Upgrade a database cluster with zero downtime

You can upgrade a DSE or HCD cluster with zero downtime by updating the MissionControlCluster manifest. This topic demonstrates upgrading the DSE version from 6.9.12 to 6.9.13.

Prerequisites

A prepared environment on either bare-metal/VM or an existing Kubernetes cluster.

Upgrade process

The existing deployment consists of 1 datacenter (dc1) with 3 nodes, each node on a separate rack.

Workflow of user and operators

Submit a modified MissionControlCluster manifest to the control plane Kubernetes cluster.
Cluster-level operator detects the change and modifies the datacenter-level resource.
Datacenter-level operator detects the change and modifies the rack-level resource.
Datacenter-level operator terminates the instance in the rack.
Datacenter-level operator bootstraps a replacement node.
Datacenter-level operator repeats steps c-e for all racks in one datacenter.
Cluster-level operator repeats steps b-e for all datacenters in a cluster, one dc at a time, in the case of a multi-datacenter cluster.
User follows Upgrade SSTables steps, if required.

Upgrades of DSE versions follow the same process as configuration change deployments. See Deploy configuration changes to a DSE cluster.

Cluster upgrade with no downtime

Modify the existing MissionControlCluster YAML (demo.yaml) in the control plane Kubernetes cluster, overwriting the value 6.9.12 in the spec.k8ssandra.cassandra.serverVersion field with 6.9.13:
```
apiVersion: missioncontrol.datastax.com/v1beta2
kind: MissionControlCluster
metadata:
  name: demo
spec:
  k8ssandra:
    cassandra:
      serverVersion: 6.9.13
...
```
Submit the version change to the Kubernetes control plane cluster with the following command:
```
kubectl apply -f demo.yaml
```
The version upgrade is done as a rolling update.

Monitor the progress of the upgrade from the data plane east cluster, looking at the status of the CassandraDatacenter with this command:

kubectl get cassandradatacenter dc1 -o yaml

Result

status:
  cassandraOperatorProgress: Updating
  conditions:
  - lastTransitionTime: "2025-08-28T16:07:10Z"
    message: ""
    reason: ""
    status: "True"
    type: Healthy
  - lastTransitionTime: "2025-08-28T16:07:10Z"
    message: ""
    reason: ""
    status: "True"
    type: Stopped
...

If the Updating operation’s conditions:status is True, then the update is in progress. When the Updating operation is complete, the cass-operator changes the conditions.status to False.

Upgrade a database cluster with zero downtime

Prerequisites

Upgrade process

Workflow of user and operators

Cluster upgrade with no downtime

See also

Was this helpful?

Give Feedback