CSPL-4005: Remove CRDs from splunk/splunk-operator helm chart (#1579)

* remove crds from helm chart, needs documentation

* add documentation about adding CRDs before helm chart installation

* repackage helm charts

* cleanup helm test workflow

Author: rlieberman-splunk

.github/workflows/automated-release-workflow.yml

@@ -90,7 +90,7 @@ jobs:
     - name: Create Release
       uses: ncipollo/release-action@40bb172bd05f266cf9ba4ff965cb61e9ee5f6d01
       with:
-        artifacts: "release-${{ github.event.inputs.release_version }}/splunk-operator-cluster.yaml,release-${{ github.event.inputs.release_version }}/splunk-operator-namespace.yaml"
+        artifacts: "release-${{ github.event.inputs.release_version }}/splunk-operator-cluster.yaml,release-${{ github.event.inputs.release_version }}/splunk-operator-namespace.yaml,release-${{ github.event.inputs.release_version }}/splunk-operator-crds.yaml"
         bodyFile: "docs/ReleaseNotes.md"
         tag: "${{ github.event.inputs.release_version }}"
         draft: true

.github/workflows/helm-test-workflow.yml

@@ -169,6 +169,14 @@ jobs:
       - name: install k8s dashboard
         run: |
           kubectl apply -f https://raw.githubusercontent.com/kubernetes/dashboard/v2.0.5/aio/deploy/recommended.yaml
+      - name: Setup Kustomize
+        run: |
+          sudo snap install kustomize
+          mkdir -p ./bin
+          cp /snap/bin/kustomize ./bin/kustomize
+      - name: Install CRDs on cluster
+        run: |
+          make install
       - name: Add splunk helm repo for main branch
         if: github.ref == 'refs/heads/main'
         run: |

Makefile

@@ -372,7 +372,12 @@ generate-artifacts-cluster: manifests kustomize ## Deploy controller to the K8s
 	cd config/manager && $(KUSTOMIZE) edit set image controller=${IMG}
 	RELATED_IMAGE_SPLUNK_ENTERPRISE=${SPLUNK_ENTERPRISE_IMAGE} WATCH_NAMESPACE=${WATCH_NAMESPACE} SPLUNK_GENERAL_TERMS=${SPLUNK_GENERAL_TERMS} $(KUSTOMIZE) build config/default > release-${VERSION}/splunk-operator-cluster.yaml
 
-generate-artifacts: generate-artifacts-namespace generate-artifacts-cluster
+
+generate-crds: manifests kustomize ## Generate CRD artifacts
+	mkdir -p release-${VERSION}
+	$(KUSTOMIZE) build config/crd > release-${VERSION}/splunk-operator-crds.yaml
+
+generate-artifacts: generate-artifacts-namespace generate-artifacts-cluster generate-crds
 	echo "artifacts generation complete"
 
 #############################

docs/ChangeLog.md

@@ -12,6 +12,7 @@
 * CSPL-3867: SHC and CM Error Message Visibility
 * CSPL-3186: Upgrade Enterprise Security Version 8.0.2
 * 1559: Added SplunkGeneralTerms acceptance
+* CSPL-4005: Remove CRDs from splunk/splunk-operator helm chart
 
 ### Supported Splunk Version
 

docs/Helm.md

@@ -11,14 +11,20 @@ helm repo update
 
 The ```splunk``` chart repository contains the ```splunk/splunk-operator``` chart to deploy the Splunk Operator and the ```splunk/splunk-enterprise``` chart to deploy Splunk Enterprise custom resources.
 
