- Deploy Resources >
- Deploy a Sharded Cluster
Deploy a Sharded Cluster¶
On this page
Supported with Cloud Manager and Ops Manager 4.0
You can use Kubernetes Operator to deploy MongoDB instances with Ops Manager version 4.0 or later and Cloud Manager. At any place in this guide that says Ops Manager, you can substitute Cloud Manager.
Sharded clusters provide horizontal scaling for large data sets and enable high throughput operations by distributing the data set across a group of servers.
To learn more about sharding, see Sharding Introduction in the MongoDB manual.
Use this procedure to deploy a new sharded cluster that Ops Manager manages. Later, you can use Ops Manager to add shards and perform other maintenance operations on the cluster.
Prerequisites¶
To deploy a sharded cluster using an object, you need to complete the following procedures:
Considerations¶
Starting in MongoDB Enterprise Kubernetes Operator version 1.3.0, you can only have one MongoDB resource per project. To learn how to migrate your project to a single-cluster configuration, see Migrate to One Resource per Project (Required for Version 1.3.0).
Do Not Deploy Monitoring Agents inside and outside Kubernetes¶
Do not mix MongoDB deployments outside Kubernetes with ones insider Kubernetes in the same Project.
Due to Kubernetes network translation, a Monitoring Agent outside Kubernetes cannot monitor MongoDB instances inside Kubernetes. For this reason, k8s and non-k8s deployments in the same Project is not supported. Use separate projects.
Procedure¶
The procedure for deploying a sharded cluster depends on whether you require the deployment to run with TLS enabled for intra-cluster communication and clients connecting to the database:
Copy the highlighted section of this sharded cluster resource.¶
Change the highlighted settings of this YAML file to match your desired sharded cluster configuration.
This is a YAML file that you can modify to meet your desired configuration. Change the highlighted settings to match your desired sharded cluster configuration.
Paste the copied example to create a new sharded cluster resource.¶
Open your preferred text editor and paste the object specification into a new text file.
Configure the settings highlighted in the preceding step as follows.¶
Key | Type | Description | Example |
---|---|---|---|
metadata.name |
string | Label for this Kubernetes sharded cluster object. See also
|
myproject |
metadata.namespace |
string | Scope of object names. Kubernetes namespace where this MongoDB Kubernetes resource and other objects are created. Using two different namespaces allows you to delete your sharded cluster or all of the resources in the namespace without affecting your Kubernetes Operator. See also
|
mongodb |
spec.shardCount |
integer | Number of shards to deploy. | 2 |
spec.mongodsPerShardCount |
integer | Number of shard members per shard. | 3 |
spec.mongosCount |
integer | Number of shard routers to deploy. | 2 |
spec.configServerCount |
integer | Number of members of the config server replica set. | 3 |
spec.version |
string | Version of MongoDB that this sharded cluster should run. The format should be To learn more about MongoDB versioning, see MongoDB Versioning in the MongoDB Manual. |
3.6.7 |
spec.opsManager.configMapRef.name |
string | Name of the ConfigMap with the Ops Manager connection
configuration. The
Value must match namespace and name of ConfigMap This value must match the value you provided for
Operator manages changes to the ConfigMap The Kubernetes Operator tracks any changes to the ConfigMap and reconciles the state of the MongoDB Kubernetes resource. |
<myproject> |
spec.credentials |
string | Name of the Kubernetes secret you created as Ops Manager API authentication credentials for the Kubernetes Operator to communicate with Ops Manager. Value must use namespace and name of Secret This value must match the value you provided for
Operator manages changes to the Secret The Kubernetes Operator tracks any changes to the Secret and reconciles the state of the MongoDB Kubernetes resource. |
<mycredentials> |
spec.type |
string | Type of MongoDB Kubernetes resource to create. | ShardedCluster |
spec.persistent |
string | Optional. Flag indicating if this MongoDB Kubernetes resource should use Persistent Volumes for storage. Persistent volumes are not deleted when the MongoDB Kubernetes resource is stopped or restarted. If this value is To change your Persistent Volume Claims configuration, configure the following collections to meet your deployment requirements:
Warning Your containers must have permissions to write to your Persistent Volume.
The Kubernetes Operator sets Note If you do not use Persistent Volumes, the Disk Usage and Disk IOPS charts cannot be displayed in either the Processes tab on the Deployment page or in the Metrics page when reviewing the data for this deployment. |
true |
Add any additional accepted settings for a sharded cluster deployment.¶
You can also add any of the following optional settings to the object specification file for a sharded cluster deployment:
Warning
You must set spec.clusterName
if your Kubernetes cluster has a
default domain
different from default cluster.local
. If you neither use the
default nor set this option, the Kubernetes Operator may not function as
expected.
For config server
spec.configSrvPodSpec.cpu
spec.configSrvPodSpec.cpuRequests
spec.configSrvPodSpec.memory
spec.configSrvPodSpec.memoryRequests
spec.configSrvPodSpec.persistence.single
spec.configSrvPodSpec.persistence.multiple.data
spec.configSrvPodSpec.persistence.multiple.journal
spec.configSrvPodSpec.persistence.multiple.logs
spec.configSrvPodSpec.nodeAffinity
spec.configSrvPodSpec.podAffinity
spec.configSrvPodSpec.podAntiAffinityTopologyKey
For shard routers
spec.mongosPodSpec.cpu
spec.mongosPodSpec.cpuRequests
spec.mongosPodSpec.memory
spec.mongosPodSpec.memoryRequests
spec.mongosPodSpec.nodeAffinity
spec.mongosPodSpec.podAffinity
spec.mongosPodSpec.podAntiAffinityTopologyKey
For shard members
spec.shardPodSpec.cpu
spec.shardPodSpec.cpuRequests
spec.shardPodSpec.memory
spec.shardPodSpec.memoryRequests
spec.shardPodSpec.nodeAffinity
spec.shardPodSpec.persistence.single
spec.shardPodSpec.persistence.multiple.data
spec.shardPodSpec.persistence.multiple.journal
spec.shardPodSpec.persistence.multiple.logs
spec.shardPodSpec.podAffinity
spec.shardPodSpec.podAntiAffinityTopologyKey
Save this file with a .yaml
file extension.¶
Start your sharded cluster deployment.¶
Invoke the following Kubernetes command to create your sharded cluster:
Check the log after running this command. If the creation was successful, you should see a message similar to the following:
Track the status of your sharded cluster deployment.¶
To check the status of your MongoDB Kubernetes resource, invoke the following command:
The -w
flag means “watch”. With the “watch” flag set, the output
refreshes immediately when something changes until the status phase
achieves the Running
state.
See Troubleshooting the Kubernetes Operator for information about the resource deployment statuses.