Navigation

Remove Shards from an Existing Sharded Cluster

To remove a shard you must ensure the shard’s data is migrated to the remaining shards in the cluster. This procedure describes how to safely migrate data and how to remove a shard.

When you remove a shard in a cluster with an uneven chunk distribution, the balancer first removes the chunks from the draining shard and then balances the remaining uneven chunk distribution.

This procedure describes how to safely remove a single shard. Do not use this procedure to migrate an entire cluster to new hardware. To migrate an entire shard to new hardware, migrate individual shards as if they were independent replica sets.

To remove a shard, first connect to one of the cluster’s mongos instances using mongo shell. Then use the sequence of tasks in this document to remove a shard from the cluster.

Considerations

A shard removal may cause an open change stream cursor to close, and the closed change stream cursor may not be fully resumable.

Ensure the Balancer Process is Enabled

To successfully migrate data from a shard, the balancer process must be enabled. Check the balancer state using the sh.getBalancerState() helper in the mongo shell. For more information, see the section on balancer operations.

Determine the Name of the Shard to Remove

To determine the name of the shard, connect to a mongos instance with the mongo shell and either:

The shards._id field lists the name of each shard.

Remove Chunks from the Shard

From the admin database, run the removeShard command. This begins “draining” chunks from the shard you are removing to other shards in the cluster. For example, for a shard named mongodb0, run:

db.adminCommand( { removeShard: "mongodb0" } )

mongos converts the write concern of the removeShard command to "majority".

This operation returns with the following response:

{
   "msg" : "draining started successfully",
   "state" : "started",
   "shard" : "mongodb0",
   "note" : "you need to drop or movePrimary these databases",
   "dbsToMove" : [
      "fiz",
      "buzz"
   ],
   "ok" : 1,
   "operationTime" : Timestamp(1575398919, 2),
   "$clusterTime" : {
      "clusterTime" : Timestamp(1575398919, 2),
      "signature" : {
         "hash" : BinData(0,"Oi68poWCFCA7b9kyhIcg+TzaGiA="),
         "keyId" : NumberLong("6766255701040824328")
      }
   }
}

The balancer begins migrating chunks from the shard named mongodb0 to other shards in the cluster. These migrations happens slowly to avoid placing undue load on the overall cluster. Depending on your network capacity and the amount of data, this operation can take from a few minutes to several days to complete.

Note

The output includes the field dbsToMove indicating the databases, if any, for which the shard is the primary shard. After all chunks have been drained from the shard, you must either movePrimary for the database(s) or alternatively, drop the databases (which deletes the associated data files).

Check the Status of the Migration

To check the progress of the migration at any stage in the process, run removeShard from the admin database again. For example, for a shard named mongodb0, run:

db.adminCommand( { removeShard: "mongodb0" } )

mongos converts the write concern of the removeShard command to "majority".

The command returns output similar to the following:

{
   "msg" : "draining ongoing",
   "state" : "ongoing",
   "remaining" : {
      "chunks" : NumberLong(2),
      "dbs" : NumberLong(2),
      "jumboChunks" : NumberLong(0)         // Available starting in 4.0.14
   },
   "note" : "you need to drop or movePrimary these databases",
   "dbsToMove" : [
      "fizz",
      "buzz"
   ],
   "ok" : 1,
   "operationTime" : Timestamp(1575399086, 1655),
   "$clusterTime" : {
      "clusterTime" : Timestamp(1575399086, 1655),
      "signature" : {
         "hash" : BinData(0,"XBrTmjMMe82fUtVLRm13GBVtRE8="),
         "keyId" : NumberLong("6766255701040824328")
      }
   }
}

In the output, the remaining field includes the following fields:

chunks Total number of chunks currently remaining on the shard.
dbs Total number of databases whose primary shard is the shard. These databases are specified in the dbsToMove output field.
jumboChunks

Of the total number of chunks, the number that are jumbo.

If the jumboChunks is greater than 0, wait until only the jumboChunks remain on the shard. Once only the jumbo chunks remain, you must manually clear the jumbo flag before the draining can complete. See Clear jumbo Flag.

New in version 4.0.14.

Continue checking the status of the removeShard command until the number of chunks remaining is 0.

db.adminCommand( { removeShard: "mongodb0" } )

Move Databases to Another Primary Shard

If the shard is the primary shard for one or more databases in the cluster, then you must make that database use a different shard as its primary shard. removeShard lists any databases that you need to move in the dbsToMove field in the command output. If the shard is not the primary shard for any databases, skip to the next task, Finalize the Migration.

Warning

Do not perform this procedure until you have finished draining the shard.

To move a database to another shard, use the movePrimary command.

Important

To ensure a smooth migration, refer to the considerations in the movePrimary command documentation before running movePrimary.

To migrate the fizz database from mongodb0 to mongodb1, issue the following command:

db.adminCommand( { movePrimary: "fizz", to: "mongodb1" })

mongos uses "majority" write concern for movePrimary.

This command does not return until MongoDB completes moving all data. The response from this command will resemble the following:

{
   "ok" : 1,
   "operationTime" : Timestamp(1575400369, 9),
   "$clusterTime" : {
   "clusterTime" : Timestamp(1575400369, 9),
   "signature" : {
      "hash" : BinData(0,"2Nz8QCcVXB0LJLm1hsXfpTCaM0M="),
      "keyId" : NumberLong("6766255701040824328")
   }
}
}

Warning

If you use the movePrimary command to move un-sharded collections, you must either restart all mongos instances, or use the flushRouterConfig command on all mongos instances before reading or writing any data to any unsharded collections that were moved. This action ensures that the mongos is aware of the new shard for these collections.

If you do not update the mongos instances’ metadata cache after using movePrimary, the mongos may miss data on reads, and may not write data to the correct shard. To recover, you must manually intervene.

Finalize the Migration

To clean up all metadata information and finalize the removal, run removeShard again. For example, for a shard named mongodb0, run:

db.adminCommand( { removeShard: "mongodb0" } )

mongos converts the write concern of the removeShard command to "majority".

A success message appears at completion:

{
    "msg" : "removeshard completed successfully",
    "state" : "completed",
    "shard" : "mongodb0",
    "ok" : 1,
    "operationTime" : Timestamp(1575400370, 2),
    "$clusterTime" : {
       "clusterTime" : Timestamp(1575400370, 2),
       "signature" : {
          "hash" : BinData(0,"JjSRciHECXDBXo0e5nJv9mdRG8M="),
          "keyId" : NumberLong("6766255701040824328")
       }
    }
}

Once the value of the state field is “completed”, you may safely stop the instances comprising the mongodb0 shard.