-Upgrading to latest version of splunk operator using helm chart will not upgrade CRDs. User need to deploy the latest CRDs manually. this is [limitation](https://helm.sh/docs/chart_best_practices/custom_resource_definitions/) from helm
+Users need to deploy the latest CRDs manually. This is a [limitation](https://helm.sh/docs/chart_best_practices/custom_resource_definitions/) from helm. The ```splunk/splunk-operator``` chart no longer contains the CRDs to install on the first deployment due to the size of the CRDs. Helm has a chart size [limit of 1MB](https://helm.sh/docs/topics/advanced/#sql-storage-backend) due to internal limits in Kubernetes' underlying etcd key-value store, and the Splunk Operator for Kubernetes CRDs are too big to fit into the helm chart to deploy the operator. To install the CRDs for the first time, or to update the CRDs to the latest versions, follow one of the following steps.
 
 ```
 git clone https://github.com/splunk/splunk-operator.git .
-git checkout release/2.8.1
+git checkout release/3.0.0
 make install
 ```
 
+OR
+
+```
+kubectl apply -f https://github.com/splunk/splunk-operator/releases/download/3.0.0/splunk-operator-crds.yaml --server-side
+```
+
 Helm provides a long list of commands to manage your deployment, we'll be going over a few useful ones in the sections to come. You can learn more about supported commands [here](https://helm.sh/docs/helm/helm/).
 
 ## Splunk Operator deployments
@@ -35,7 +41,7 @@ There are a couple ways you can configure your operator deployment
 
 1. Using a ```new_values.yaml``` file to override default values (Recommended)
 ```
-helm install -f new_values.yaml --set installCRDs=true <RELEASE_NAME> splunk/splunk-operator -n <RELEASE_NAMESPACE>
+helm install -f new_values.yaml <RELEASE_NAME> splunk/splunk-operator -n <RELEASE_NAMESPACE>
 ```
 
 2. Using the Helm CLI directly to set new values
@@ -46,7 +52,7 @@ helm install --set <KEY>=<VALUE> <RELEASE_NAME> splunk/splunk-operator -n <RELEA
 If the release already exists, we can use ```helm upgrade``` to configure and upgrade the deployment using a file or the CLI directly as above.
 
 ```
-helm upgrade -f new_values.yaml --set installCRDs=true <RELEASE_NAME> splunk/splunk-operator -n <RELEASE_NAMESPACE>
+helm upgrade -f new_values.yaml <RELEASE_NAME> splunk/splunk-operator -n <RELEASE_NAMESPACE>
 ```
 
 Read more about configuring values [here](https://helm.sh/docs/intro/using_helm/).
@@ -74,7 +80,7 @@ The ```helm list``` command can be used to retrieve all deployed releases.
 
 By default, the Splunk Operator has cluster-wide access. Let's upgrade the ```splunk-operator-test``` release by revoking cluster-wide access:
 ```
-helm upgrade --set splunkOperator.clusterWideAccess=false --set installCRDs=true splunk-operator-test splunk/splunk-operator -n splunk-operator
+helm upgrade --set splunkOperator.clusterWideAccess=false splunk-operator-test splunk/splunk-operator -n splunk-operator
 ```
 ```
 NAME: splunk-operator-test
@@ -100,7 +106,7 @@ helm dependency build splunk/splunk-enterprise
 ```
 If the operator is already installed then you will need to disable the dependency:
 ```
-helm install --set splunk-operator.enabled=false --set installCRDs=true <RELEASE_NAME> splunk/splunk-enterprise -n <RELEASE_NAMESPACE>
+helm install --set splunk-operator.enabled=false <RELEASE_NAME> splunk/splunk-enterprise -n <RELEASE_NAMESPACE>
 ```
 Installing ```splunk/splunk-enterprise``` will deploy Splunk Enterprise custom resources according to your configuration, the following ```new_values.yaml``` file specifies override configurations to deploy a Cluster Manager, an Indexer Cluster and a Search Head Cluster.
 
@@ -124,7 +130,7 @@ helm show values splunk/splunk-enterprise
 
 To install a Splunk Enterprise deployment according to our configurations above:
 ```
-helm install --set installCRDs=true -f new_values.yaml splunk-enterprise-test splunk/splunk-enterprise -n splunk-operator
+helm install -f new_values.yaml splunk-enterprise-test splunk/splunk-enterprise -n splunk-operator
 ```
 ```
 NAME: splunk-enterprise-test
@@ -156,6 +162,17 @@ release "splunk-enterprise-test" uninstalled
 
 Note: Helm by default does not cleanup Custom Resource Definitions and Persistent Volume Claims. Splunk Admin needs to manually clean them.
 
+### Troubleshooting Splunk Enterprise Deployments
+
+#### CRDs are not installed
+If you attempt to install a Splunk Enterprise deployment, and there is an error that says:
+```
+Error: INSTALLATION FAILED: unable to build kubernetes objects from release manifest: resource mapping not found for name: "release-name" namespace: "release-namespace" from "": no matches for kind "Standalone" in version "enterprise.splunk.com/v4"
+ensure CRDs are installed first
+```
+
+Verify that the CRDs have been installed with the instructions at the [top of this documentation](#splunk-operator-helm-chart-repository).
+
 ## Splunk Validated Architecture deployments
 
 The Splunk Enterprise chart has support for three Splunk Validated Architectures:
@@ -166,6 +183,6 @@ The Splunk Enterprise chart has support for three Splunk Validated Architectures
 
 Install a Standalone deployment using the following command:
 ```
-helm install --set s1.enabled=true --set installCRDs=true <RELEASE_NAME> splunk/splunk-enterprise -n <RELEASE_NAMESPACE>
+helm install --set s1.enabled=true <RELEASE_NAME> splunk/splunk-enterprise -n <RELEASE_NAMESPACE>
 ```
 Visit the Splunk Operator github repository to learn more about the configurable values of [splunk/splunk-operator](https://github.com/splunk/splunk-operator/blob/develop/helm-chart/splunk-operator/values.yaml) and [splunk/splunk-enterprise](https://github.com/splunk/splunk-operator/blob/develop/helm-chart/splunk-enterprise/values.yaml).

docs/README.md

@@ -136,7 +136,9 @@ splunk-operator-75f5d4d85b-8pshn   1/1     Running   0          5s
 
 ### Installation using Helm charts
 
-Installing the Splunk Operator using Helm allows you to quickly deploy the operator and Splunk Enterprise in a Kubernetes cluster. The operator and custom resources are easily configurable allowing for advanced installations including support for Splunk Validated Architectures. Helm also provides a number of features to manage the operator and custom resource lifecycle. The [Installation using Helm](Helm.md) page will walk you through installing and configuring Splunk Enterprise deployments using Helm charts.
+Installing the Splunk Operator using Helm allows you to quickly deploy the operator and Splunk Enterprise in a Kubernetes cluster. The operator is easily configurable allowing for advanced installations including support for Splunk Validated Architectures. Helm also provides a number of features to manage the operator lifecycle. The [Installation using Helm](Helm.md) page will walk you through installing and configuring Splunk Enterprise deployments using Helm charts.
+
+Splunk Operator CRDs are not deployed as part of the helm installation. Users need to deploy the latest CRDs manually. See the [Installation using Helm](Helm.md) documentation on how to deploy the CRDs before installing the helm charts.
 
 ## Upgrading the Splunk Operator
 

docs/SplunkGeneralTermsMigration.md

@@ -43,7 +43,7 @@ make deploy IMG=docker.io/splunk/splunk-operator:<tag name> SPLUNK_GENERAL_TERMS
 ```
 3. Set the value in a `helm install` command
 ```
-helm install -f new_values.yaml --set installCRDs=true --set splunkOperator.splunkGeneralTerms="[required value]" <RELEASE_NAME> splunk/splunk-operator -n <RELEASE_NAMESPACE>
+helm install -f new_values.yaml --set splunkOperator.splunkGeneralTerms="[required value]" <RELEASE_NAME> splunk/splunk-operator -n <RELEASE_NAMESPACE>
 ```
 4. Edit the splunk-operator-controller-manager deployment after it is deployed
 ```