:sparkles: How to upgrade Strimzi Operator using the CLI
Developing
Photo by ThisIsEngineering from Pexels

:sparkles: How to upgrade Strimzi Operator using the CLI

2020, Sep 25    

Operator Lifecycle Manager (OLM) helps users install, update, and manage the lifecycle of all Operators and their associated services running across their clusters. It is part of the Operator Framework, an open source toolkit designed to manage Kubernetes native applications (Operators) in an effective, automated, and scalable way.

OLM is responsible for managing the custom resources definitions (CRDs) that are the basis for the OLM framework:

  • OperatorGroup (og): OLM used to group multiple namespaces and prepare for use by an operator.
  • Subscription (sub): Catalog used to keep CSVs up to date by tracking a channel in a package.
  • ClusterServiceVersion (csv):OLM application metadata: name, version, icon, required resources, installation, etc…
  • InstallPlan (ip): Catalog calculated list of resources to be created in order to automatically install/upgrade a CSV.

This article describes the upgrade process for the Strimzi Operator (or Red Hat AMQ Streams Operator) in an OpenShift Platform.

Note 1: Only cluster-admin users could manage the OLM objects.

Note 2: The article only covers the upgrade process of the Strimzi Operator, does not cover the strategies to upgrade the Apache Kafka cluster version.

Operator Subscription

To deploy the Strimzi Operators only to inspect a namespace (namespaced scope), we need to use an OperatorGroup. An OperatorGroup is an OLM resource that provides multitenant configuration to OLM-installed Operators.

This is a definition for manage operators in myproject namespace:

---
apiVersion: operators.coreos.com/v1
kind: OperatorGroup
metadata:
  name: myproject-og
  namespace: myproject
spec:
  targetNamespaces:
    - myproject

To install it:

❯ oc apply -f operatorgroup.yml

A Subscription represents an intention to install an Operator. It is the custom resource that relates an Operator to a CatalogSource. Subscriptions describe which channel of an Operator package to subscribe to, and whether to perform updates automatically or manually. If set to automatic, the Subscription ensures OLM manages and upgrades the Operator to ensure that the latest version is always running in the cluster.

This is a Subscription definition to install Strimzi Operator:

---
apiVersion: operators.coreos.com/v1alpha1
kind: Subscription
metadata:
  name: strimzi-kafka-operator
  namespace: myproject
spec:
  channel: stable
  installPlanApproval: Manual
  name: strimzi-kafka-operator
  source: community-operators
  sourceNamespace: openshift-marketplace
  startingCSV: strimzi-cluster-operator.v0.17.0

In the case of Red Hat AMQ Streams the definition is:

---
apiVersion: operators.coreos.com/v1alpha1
kind: Subscription
metadata:
  name: amq-streams
  namespace: myproject
spec:
  channel: stable
  installPlanApproval: Manual
  name: amq-streams
  source: redhat-operators
  sourceNamespace: openshift-marketplace
  startingCSV: amqstreams.v1.4.0

In both cases the following attributes are the important:

The channel, such as alpha, beta, or stable, helps determine which Operator stream should be installed from the CatalogSource.

  • channel: Identifies which Operator stream should be installed from the CatalogSourceThe channel. Values: alpha, beta, or stable.
  • installPlanApproval: Identifies the approval mode of the installPlan. If you defined as Automatic then the upgrade of the operator will be done as soon a new version is available.
  • startingCSV: Identifies the initial version to install of the operator.

:bulb: If you don’t know to values needed for these attributes, the following command will show you that information easily:

❯ oc get packagemanifest strimzi-kafka-operator \
	-o jsonpath='{range .status.channels[*]}{"Chanel: "}{.name}{" -- CSV: "}{.currentCSV}{" -- Version: "}{.currentCSVDesc.version}{" -- skipRange: "}{.currentCSVDesc.annotations.olm\.skipRange}{"\n"}'
Chanel: stable -- CSV: strimzi-cluster-operator.v0.19.0 -- Version: 0.19.0 -- skipRange: 
Chanel: strimzi-0.19.x -- CSV: strimzi-cluster-operator.v0.19.0 -- Version: 0.19.0 -- skipRange: 

For Red Hat AMQ Streams should be similar to:

❯ oc get packagemanifest amq-streams \
	-o jsonpath='{range .status.channels[*]}{"Chanel: "}{.name}{" -- CSV: "}{.currentCSV}{" -- Version: "}{.currentCSVDesc.version}{" -- skipRange: "}{.currentCSVDesc.annotations.olm\.skipRange}{"\n"}'
Chanel: amq-streams-1.5.x -- CSV: amqstreams.v1.5.3 -- Version: 1.5.3 -- skipRange: 
Chanel: amq-streams-1.x -- CSV: amqstreams.v1.5.3 -- Version: 1.5.3 -- skipRange: 
Chanel: stable -- CSV: amqstreams.v1.5.3 -- Version: 1.5.3 -- skipRange: 

:bulb: This article is focused to show how to update the Strimzi Operator manually, so this is the reason that we will use Manual installation plan approval mode. It could be a good practice to declare your subscriptions manually to control in which moment update your operators.

This command will install the subscription to deploy the operator:

❯ oc apply -f subscription.yml

After you create the subscription we could check it:

❯ oc get sub
NAME                     PACKAGE                  SOURCE                CHANNEL
strimzi-kafka-operator   strimzi-kafka-operator   community-operators   stable

The install plan is created, but it is not approved, so the operator is not installed jet:

❯ oc get ip
NAME            CSV                                APPROVAL   APPROVED
install-kt477   strimzi-cluster-operator.v0.17.0   Manual     false

This command will approve it to start the deployment of the operator:

❯ oc patch installplan install-kt477 --type merge --patch '{"spec":{"approved":true}}'

The install plan will be updated as:

❯ oc get ip
NAME            CSV                                APPROVAL   APPROVED
install-kt477   strimzi-cluster-operator.v0.17.0   Manual     true

After some minutes we could check the status of the CSV:

❯ oc get csv
NAME                               DISPLAY   VERSION   REPLACES   PHASE
strimzi-cluster-operator.v0.17.0   Strimzi   0.17.0               Succeeded

Now you are ready to deploy an Apache Kafka cluster using the Kafka CRD provided by this operator:

❯ oc apply -f kafka.yml

Upgrading the Operator

When a new version of your operator is released, OLM will check it and then a new install plan will be created for the new version.

If there is a new version available, the install plan should be similar to:

❯ oc get ip
NAME            CSV                                APPROVAL   APPROVED
install-2lxhr   strimzi-cluster-operator.v0.18.0   Manual     false
install-kt477   strimzi-cluster-operator.v0.17.0   Manual     true

Here you can see that there is a new strimzi-cluster-operator.v0.18.0 version available, so we could approved in the same way as the first installation:

❯ oc patch installplan install-2lxhr --type merge --patch '{"spec":{"approved":true}}'

When the new version is active the CSV will be updated and then you could check the upgrading history:

❯ oc get csv
NAME                               DISPLAY   VERSION   REPLACES                           PHASE
strimzi-cluster-operator.v0.17.0   Strimzi   0.17.0                                       Replacing
strimzi-cluster-operator.v0.18.0   Strimzi   0.18.0    strimzi-cluster-operator.v0.17.0   Installing

When the upgrade concluded the final status of the CSV will be:

❯ oc get csv
NAME                               DISPLAY   VERSION   REPLACES                           PHASE
strimzi-cluster-operator.v0.18.0   Strimzi   0.18.0    strimzi-cluster-operator.v0.17.0   Succeeded

The operator pod it will be recreated with the new version as you can see in the following watcher command:

❯ oc get pod -w
NAME                                                  READY   STATUS              RESTARTS   AGE
strimzi-cluster-operator-v0.17.0-5d9c7757c9-2blvf   1/1     Running             0          2m40s
strimzi-cluster-operator-v0.18.0-647d7bccc-bh66x    0/1     Pending             0          0s
strimzi-cluster-operator-v0.18.0-647d7bccc-bh66x    0/1     ContainerCreating   0          0s
strimzi-cluster-operator-v0.18.0-647d7bccc-bh66x    1/1     Running             0          2m36s
strimzi-cluster-operator-v0.17.0-5d9c7757c9-2blvf   1/1     Terminating         0          6m22s

This process will be repeated every time a new version of the operator is available. You only need to approve each version.

Versions Upgrading Workflow

Strimzi Operator is a very matured operator and includes a large list of versions. Each version includes a set of new features. If you start from an older version, OLM will prompt you to upgrade to the next version.

The upgrade pathway for Strimzi Operator versions starting from v0.17.0:

v0.17.0 -> v0.18.0 -> v0.19.0 -> vX.Y.Z

In the case of Red Hat AMQ Streams starting from v1.4.0 version is:

v1.4.0 -> v1.4.1 -> v1.5.0 -> v1.5.1 -> v1.5.2 -> v1.5.3 -> vX.Y.Z

NOTE: This list is not updated to the latest versions.

References