Helm Management Guide
When you install a Helm chart, Helm creates a release. You can manage the release by updating, upgrading, or uninstalling it.
When you install a Helm chart, Tiller (the Helm server) creates an instance of that chart in your Kubernetes cluster. This instance is called a release, and Tiller uses it to track all of the objects and resources that the chart creates. So when you need to manage the Operator or Couchbase Server using Helm, you do so by making changes to the release.
The rest of this page describes the best practices for managing Helm releases in your Kubernetes or OpenShift environment.
After installation, you may find yourself needing to make updates to the Operator or Couchbase configuration.
Similar to installing the chart, customizations are made in the form of overrides via the --values or --set options.
However, instead of
helm install, you’ll use
Here’s an example:
helm upgrade --values myvalues.yaml <release-name> couchbase/couchbase-operator
It’s important that you make your updates using the
helm upgrade command instead of using
kubectl or manually editing chart resources.
This is to ensure that all resources are updated appropriately.
Upgrading the Operator and admission controller to a newer version requires that you upgrade the Operator release to a newer version of its chart. This is to ensure that any other dependencies related to the Operator upgrade are also updated appropriately.
|For production deployments where you’ve manually installed the CRD, you’ll need to manually update the CRD before you upgrade the Operator via Helm.|
helm upgrade --version <version> <release-name> couchbase/couchbase-operator
<version> is the version of the chart that you want to upgrade to, and
<release-name> is the name of the release that is managing the instance of the Operator that you are trying to upgrade.
Refer to the Operator upgrade documentation for more information about the upgrade process.
|If you didn’t originally install the Operator using Helm, then you cannot upgrade the Operator using Helm. At this time, Operator installations that weren’t created with Helm cannot be ported over to using Helm.|
There are two ways to upgrade the Couchbase Server cluster with Helm. The first way is to upgrade the entire cluster chart to a version that uses the desired Couchbase Server version as a default. The second way is to update the existing chart to use a newer Couchbase Server image. The preferred way is to upgrade the entire cluster chart, since this will ensure that any parameters required to configure the cluster are defined.
|When upgrading the Couchbase Server cluster, you should first upgrade the Operator to the latest compatible version so as to ensure that the cluster can be properly managed.|
Check the default Couchbase Server version of a chart to make sure it uses the version that you want to upgrade your cluster to.
helm inspect --version <version> values couchbase/couchbase-cluster | grep "version:"
<version>is the version of the chart.
Upgrade the Couchbase Server cluster using the newer chart.
helm upgrade --version <version> <release-name> couchbase/couchbase-cluster
helm upgrade --set couchbaseCluster.version=<version> <release-name> couchbase/couchbase-cluster
<version> is the name of the container image.
For example, to upgrade a cluster to Couchbase Server 6.0.1:
helm upgrade --set couchbaseCluster.version=enterprise-6.0.1 <release-name> couchbase/couchbase-cluster