Navigation

db.collection.getShardDistribution()

On this page

Definition

db.collection.getShardDistribution()

Prints the data distribution statistics for a sharded collection.

Tip

Before running the method, use the flushRouterConfig command to refresh the cached routing table to avoid returning stale distribution information for the collection. Once refreshed, run db.collection.getShardDistribution() for the collection you wish to build the index.

For example:

db.adminCommand( { flushRouterConfig: "test.myShardedCollection" } );
db.getSiblingDB("test").myShardedCollection.getShardDistribution();

See also

Sharding

Output

Sample Output

The following is a sample output for the distribution of a sharded collection:

Shard shard-a at shard-a/MyMachine.local:30000,MyMachine.local:30001,MyMachine.local:30002
data : 38.14Mb docs : 1000003 chunks : 2
estimated data per chunk : 19.07Mb
estimated docs per chunk : 500001

Shard shard-b at shard-b/MyMachine.local:30100,MyMachine.local:30101,MyMachine.local:30102
data : 38.14Mb docs : 999999 chunks : 3
estimated data per chunk : 12.71Mb
estimated docs per chunk : 333333

Totals
data : 76.29Mb docs : 2000002 chunks : 5
Shard shard-a contains 50% data, 50% docs in cluster, avg obj size on shard : 40b
Shard shard-b contains 49.99% data, 49.99% docs in cluster, avg obj size on shard : 40b

Output Fields

Shard <shard-a> at <host-a>
 data : <size-a> docs : <count-a> chunks : <number of chunks-a>
 estimated data per chunk : <size-a>/<number of chunks-a>
 estimated docs per chunk : <count-a>/<number of chunks-a>

Shard <shard-b> at <host-b>
 data : <size-b> docs : <count-b> chunks : <number of chunks-b>
 estimated data per chunk : <size-b>/<number of chunks-b>
 estimated docs per chunk : <count-b>/<number of chunks-b>

Totals
 data : <stats.size> docs : <stats.count> chunks : <calc total chunks>
 Shard <shard-a> contains  <estDataPercent-a>% data, <estDocPercent-a>% docs in cluster, avg obj size on shard : stats.shards[ <shard-a> ].avgObjSize
 Shard <shard-b> contains  <estDataPercent-b>% data, <estDocPercent-b>% docs in cluster, avg obj size on shard : stats.shards[ <shard-b> ].avgObjSize

The output information displays:

  • <shard-x> is a string that holds the shard name.

  • <host-x> is a string that holds the host name(s).

  • <size-x> is a number that includes the size of the data, including the unit of measure (e.g. b, Mb).

  • <count-x> is a number that reports the number of documents in the shard.

  • <number of chunks-x> is a number that reports the number of chunks in the shard.

  • <size-x>/<number of chunks-x> is a calculated value that reflects the estimated data size per chunk for the shard, including the unit of measure (e.g. b, Mb).

  • <count-x>/<number of chunks-x> is a calculated value that reflects the estimated number of documents per chunk for the shard.

  • <stats.size> is a value that reports the total size of the data in the sharded collection, including the unit of measure.

  • <stats.count> is a value that reports the total number of documents in the sharded collection.

  • <calc total chunks> is a calculated number that reports the number of chunks from all shards, for example:

    <calc total chunks> = <number of chunks-a> + <number of chunks-b>
    
  • <estDataPercent-x> is a calculated value that reflects, for each shard, the data size as the percentage of the collection’s total data size, for example:

    <estDataPercent-x> = <size-x>/<stats.size>
    
  • <estDocPercent-x> is a calculated value that reflects, for each shard, the number of documents as the percentage of the total number of documents for the collection, for example:

    <estDocPercent-x> = <count-x>/<stats.count>
    
  • stats.shards[ <shard-x> ].avgObjSize is a number that reflects the average object size, including the unit of measure, for the shard.