Error Description

During the installation of the dbo chart, an internal error may occur due to a certificate conflict (with cert-manager).

Error: failed post-install: warning: Hook post-install druva-backup-operator/templates/cluster.yaml failed: 
Internal error occurred: failed calling webhook "mcluster.kb.io": Post 
"https://druva-webhook-service.druva-system.svc:443/mutate-backup-druva-com-v1alpha1-cluster?timeout=10s": 
no endpoints available for service "druva-webhook-service"

Cause

This issue arises since the Dbo chart does not include the webhook.securePort parameter. You will need to add this parameter for the dbo chart readme during cert-manager installation:

--set webhook.hostNetwork=true,webhook.securePort=10260

For more information on platform compatibility and troubleshooting Compatibility with Kubernetes Platform Providers

Workaround

For installations using Helm, you will need to set the webhook.securePort parameter when installing cert-manager. To modify the cert-manager install command, add this parameter:

 --set webhook.hostNetwork=true,webhook.securePort=10260

Error Description

During dbo chart installation, using the --dry-run parameter fails and no dbo pods are created.

Cause

The --dry-run parameter is used to simulate a chart installation. Using this option in Helm only simulates an installation, but in effect no pods are created.

Note: This is the expected outcome of --dry run, which is only a simulation. Use this parameter to verify the chart deployment output or to render the templates.

Workaround

To verify the components that the dbo chart installs, add two additional parameters with --dry-run:

--set disableChecks=true and --set druva-backup-config.disableChecks=true

helm install druva-backup-operator druva-backup-operator --namespace druva-system 
--create-namespace --debug --dry-run --set disableChecks=true --set druva-backup-config.disableChecks=true 
--set druva-backup-config.nameOverride=druva-backup-config --set s3-catalogue.nameOverride=druva-s3-catalogue 
--set global.image.registry=$IMAGE_REGISTRY --set global.image.tag=$IMAGE_TAG --set bootstrap.token=$TOKEN 
--set bootstrap.clusterURI=$CLUSTER_URI --set bootstrap.clusterURL=$CLUSTER_URL --set catalogue.url=$CATALOGUE_URL
--set prometheus.monitoring.enabled=false --set catalogue.planName=standard 
--set global.image.namespace=qa.druva.com/images --set druva-backup-config.catalogue.type=s3

helm install --debug --dry-run ./mychart

Error Description

When running the kubectl get cluster command, the ClusterID field returns blank, Available is False and Reason is blank. This indicates that the DBO installation has failed.

Cause

This is likely because of an incorrect bootstrap token being provided during DBO installation.
In this case, running kubectl get cluster returns the following output.

NAMESPACE      NAME            CLUSTERID                              CLUSTERNAME   CLOUD   REGION       CLUSTERTYPE   CLUSTERURI    CLUSTERVERSION   OPERATORVERSION   AVAILABLE   REASON
druva-system   druva-cluster      CLUSTER_URI   AWS     ap-south-1   kOps          CLUSTER_URI   v1.20.11         n/a               False

Resolution

1. Uninstall the Druva Backup Operator Helm chart.
2. Copy the bootstrap token from the CloudRanger Register Cluster page.

Note: it is recommended that you copy-paste the entire command as is. Alternatively, the Cloud Admin can provide the correct bootstrap token to authenticate clusters.

3. Reinstall the DBO Helm chart with the accurate bootstrap token.

Post successful installation, run the following command:

$ kubectl get cluster -A -o wide

The output will now return the appropriate ClusterID and Available as True.

NAMESPACE      NAME            CLUSTERID                              CLUSTERNAME   CLOUD   REGION       CLUSTERTYPE   CLUSTERURI    CLUSTERVERSION   OPERATORVERSION   AVAILABLE   REASON
druva-system   druva-cluster   9cd9c6c8-5521-4712-ad3c-faefe2165f1e   CLUSTER_URI   AWS     ap-south-1   kOps          CLUSTER_URI   v1.20.11         n/a              True        Registered