Upgrade of Scylla Operator

This page describes Scylla Operator upgrade procedures.There are two generic update procedures - via Helm and via kubectl. Before upgrading, please check this page to find out if your target version requires additional upgrade steps.

Upgrade via Helm

Helm doesn’t support managing CustomResourceDefinition resources (#5871, #7735)These are only created on first install and never updated. In order to update them, users have to do it manually.

Replace <release_name> with the name of your Helm release for Scylla Operator and replace <version> with the version number you want to install:

  1. Make sure Helm chart repository is up-to-date:

    helm repo add scylla-operator https://storage.googleapis.com/scylla-operator-charts/stable
    helm repo update
    
  2. Update CRD resources. We recommend using --server-side flag for kubectl apply, if your version supports it.

    tmpdir=$( mktemp -d ) \
      && helm pull scylla-operator/scylla-operator --version <version> --untar --untardir "${tmpdir}" \
      && find "${tmpdir}"/scylla-operator/crds/ -name '*.yaml' -printf '-f=%p ' \
      | xargs kubectl apply
    
  3. Update Scylla Operator

    helm upgrade --version <version> <release_name> scylla-operator/scylla-operator
    

Upgrade via kubectl

Replace <version> with the version number you want to install:

  1. Checkout source code of version you want to use:

    git checkout <version>
    
  2. Manifests use rolling minor version tag, you may want to pin it to specific version:

    find deploy/operator -name "*.yaml" | xargs sed --follow-symlinks -i -E "s^docker.io/scylladb/scylla-operator:[0-9]+\.[0-9]+^docker.io/scylladb/scylla-operator:<version>^g"
    
  3. Update Scylla Operator. We recommend using --server-side flag for kubectl apply, if your version supports it.

    kubectl apply -f deploy/operator
    

v1.2.0 -> v1.3.0

Sidecar image is going to be upgraded automatically, so a rolling restart of your Scylla clusters is expected during the upgrade procedure.

  1. Checkout source code of v1.3.0:

    git checkout v1.3.0
    
  2. Update Scylla Operator from deploy directory:

    kubectl -n scylla-operator apply -f deploy/operator
    
  3. Wait until Scylla Operator is up and running:

    kubectl wait --for condition=established crd/scyllaclusters.scylla.scylladb.com
    kubectl -n scylla-operator rollout status deployment.apps/scylla-operator
    

v1.1.0 -> v1.2.0

1.2.0 release brought a lot of changes to the Scylla Operator deployment process. To properly update Scylla Operator one must delete old objects and install updated ones.

Sidecar image is going to be upgraded automatically, so a rolling restart of your Scylla clusters is expected during the upgrade procedure.

  1. Checkout source code of v1.2.0:

    git checkout v1.2.0
    
  2. Remove old scylla operator namespace - in our case it’s called scylla-operator-system:

    kubectl delete namespace scylla-operator-system --wait=true
    
  3. Remove old webhooks:

    kubectl delete MutatingWebhookConfiguration scylla-operator-mutating-webhook-configuration
    kubectl delete ValidatingWebhookConfiguration scylla-operator-validating-webhook-configuration
    
  4. Install Scylla Operator from deploy directory:

    kubectl -n scylla-operator apply -f deploy/operator
    
  5. Wait until Scylla Operator is up and running:

    kubectl wait --for condition=established crd/scyllaclusters.scylla.scylladb.com
    kubectl -n scylla-operator rollout status deployment.apps/scylla-operator
    

v1.0.0 -> v1.1.0

During this update we will change probes and image for Scylla Operator. A new version brings an automation for upgrade of sidecar image, so a rolling restart of managed Scylla clusters is expected.

  1. Get name of StatefulSet managing Scylla Operator

    kubectl --namespace scylla-operator-system get sts --selector="control-plane=controller-manager"
    
    NAME                                 READY   AGE
    scylla-operator-controller-manager   1/1     95m
    
  2. Change probes and used container image by applying following patch:

    spec:
      template:
        spec:
          containers:
          - name: manager
            image: docker.io/scylladb/scylla-operator:1.1.0
            livenessProbe:
              httpGet:
                path: /healthz
                port: 8080
                scheme: HTTP
            readinessProbe:
              $retainKeys:
              - httpGet
              httpGet:
                path: /readyz
                port: 8080
                scheme: HTTP
    

    To apply above patch save it to file (operator-patch.yaml for example) and apply to Operator StatefulSet:

    kubectl -n scylla-operator-system patch sts scylla-operator-controller-manager --patch "$(cat operator-patch.yaml)"
    

v0.3.0 -> v1.0.0

Note: There’s an experimental migration procedure available here.

v0.3.0 used a very common name as a CRD kind (Cluster). In v1.0.0 this issue was solved by using less common kind which is easier to disambiguate. (ScyllaCluster). This change is backward incompatible, so Scylla cluster must be turned off and recreated from scratch. In case you need to preserve your data, refer to backup and restore guide.

  1. Get list of existing Scylla clusters

    kubectl -n scylla get cluster.scylla.scylladb.com
    
    NAME             AGE
    simple-cluster   30m
    
  2. Delete each one of them

    kubectl -n scylla delete cluster.scylla.scylladb.com simple-cluster
    
  3. Make sure you’re on v0.3.0 branch

    git checkout v0.3.0
    
  4. Delete existing CRD and Operator

    kubectl delete -f examples/generic/operator.yaml
    
  5. Checkout v1.0.0 version

    git checkout v1.0.0
    
  6. Install new CRD and Scylla Operator

    kubectl apply -f examples/common/operator.yaml
    
  7. Migrate your existing Scylla Cluster definition. Change apiVersion and kind from:

    apiVersion: scylla.scylladb.com/v1alpha1
    kind: Cluster
    

    to:

    apiVersion: scylla.scylladb.com/v1
    kind: ScyllaCluster
    
  8. Once your cluster definition is ready, use kubectl apply to install fresh Scylla cluster.