Troubleshooting

Global pull secret not updated for Red Hat OpenShift on IBM Cloud

Updating the global pull secret requires recycling all the nodes in your cluster to be effective. Pod logs will show ImagePullErr and the details of the error will contain unauthorized: authentication required. To pick up the global configuration changes for Red Hat OpenShift on IBM Cloud, reload all of the worker nodes in your cluster. Learn more here.

Install IBM Metrics Operator script was successful but cluster status is not Registered

If the install IBM Metrics Operator script reported success, but your cluster status has not changed to Registered after 1 hour, try these steps:

oc -n ibm-software-central delete pod -l redhat.marketplace.com/name=ibm-metrics-operator

The IBM Metrics Operator cluster registration reconciler runs on a 1 hour period. Restarting the pod forces the reconciler to re-run.

Timeout error during IBM Metrics Operator installation

Installation of the IBM Metrics Operator may result in a timeout error such as Timeout: request did not complete within allowed duration This could be caused by not having an OpenShift cluster at the minimum supported version. Please confirm your cluster is at major version 4 with any available supported minor version or newer.

Unregistered cluster shows registered status

You ran the unregister cluster command, the cluster status showed To be unregistered, and now the status

shows Registered. To resolve, ensure that the Red Hat Marketplace agent is

completely removed by running oc get pods -n ibm-software-central. You should not see any resources returned. If you do, try running the unregister cluster command again to completely clean up. If the issue is still not resolved, create a support case.
You clicked Confirm, but didn’t run the command. The cluster status showed To be unregistered, and now the status shows Registered. To resolve, run the unregister cluster command. Then unregister the cluster again and click Confirm.

OpenShift Container Platform web console error

You clicked the Cluster Console button, but your OpenShift Container Platform web console doesn’t show. There’s a broken link between your Red Hat Marketplace registered cluster and your OpenShift web console. The link may not be active for several minutes after successful cluster registration.

Register Red Hat CodeReady Containers, 1.9, warning message

When you register a cluster that runs on OpenShift Code Ready Container, version 1.9, the following warning shows: --dry-run=true. To be compatible with earlier versions, the parameter true shows instead of client.

Install Operator errors

Error messages show more than what is shown in this topic; however, the key phrases given explain how to resolve the installation error.

Failed to pull image: unauthorized: authentication required

Verify the pull secret is correct. Run:
```
docker login -u cp -p <pull_secret> registry.marketplace.redhat.com
```
Look for output Login succeeded. If an error occurs instead, the pull secret is invalid. Try creating a new secret.
Verify product subscription is active. In Red Hat Marketplace, click Workspace, and then click My software. Verify that the product being installed is listed and not pending, expired, or suspended.
When the pull secret is correct and product subscription is active, verify that the global pull secret has been configured correctly on the cluster. Login to the cluster as a cluster admin, and run:
```
oc get secret pull-secret -n openshift-config --output="jsonpath={.data.\.dockerconfigjson}" | base64 --decode
```
Look for an entry for “registry.marketplace.redhat.com”. Verify that the “password” value is your pull secret value. If it is missing or incorrect, run the following command to update the global pull secret on your cluster:
```
curl -sL https://marketplace.redhat.com/provisioning/v1/scripts/update-global-pull-secret | bash -s <pull_secret>
```
If the global pull secret is set up correctly, there are additional steps required to apply the global pull secret, if an IBM Cloud or CodeReady Containers cluster is being used.
- IBM Cloud: All worker nodes must be reloaded. Learn more here.
- CodeReady Containers: You must update the kubelet config and restart the CodeReady Containers cluster. Follow the Procedure here.
If nodes have been reloaded, or if you do not have an IBM Cloud or CodeReady Containers cluster, open a support case.

Failed to pull image: manifest unknown: manifest unknown

When you see this error, open a support case.

Manual installation of Operator into single namespace fails

Manual installation of an operator into a single namespace could fail if there is an existing operator also installed in that namespace from outside Red Hat Marketplace. Users may see an error such as Failed: csv created in namespace with multiple operatorgroups, can't pick one automatically.

Error shows when you install an operator to a single namespace that already has an operator installed from OpenShift Container Platform.

To resolve:

remove the OperatorGroup from the cluster, or
install the operator to a different namespace.

Software Operator doesn’t show on OpenShift Container Platform web console

To resolve, contact your OpenShift cluster administrator. When an installed operator doesn’t show on OpenShift, it could be an access issue that your cluster administrator can resolve.