Troubleshooting user interface

Known issues in the Managed services user interface.

• When all Managed services pods are running, but you cannot still log in to the user interface, then it might be an issue with oauthcache.

• The plan functionality fails to work whenever you do the following steps:

  1. Return to the modify tab of a deployed instance wherein a plan is already run.
  2. Click Back or Edit parameters.

  As a workaround, do the following steps:

  1. Click select version step.
  2. Edit parameters page.
  3. Click Plan again to start the plan process

• The following error might occur when you start the console in spite of Managed services services running in IBM Cloud Pak for AIOps:

  CWOAU0061E: The OAuth service provider could not find the client because the client name is not valid.
  

  To resolve the problem, run the following commands to restart the cam-iaas pod and wait for the pod to start:

  kubectl get pods -n cp4aiops | grep iaas
  kubectl delete po/<pod_name> -n cp4aiops
  

  Wait for the pod to start. Run the following command to confirm whether it is in running state:

  kubectl get pods -n cp4aiops | grep iaas
  

• If you are redirected to the getting started page in Managed services user interface after the restart of IBM Cloud Pak for AIOps or you encounter other user interface related issues, then restart the user interface microservices. Run the following command to restart the user interface pods:

  kubectl delete -n cp4aiops -l 'name in (cam-portal-ui,cam-service-composer-ui,cam-ui-basic,cam-ui-connections,cam-ui-instances,cam-ui-templates)' pods
  

• As an LDAP user you might get redirected to the getting started page. To avoid such erroneous redirection, go to IBM Cloud Pak for AIOps user interface and ensure that the user is assigned to a team that has access to one or more namespaces.

• There might be performance issues in Managed services user interface whenever you use Mozilla Firefox browser. If this issue affects all websites wherein you use self-signed certificates, then the problem might be because of SSL certificates. If a service generates a new certificate for a particular website every time it is restarted, then Mozilla Firefox stores all these certificates in its local database and verifies whether the path is valid. Multiple self-signed certificates of the same name can slow down the turn around time of your browser. In addition, the TLS handshake operation ceases to respond. To identify the root cause of the problem and resolve it temporarily, open the profile folder of your Mozilla Firefox browser and locate the cert8.db file. Replace the file with a different name and revisit the same site that had problems. If the page displays properly, then the issue is because of local certificate database having too many self-signed certificates of the same name. Do the workaround often to avoid problem recurrence.

• Turn on the inspect for the browser and look for any errors returned on the API. For example, an API call might return a 401 (unauthorized), or might return a 400 (bad input) with an error message in the response, which is not displayed on the API. Page redirects, looping/refreshing pages might indicate a bad API call being returned to the user interface.

  To debug such errors, examine the following logs:

  • cam-ui-basic
  • cam-ui-connections
  • cam-ui-instances
  • cam-ui-templates
  • cam-portal-ui
  • cam-service-composer-ui

• The logout button might not be visible with some screen resolutions or if you zoom in the Managed services user interface. To solve the problem zoom out the user interface or access directly the logout page at https://<CAM_IP_address>:30000/logout.

• The Internet Explorer browser is not supported.

• Managed services UI shows one of the following errors:

  • Topology destroyed error while deployment
    • To resolve this issue, check the logs of cam-orchestration and if there are any errors, restart the pod.
  • Pool destroyed error on UI while accessing Managed services
    • To resolve this issue, check the logs of cam-ui-basic and if there are any errors, restart the pod.
  • