summaryrefslogtreecommitdiff
path: root/src/mongo/db/commands/set_feature_compatibility_version_command.cpp
blob: c2e9cfe5cec763c99a1069055ebf27152cdc3aaa (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
/**
 *    Copyright (C) 2018-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.
 */

#define MONGO_LOGV2_DEFAULT_COMPONENT ::mongo::logv2::LogComponent::kDefault

#include "mongo/platform/basic.h"

#include "mongo/db/auth/authorization_session.h"
#include "mongo/db/catalog/coll_mod.h"
#include "mongo/db/catalog/database.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/catalog/drop_indexes.h"
#include "mongo/db/catalog/index_catalog.h"
#include "mongo/db/commands.h"
#include "mongo/db/commands/feature_compatibility_version.h"
#include "mongo/db/commands/feature_compatibility_version_documentation.h"
#include "mongo/db/commands/feature_compatibility_version_parser.h"
#include "mongo/db/commands/set_feature_compatibility_version_gen.h"
#include "mongo/db/concurrency/d_concurrency.h"
#include "mongo/db/db_raii.h"
#include "mongo/db/dbdirectclient.h"
#include "mongo/db/index_builds_coordinator.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/ops/write_ops.h"
#include "mongo/db/read_write_concern_defaults.h"
#include "mongo/db/repl/repl_client_info.h"
#include "mongo/db/repl/replication_coordinator.h"
#include "mongo/db/s/active_migrations_registry.h"
#include "mongo/db/s/config/sharding_catalog_manager.h"
#include "mongo/db/s/migration_util.h"
#include "mongo/db/s/sharding_state.h"
#include "mongo/db/server_options.h"
#include "mongo/logv2/log.h"
#include "mongo/rpc/get_status_from_command_result.h"
#include "mongo/s/database_version_helpers.h"
#include "mongo/stdx/unordered_set.h"
#include "mongo/util/exit.h"
#include "mongo/util/fail_point.h"
#include "mongo/util/scopeguard.h"

namespace mongo {

using FCVP = FeatureCompatibilityVersionParser;
using FeatureCompatibilityParams = ServerGlobalParams::FeatureCompatibility;

namespace {

MONGO_FAIL_POINT_DEFINE(featureCompatibilityDowngrade);
MONGO_FAIL_POINT_DEFINE(featureCompatibilityUpgrade);
MONGO_FAIL_POINT_DEFINE(failUpgrading);
MONGO_FAIL_POINT_DEFINE(hangWhileUpgrading);
MONGO_FAIL_POINT_DEFINE(failDowngrading);
MONGO_FAIL_POINT_DEFINE(hangWhileDowngrading);

/**
 * Deletes the persisted default read/write concern document.
 */
void deletePersistedDefaultRWConcernDocument(OperationContext* opCtx) {
    DBDirectClient client(opCtx);
    const auto commandResponse = client.runCommand([&] {
        write_ops::Delete deleteOp(NamespaceString::kConfigSettingsNamespace);
        deleteOp.setDeletes({[&] {
            write_ops::DeleteOpEntry entry;
            entry.setQ(BSON("_id" << ReadWriteConcernDefaults::kPersistedDocumentId));
            entry.setMulti(false);
            return entry;
        }()});
        return deleteOp.serialize({});
    }());
    uassertStatusOK(getStatusFromWriteCommandReply(commandResponse->getCommandReply()));
}

/**
 * Sets the minimum allowed feature compatibility version for the cluster. The cluster should not
 * use any new features introduced in binary versions that are newer than the feature compatibility
 * version set.
 *
 * Format:
 * {
 *   setFeatureCompatibilityVersion: <string version>
 * }
 */
class SetFeatureCompatibilityVersionCommand : public BasicCommand {
public:
    SetFeatureCompatibilityVersionCommand()
        : BasicCommand(SetFeatureCompatibilityVersion::kCommandName) {}

    AllowedOnSecondary secondaryAllowed(ServiceContext*) const override {
        return AllowedOnSecondary::kNever;
    }

    virtual bool adminOnly() const {
        return true;
    }

    virtual bool supportsWriteConcern(const BSONObj& cmd) const override {
        return true;
    }

    std::string help() const override {
        std::stringstream h;
        h << "Set the featureCompatibilityVersion used by this cluster. If set to '"
          << FCVP::kLastLTS << "', then features introduced in versions greater than '"
          << FCVP::kLastLTS << "' will be disabled";
        if (FCVP::kLastContinuous != FCVP::kLastLTS) {
            h << " If set to '" << FCVP::kLastContinuous << "', then features introduced in '"
              << FCVP::kLatest << "' will be disabled.";
        }
        h << " If set to '" << FCVP::kLatest << "', then '" << FCVP::kLatest
          << "' features are enabled, and all nodes in the cluster must be binary version "
          << FCVP::kLatest << ". See "
          << feature_compatibility_version_documentation::kCompatibilityLink << ".";
        return h.str();
    }

    Status checkAuthForCommand(Client* client,
                               const std::string& dbname,
                               const BSONObj& cmdObj) const override {
        if (!AuthorizationSession::get(client)->isAuthorizedForActionsOnResource(
                ResourcePattern::forClusterResource(),
                ActionType::setFeatureCompatibilityVersion)) {
            return Status(ErrorCodes::Unauthorized, "Unauthorized");
        }
        return Status::OK();
    }

    bool run(OperationContext* opCtx,
             const std::string& dbname,
             const BSONObj& cmdObj,
             BSONObjBuilder& result) {
        // Always wait for at least majority writeConcern to ensure all writes involved in the
        // upgrade process cannot be rolled back. There is currently no mechanism to specify a
        // default writeConcern, so we manually call waitForWriteConcern upon exiting this command.
        //
        // TODO SERVER-25778: replace this with the general mechanism for specifying a default
        // writeConcern.
        ON_BLOCK_EXIT([&] {
            // Propagate the user's wTimeout if one was given.
            auto timeout =
                opCtx->getWriteConcern().usedDefault ? INT_MAX : opCtx->getWriteConcern().wTimeout;
            WriteConcernResult res;
            auto waitForWCStatus = waitForWriteConcern(
                opCtx,
                repl::ReplClientInfo::forClient(opCtx->getClient()).getLastOp(),
                WriteConcernOptions(repl::ReplSetConfig::kMajorityWriteConcernModeName,
                                    WriteConcernOptions::SyncMode::UNSET,
                                    timeout),
                &res);
            CommandHelpers::appendCommandWCStatus(result, waitForWCStatus, res);
        });

        // Ensure that this operation will be killed by the RstlKillOpThread during step-up or
        // stepdown.
        opCtx->setAlwaysInterruptAtStepDownOrUp();

        // Only allow one instance of setFeatureCompatibilityVersion to run at a time.
        invariant(!opCtx->lockState()->isLocked());
        Lock::ExclusiveLock lk(opCtx->lockState(), FeatureCompatibilityVersion::fcvLock);

        auto request = SetFeatureCompatibilityVersion::parse(
            IDLParserErrorContext("setFeatureCompatibilityVersion"), cmdObj);
        const auto requestedVersion = request.getCommandParameter();
        const auto requestedVersionString = FCVP::serializeVersion(requestedVersion);
        FeatureCompatibilityParams::Version actualVersion =
            serverGlobalParams.featureCompatibility.getVersion();
        if (request.getDowngradeOnDiskChanges() &&
            (requestedVersion != FeatureCompatibilityParams::kLastContinuous ||
             actualVersion < requestedVersion)) {
            std::stringstream downgradeOnDiskErrorSS;
            downgradeOnDiskErrorSS
                << "cannot set featureCompatibilityVersion to " << requestedVersionString
                << " with '" << SetFeatureCompatibilityVersion::kDowngradeOnDiskChangesFieldName
                << "' set to true. This is only allowed when downgrading to "
                << FCVP::kLastContinuous;
            uasserted(ErrorCodes::IllegalOperation, downgradeOnDiskErrorSS.str());
        }

        if (actualVersion == requestedVersion) {
            // Set the client's last opTime to the system last opTime so no-ops wait for
            // writeConcern.
            repl::ReplClientInfo::forClient(opCtx->getClient()).setLastOpToSystemLastOpTime(opCtx);
            return true;
        }

