checkMetadataConsistency

~~On this page~~本页内容

Definition
Syntax
Command Fields
Output
Behavior
Batch Results
Check Indexes
Example

Definition

checkMetadataConsistency

Performs a series of consistency checks on sharding metadata for a cluster, database, or collection. The command returns a cursor with either all or a batch of the inconsistency results found.

Run this command after major maintenance operations, such as upgrades and downgrades, to check the state of the catalog.

By default, the command does not check indexes for consistency across the shards. To check indexes, set the checkIndexes option.

New in version 7.0.

Syntax

To check the entire cluster for sharding metadata inconsistencies, run the command from the admin database.
```
db.getSiblingDB("admin").runCommand( {
   checkMetadataConsistency: 1
} )
```
To check the database for sharding metadata inconsistencies, run the command from the database context:
```
use cars
db.runCommand( {
   checkMetadataConsistency: 1
} )
```
To check a collection for sharding metadata inconsistencies, run the command with the coll option:
```
use library
db.runCommand( {
   checkMetadataConsistency: 1,
   coll: "authors",
} )
```

Command Fields

Field	Type	Description
`checkIndexes`	boolean	Sets whether the command also checks indexes in sharding metadata. For more information, see Check Indexes.
`coll`	string	Sets the collection to check for sharding metadata inconsistencies.
`cursor`	document	Configures the return cursor.
`cursor.batchSize`	integer	Maximum number of inconsistency results to include in each batch.

Output

The checkMetadataConsistency command returns a cursor with a document for each inconsistency found in sharding metadata.

The return document has the following fields:

Field	Type	Description
`cursor`	document	Cursor with the results of the inconsistency checks.
`cursor.id`	integer	A 64-bit integer indicated the cursor ID. Use the `cursor.id` value with the `getMore` command to retrieve the next batch of inconsistencies. If the cursor returns an ID of `0`, it indicates that there are no more batches of information.
`cursor.ns`	string	The database and collection checked for inconsistencies.
`cursor.firstBatch`	array	Results of metadata consistency checks.
`ok`	boolean	Indicates whether the command was successful.

Behavior

Batch Results

The checkMetadataConsistency command returns results in batches. To customize the batch size, the batchSize option:

var cur = db.runCommand( {
   checkMetadataConsistency: 1,
   cursor: {
      batchSize: 10
   }
} )

If the cursor.id field is greater than 0, you can use with the getMore command to retrieve the next batch of results.

Check Indexes

The checkMetadataConsistency command does not check indexes by default. To check metadata consistency and indexes, use the checkIndexes option:

db.runCommand( {
   checkMetadataConsistency: 1,
   checkIndexes: true
} )

Example

Use runCommand() to run the checkMetadataConsistency command:

db.runCommand( { checkMetadataConsistency: 1 } )

Example Output:

{
   cursor: {
      id: Long("0"),
      ns: "test.$cmd.aggregate",
      firstBatch: [
         {
            type: "MisplacedCollection",
            description: "Unsharded collection found on shard different from database primary shard",
            details: {
               namespace: "test.authors",
               shard: "shard02",
               localUUID: new UUID("1ad56770-61e2-48e9-83c6-8ecefe73cfc4")
            }
         }
      ],
   },
   ok: 1
}

← balancerStop checkShardingIndex →