OplogEntryOrGroupedInserts& entryOrGroupedInserts,
OplogApplication::Mode oplogApplicationMode,
bool isDataConsistent,
IncrementOpsAppliedStatsFn incrementOpsAppliedStats,
OpCounters* opCounters);
/**
* The logic for oplog batch application which is shared between standard and tenant oplog
* application.
*/
static Status applyOplogBatchCommon(
OperationContext* opCtx,
std::vector<ApplierOperation>* ops,
OplogApplication::Mode oplogApplicationMode,
bool allowNamespaceNotFoundErrorsOnCrudOps,
bool isDataConsistent,
InsertGroup::ApplyFunc applyOplogEntryOrGroupedInserts) noexcept;
};
} // namespace repl
} // namespace mongo
|