        auto isFromConfigServer = request.getFromConfigServer().value_or(false);
        FeatureCompatibilityVersion::validateSetFeatureCompatibilityVersionRequest(
            actualVersion, requestedVersion, isFromConfigServer);

        if (actualVersion < requestedVersion) {
            FeatureCompatibilityVersion::updateFeatureCompatibilityVersionDocument(
                opCtx, actualVersion, requestedVersion, isFromConfigServer);
            {
                // Take the global lock in S mode to create a barrier for operations taking the
                // global IX or X locks. This ensures that either
                //   - The global IX/X locked operation will start after the FCV change, see the
                //     upgrading to the latest FCV and act accordingly.
                //   - The global IX/X locked operation began prior to the FCV change, is acting on
                //     that assumption and will finish before upgrade procedures begin right after
                //     this.
                Lock::GlobalLock lk(opCtx, MODE_S);
            }

            if (failUpgrading.shouldFail())
                return false;

            if (serverGlobalParams.clusterRole == ClusterRole::ShardServer) {
                const auto shardingState = ShardingState::get(opCtx);
                if (shardingState->enabled()) {
                    LOGV2(20500, "Upgrade: submitting orphaned ranges for cleanup");
                    migrationutil::submitOrphanRangesForCleanup(opCtx);
                }
            }

            // Delete any haystack indexes if we're upgrading to an FCV of 4.9 or higher.
            // TODO SERVER-51871: This block can removed once 5.0 becomes last-lts.
            if (requestedVersion >= FeatureCompatibilityParams::Version::kVersion49) {
                _deleteHaystackIndexesOnUpgrade(opCtx);
            }

            // Upgrade shards before config finishes its upgrade.
            if (serverGlobalParams.clusterRole == ClusterRole::ConfigServer) {
                uassertStatusOK(
                    ShardingCatalogManager::get(opCtx)->setFeatureCompatibilityVersionOnShards(
                        opCtx, CommandHelpers::appendMajorityWriteConcern(request.toBSON({}))));
            }

            hangWhileUpgrading.pauseWhileSet(opCtx);
            // Completed transition to requestedVersion.
            FeatureCompatibilityVersion::updateFeatureCompatibilityVersionDocument(
                opCtx,
                serverGlobalParams.featureCompatibility.getVersion(),
                requestedVersion,
                isFromConfigServer);
        } else {
            auto replCoord = repl::ReplicationCoordinator::get(opCtx);
            const bool isReplSet =
                replCoord->getReplicationMode() == repl::ReplicationCoordinator::modeReplSet;

            uassert(ErrorCodes::ConflictingOperationInProgress,
                    str::stream() << "Cannot downgrade the cluster when the replica set config "
                                  << "contains 'newlyAdded' members; wait for those members to "
                                  << "finish its initial sync procedure",
                    !(isReplSet && replCoord->replSetContainsNewlyAddedMembers()));

            // We should make sure the current config w/o 'newlyAdded' members got replicated
            // to all nodes.
            LOGV2(4637904, "Waiting for the current replica set config to propagate to all nodes.");
            // If a write concern is given, we'll use its wTimeout. It's kNoTimeout by default.
            WriteConcernOptions writeConcern(repl::ReplSetConfig::kConfigAllWriteConcernName,
                                             WriteConcernOptions::SyncMode::NONE,
                                             opCtx->getWriteConcern().wTimeout);
            writeConcern.checkCondition = WriteConcernOptions::CheckCondition::Config;
            repl::OpTime fakeOpTime(Timestamp(1, 1), replCoord->getTerm());
            uassertStatusOKWithContext(
                replCoord->awaitReplication(opCtx, fakeOpTime, writeConcern).status,
                "Failed to wait for the current replica set config to propagate to all "
                "nodes");
            LOGV2(4637905, "The current replica set config has been propagated to all nodes.");

            FeatureCompatibilityVersion::updateFeatureCompatibilityVersionDocument(
                opCtx, actualVersion, requestedVersion, isFromConfigServer);

            {
                // Take the global lock in S mode to create a barrier for operations taking the
                // global IX or X locks. This ensures that either
                //   - The global IX/X locked operation will start after the FCV change, see the
                //     downgrading to 4.4 FCV and act accordingly.
                //   - The global IX/X locked operation began prior to the FCV change, is acting on
                //     that assumption and will finish before downgrade procedures begin right after
                //     this.
                Lock::GlobalLock lk(opCtx, MODE_S);
            }

            if (failDowngrading.shouldFail())
                return false;

            if (serverGlobalParams.clusterRole == ClusterRole::ShardServer) {
                LOGV2(20502, "Downgrade: dropping config.rangeDeletions collection");
                migrationutil::dropRangeDeletionsCollection(opCtx);
            } else if (isReplSet || serverGlobalParams.clusterRole == ClusterRole::ConfigServer) {
                // The default rwc document should only be deleted on plain replica sets and the
                // config server replica set, not on shards or standalones.
                deletePersistedDefaultRWConcernDocument(opCtx);
            }

            // Downgrade shards before config finishes its downgrade.
            if (serverGlobalParams.clusterRole == ClusterRole::ConfigServer) {
                uassertStatusOK(
                    ShardingCatalogManager::get(opCtx)->setFeatureCompatibilityVersionOnShards(
                        opCtx, CommandHelpers::appendMajorityWriteConcern(request.toBSON({}))));
            }

            hangWhileDowngrading.pauseWhileSet(opCtx);
            // Completed transition to requestedVersion.
            FeatureCompatibilityVersion::updateFeatureCompatibilityVersionDocument(
                opCtx,
                serverGlobalParams.featureCompatibility.getVersion(),
                requestedVersion,
                isFromConfigServer);

            if (request.getDowngradeOnDiskChanges()) {
                invariant(requestedVersion == FeatureCompatibilityParams::kLastContinuous);
                _downgradeOnDiskChanges();
                LOGV2(4875603, "Downgrade of on-disk format complete.");
            }
        }

        return true;
    }

private:
    /*
     * Rolls back any upgraded on-disk changes to reflect the disk format of the last-continuous
     * version.
     */
    void _downgradeOnDiskChanges() {
        LOGV2(4975602,
              "Downgrading on-disk format to reflect the last-continuous version.",
              "last_continuous_version"_attr = FCVP::kLastContinuous);
    }

    /**
     * Removes all haystack indexes from the catalog.
     *
     * TODO SERVER-51871: This method can be removed once 5.0 becomes last-lts.
     */
    void _deleteHaystackIndexesOnUpgrade(OperationContext* opCtx) {
        auto& collCatalog = CollectionCatalog::get(opCtx);
        for (const auto& db : collCatalog.getAllDbNames()) {
            for (auto collIt = collCatalog.begin(opCtx, db); collIt != collCatalog.end(opCtx);
                 ++collIt) {
                NamespaceStringOrUUID collName(
                    collCatalog.lookupNSSByUUID(opCtx, collIt.uuid().get()).get());
                AutoGetCollectionForRead coll(opCtx, collName);
                auto idxCatalog = coll->getIndexCatalog();
                std::vector<const IndexDescriptor*> haystackIndexes;
                idxCatalog->findIndexByType(opCtx, IndexNames::GEO_HAYSTACK, haystackIndexes);

                // Continue if 'coll' has no haystack indexes.
                if (haystackIndexes.empty()) {
                    continue;
                }

                // Construct a dropIndexes command to drop the indexes in 'haystackIndexes'.
                BSONObjBuilder dropIndexesCmd;
                dropIndexesCmd.append("dropIndexes", collName.nss()->coll());
                BSONArrayBuilder indexNames;
                for (auto&& haystackIndex : haystackIndexes) {
                    indexNames.append(haystackIndex->indexName());
                }
                dropIndexesCmd.append("index", indexNames.arr());

                BSONObjBuilder response;  // This response is ignored.
                uassertStatusOK(
                    dropIndexes(opCtx,
                                *collName.nss(),
                                CommandHelpers::appendMajorityWriteConcern(dropIndexesCmd.obj()),
                                &response));
            }
        }
    }

} setFeatureCompatibilityVersionCommand;

}  // namespace
}  // namespace mongo