moveChunk¶

On this page

Definition
Considerations
Behavior

Definition¶

moveChunk¶

Internal administrative command. Moves chunks between shards. Issue the moveChunk command via a mongos instance while using the admin database. Use the following forms:

db.runCommand( { moveChunk : <namespace> ,
                 find : <query> ,
                 to : <string>,
                 _secondaryThrottle : <boolean>,
                 writeConcern: <document>,
                 _waitForDelete : <boolean> } )

Alternately:

db.runCommand( { moveChunk : <namespace> ,
                 bounds : <array> ,
                 to : <string>,
                 _secondaryThrottle : <boolean>,
                 writeConcern: <document>,
                 _waitForDelete : <boolean> } )

The moveChunk command has the following fields:

Field	Type	Description
`moveChunk`	string	The namespace of the collection where the chunk exists. Specify the collection’s full namespace, including the database name.
`find`	document	An equality match on the shard key that specifies the shard-key value of the chunk to move. Specify either the `bounds` field or the `find` field but not both. Do not use the `find` field to select chunks in collections that use a hashed shard key.
`bounds`	array	The bounds of a specific chunk to move. The array must consist of two documents that specify the lower and upper shard key values of a chunk to move. Specify either the `bounds` field or the `find` field but not both. Use `bounds` to select chunks in collections that use a hashed shard key.
`to`	string	The name of the destination shard for the chunk.
`_secondaryThrottle`	boolean	Optional. Starting in MongoDB 3.4, for WiredTiger, defaults to `false`. For MMAPv1, the default value remains `true`. If `true`, then by default, each document move during chunk migration propagates to at least one secondary before the balancer proceeds with the next document. This is equivalent to a write concern of `{ w: 2 }`. See also the `writeConcern` parameter. If `false`, the balancer does not wait for replication to a secondary and instead continues with the next document. For more information, see Change Replication Behavior for Chunk Migration.
`writeConcern`	document	Optional. A document that expresses the write concern that the `_secondaryThrottle` will use to wait for secondaries during the chunk migration. Any specified `writeConcern` implies `_secondaryThrottle` and will take precedent over a contradictory `_secondaryThrottle` setting.
`_waitForDelete`	boolean	Optional. Internal option for testing purposes. The default is `false`. If set to `true`, the delete phase of a `moveChunk` operation blocks.

The value of bounds takes the form:

[ { hashedField : <minValue> } ,
  { hashedField : <maxValue> } ]

The chunk migration section describes how chunks move between shards on MongoDB.

参见

split, sh.moveChunk(), sh.splitAt(), and sh.splitFind().

Considerations¶

Only use the moveChunk in special circumstances such as preparing your sharded cluster for an initial ingestion of data, or a large bulk import operation. In most cases allow the balancer to create and balance chunks in sharded clusters. See 在集群中创建数据块 for more information.

Behavior¶

Indexes¶

在 3.0.0 版更改: In previous versions, moveChunk would build indexes as part of the migrations.

moveChunk requires that all indexes exist on the target (i.e. to ) shard before migration and returns an error if a required index does not exist.

Meta Data Error¶

moveChunk returns the following error message if another metadata operation is in progress on the chunks collection:

errmsg: "The collection's metadata lock is already taken."

If another process, such as a balancer process, changes meta data while moveChunk is running, you may see this error. You may retry the moveChunk operation without side effects.

← medianKey movePrimary →