summaryrefslogtreecommitdiff
path: root/src/mongo/db/s/config/configsvr_commit_chunk_migration_command.cpp
blob: 10da0444a4ad2f72972e3900c9f2baa96664ef1c (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
/**
 *    Copyright (C) 2016 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.
 */

#define MONGO_LOG_DEFAULT_COMPONENT ::mongo::logger::LogComponent::kSharding

#include "mongo/platform/basic.h"

#include "mongo/base/status_with.h"
#include "mongo/db/auth/authorization_session.h"
#include "mongo/db/commands.h"
#include "mongo/db/concurrency/d_concurrency.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/repl/read_concern_args.h"
#include "mongo/db/s/chunk_move_write_concern_options.h"
#include "mongo/db/s/sharding_state.h"
#include "mongo/rpc/get_status_from_command_result.h"
#include "mongo/s/catalog/sharding_catalog_manager.h"
#include "mongo/s/catalog/type_chunk.h"
#include "mongo/s/catalog/type_locks.h"
#include "mongo/s/chunk_version.h"
#include "mongo/s/client/shard_registry.h"
#include "mongo/s/grid.h"
#include "mongo/s/request_types/commit_chunk_migration_request_type.h"
#include "mongo/util/fail_point.h"
#include "mongo/util/fail_point_service.h"
#include "mongo/util/log.h"

namespace mongo {
namespace {

MONGO_FP_DECLARE(migrationCommitError);  // delete this in 3.5

/**
 * This command takes the chunk being migrated ("migratedChunk") and generates a new version for it
 * that is written along with its new shard location ("toShard") to the chunks collection. It also
 * takes a control chunk ("controlChunk") and assigns it a new version as well so that the source
 * ("fromShard") shard's shardVersion will increases. If there is no control chunk, then the chunk
 * being migrated is the source shard's only remaining chunk.
 *
 * The new chunk version is generated by querying the highest chunk version of the collection, and
 * then incrementing that major value for both migrated and control chunks and setting the minor to
 * 0 for the migrated chunk and 1 for the control chunk. A global exclusive lock is held for the
 * duration of generating the new chunk version and writing to the chunks collection so that
 * yielding cannot occur. This assures that generated ChunkVersions are strictly monotonically
 * increasing -- a second process will not be able to query for max chunk version until the first
 * finishes writing the new highest chunk version it generated.
 *
 * Command Format:
 * {
 *   _configsvrCommitChunkMigration: <database>.<collection>,
 *   fromShard: "<from_shard_name>",
 *   toShard: "<to_shard_name>",
 *   migratedChunk: {min: <min_value>, max: <max_value>, etc. },
 *   controlChunk: {min: <min_value>, max: <max_value>, etc. },  (optional)
 *   fromShardCollectionVersion: { shardVersionField: <version> }, (for backward compatibility only)
 *   shardHasDistributedLock: true/false (for backward compatibility only)
 * }
 *
 * Returns:
 * {
 *   migratedChunkVersion: <ChunkVersion_BSON>,
 *   controlChunkVersion: <ChunkVersion_BSON>, (only present if a controlChunk is defined)
 * }
 *
 */
class ConfigSvrCommitChunkMigrationCommand : public Command {
public:
    ConfigSvrCommitChunkMigrationCommand() : Command("_configsvrCommitChunkMigration") {}

    void help(std::stringstream& help) const override {
        help << "should not be calling this directly";
    }

    bool slaveOk() const override {
        return false;
    }

    bool adminOnly() const override {
        return true;
    }

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

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

    std::string parseNs(const std::string& dbname, const BSONObj& cmdObj) const override {
        return parseNsFullyQualified(dbname, cmdObj);
    }

    /**
     * Assures that the balancer still holds the collection distributed lock for this collection. If
     * it no longer does, fail because we don't know if the collection state has changed -- e.g.
     * whether it was/is dropping, whether another imcompatible migration is running, etc..
     */
    static Status checkBalancerHasDistLock(OperationContext* txn,
                                           const NamespaceString& nss,
                                           const ChunkType& chunk) {
        auto balancerDistLockProcessID =
            Grid::get(txn)->catalogClient(txn)->getDistLockManager()->getProcessID();

        // Must use local read concern because we're going to perform subsequent writes.
        auto lockQueryResponseWith =
            Grid::get(txn)->shardRegistry()->getConfigShard()->exhaustiveFindOnConfig(
                txn,
                ReadPreferenceSetting{ReadPreference::PrimaryOnly},
                repl::ReadConcernLevel::kLocalReadConcern,
                NamespaceString(LocksType::ConfigNS),
                BSON(LocksType::process(balancerDistLockProcessID) << LocksType::name(nss.ns())),
                BSONObj(),
                boost::none);
        if (!lockQueryResponseWith.isOK()) {
            return lockQueryResponseWith.getStatus();
        }

        invariant(lockQueryResponseWith.getValue().docs.size() <= 1);

        if (MONGO_FAIL_POINT(migrationCommitError)) {
            lockQueryResponseWith.getValue().docs.clear();
        }

        if (lockQueryResponseWith.getValue().docs.size() != 1) {
            return Status(
                ErrorCodes::BalancerLostDistributedLock,
                str::stream() << "The distributed lock for collection '" << nss.ns()
                              << "' was lost by the balancer since this migration began. Cannot "
                              << "proceed with the migration commit for chunk ("
                              << chunk.getRange().toString()
                              << ") because it could corrupt other operations.");
        }
        return Status::OK();
    }

    bool run(OperationContext* txn,
             const std::string& dbName,
             BSONObj& cmdObj,
             int options,
             std::string& errmsg,
             BSONObjBuilder& result) override {

        const NamespaceString nss = NamespaceString(parseNs(dbName, cmdObj));

        auto commitRequest =
            uassertStatusOK(CommitChunkMigrationRequest::createFromCommand(nss, cmdObj));

        if (!commitRequest.shardHasDistributedLock()) {
            auto check = checkBalancerHasDistLock(txn, nss, commitRequest.getMigratedChunk());
            if (!check.isOK()) {
                return appendCommandStatus(result, check);
            }
        }

        StatusWith<BSONObj> response = Grid::get(txn)->catalogManager()->commitChunkMigration(
            txn,
            nss,
            commitRequest.getMigratedChunk(),
            commitRequest.getControlChunk(),
            commitRequest.getCollectionEpoch(),
            commitRequest.getFromShard(),
            commitRequest.getToShard());
        if (!response.isOK()) {
            return appendCommandStatus(result, response.getStatus());
        }
        result.appendElements(response.getValue());
        return true;
    }

} configsvrCommitChunkMigrationCommand;

}  // namespace
}  // namespace mongo