Redis Enterprise clusters (REC)

Create and manage Redis Enterprise clusters (REC) on Kubernetes using the Redis Enterprise operator.

Redis Enterprise for Kubernetes

A Redis Enterprise cluster (REC) is a custom Kubernetes resource that represents a Redis Enterprise cluster deployment. The Redis Enterprise operator manages the lifecycle of REC resources, including deployment, scaling, upgrades, and recovery operations.

REC resources define the cluster configuration, including node specifications, storage requirements, security settings, and networking configuration. After you deploy the cluster, it provides a foundation for creating and managing Redis Enterprise databases (REDB).

Cluster management

Manage your Redis Enterprise cluster lifecycle and configuration:

Disconnect clients on password removal

Redis Software 8.0.2 and later can actively disconnect client connections that authenticated with a removed, revoked, or rotated password. In Redis Software, you configure this setting through the REST API. See the cluster object and database (bdb) object. Starting with operator version 8.2.6-21, the operator configures it through the custom resources instead:

  • Cluster-wide policy: spec.disconnectClientsOnPasswordRemoval on the REC (Enabled, Disabled, or Auto).
  • Per-database setting: spec.disconnectClientsOnPasswordRemoval on the REDB (boolean; the Redis Software default is false). Takes effect only when the cluster-wide policy is Auto.
  • Active-Active databases: spec.globalConfigurations.disconnectClientsOnPasswordRemoval on the REAADB, propagated to all participating clusters. See Set global database configurations. Takes effect when the cluster-wide policy on the participating clusters is Auto.

Set the cluster-wide policy in your REC specification:

spec:
  disconnectClientsOnPasswordRemoval: Enabled

Allowed values are Enabled, Disabled, and Auto. Auto (the Redis Software default) defers the decision to each database's own setting. Enabled and Disabled force the behavior for all databases in the cluster, ignoring the per-database setting.

Whether the operator owns the setting depends on your upgrade path:

  • Resources created by operator version 8.2.6-21 or later: the operator owns the setting from creation. Configure it only through the custom resource. Changes made through the Redis Software REST API or admin console are reverted the next time the operator reconciles the resource. Removing the field from the spec reverts the setting to its Redis Software default (Auto for the cluster, false for a database); it does not keep the last value.
  • Resources that existed before you upgraded to operator version 8.2.6-21: values configured through the Redis Software REST API are preserved. The operator takes ownership only once you set the field in the custom resource. From then on, ownership is permanent. If you later remove the field, the setting reverts to the Redis Software default, not to the previously configured value.

Storage and performance

Optimize storage and performance for your Redis Enterprise cluster:

  • Redis Flex - Configure automatic data tiering between RAM and flash storage
  • Expand PVC - Expand persistent volume claims for additional storage

Monitoring and observability

Monitor cluster health and performance:

Call home client

The call home client sends health or error data from your deployment(s) back to Redis. You can disable it by adding the following to your REC specification:

spec:
  usageMeter:
    callHomeClient:
      disabled: true
Note:

The REST API approach used for Redis Software deployments will have no effect on Kubernetes deployments. You must use the REC specification method shown above.

Recovery and troubleshooting

Handle cluster recovery and troubleshooting scenarios:

RATE THIS PAGE
Back to top ↑