Distribute Collections Using Zones

On this page

Prerequisites

Scenario
Architecture
Zones
Shard Key
Balancer
Steps
Add each shard to the appropriate zone.
Add zone ranges for the relevant collections.
Shard the collections.
Review the changes.
Learn More

In sharded clusters, you can create zones of sharded data based on the shard key. You can associate each zone with one or more shards in the cluster. A shard can associate with any number of zones. In a balanced cluster, MongoDB migrates chunks covered by a zone only to those shards associated with the zone.

You can use zone sharding to distribute collections across a sharded cluster and designate which shards store data for each collection. You can distribute collections based on shard properties, such as physical resources and available memory, to ensure that each collection is stored on the optimal shard for that data.

Prerequisites

To complete this tutorial, you must:

Deploy a sharded cluster. This tutorial uses a sharded cluster with three shards.
Connect to a mongos. You cannot create zones or zone ranges by connecting directly to a shard.
Authenticate as a user with at least the clusterManager role on the admin database. To view user permissions, use the db.getUser() method.

Scenario

You have a database called shardDistributionDB that contains two sharded collections:

bigData, which contains a large amount of data.
manyIndexes, which contains many large indexes.

You want to limit each collection to a subset of shards so that each collection can use the shards' different physical resources.

Architecture

The sharded cluster has three shards. Each shard has unique physical resources:

Shard Name	Physical Resources
`shard0`	High memory capacity
`shard1`	Fast flash storage
`shard2`	High memory capacity and fast flash storage

Zones

To distribute collections based on physical resources, use shard zones. A shard zone associates collections with a specific subset of shards, which restricts the shards that store the collection's data. In this example, you need two shard zones:

Zone Name	Description	Collections in this Zone
`HI_RAM`	Servers with high memory capacity.	Collections requiring more memory, such as collections with large indexes, should be on the `HI_RAM` shards.
`FLASH`	Servers with flash drives for fast storage speeds.	Large collections requiring fast data retrieval should be on the `FLASH` shards.

Shard Key

In this tutorial, the shard key you will use to shard each collection is { _id: "hashed" }. You will configure shard zones before you shard the collections. As a result, each collection's data only ever exists on the shards in the corresponding zone.

With hashed sharding, if you shard collections before you configure zones, MongoDB assigns chunks evenly between all shards when sharding is enabled. This means that chunks may be temporarily assigned to a shard poorly suited to handle that chunk's data.

Balancer

The balancer migrates chunks to the appropriate shard, respecting any configured zones. When balancing is complete, shards only contain chunks whose ranges match its assigned zones.

Important

Performance

Adding, removing, or changing zones or zone ranges can result in chunk migrations. Depending on the size of your dataset and the number of chunks a zone or zone range affects, these migrations may impact cluster performance. Consider running the balancer during specific scheduled windows. To learn how to set a scheduling window, see Schedule the Balancing Window.

Steps

Use the following procedure to configure shard zones and distribute collections based on shard physical resources.

Add each shard to the appropriate zone.

To configure the shards in each zone, use the addShardToZone command.

Add shard0 and shard2 to the HI_RAM zone:

sh.addShardToZone("shard0", "HI_RAM")
sh.addShardToZone("shard2", "HI_RAM")

Add shard1 and shard2 to the FLASH zone:

sh.addShardToZone("shard1", "FLASH")
sh.addShardToZone("shard2", "FLASH")

Add zone ranges for the relevant collections.

To associate a range of shard keys to a zone, use sh.updateZoneKeyRange().

In this scenario, you want to associate all documents in a collection to the appropriate zone. To associate all collection documents to a zone, specify the following zone range:

a lower bound of { "_id" : MinKey }
an upper bound of { "_id" : MaxKey }

For the bigData collection, set:

The namespace to shardDistributionDB.bigData,
The lower bound to MinKey,
The upper bound to MaxKey,
The zone to FLASH

sh.updateZoneKeyRange(
   "shardDistributionDB.bigData",
   { "_id" : MinKey },
   { "_id" : MaxKey },
   "FLASH"
)

For the manyIndexes collection, set:

The namespace to shardDistributionDB.manyIndexes,
The lower bound to MinKey,
The upper bound to MaxKey,
The zone to HI_RAM

sh.updateZoneKeyRange(
   "shardDistributionDB.manyIndexes",
   { "_id" : MinKey },
   { "_id" : MaxKey },
   "HI_RAM"
)

Shard the collections.

To shard both collections (bigData and manyIndexes), specify a shard key of { _id: "hashed" }.

Run the following commands:

sh.shardCollection(
   "shardDistributionDB.bigData", { _id: "hashed" }
)
sh.shardCollection(
   "shardDistributionDB.manyIndexes", { _id: "hashed" }
)

Review the changes.

To view chunk distribution and shard zones, use the sh.status() method:

sh.status()

The next time the balancer runs, it splits chunks where necessary and migrates chunks across the shards, respecting the configured zones. The amount of time the balancer takes to complete depends on several factors, including number of shards, available memory, and IOPS.

When balancing finishes:

Chunks for documents in the manyIndexes collection reside on shard0 and shard2
Chunks for documents in the bigData collection reside on shard1 and shard2.

Learn More

To learn more about sharding and balancing, see the following pages:

Back

Distributed Local Writes for Insert-Only Workloads

Data Partitioning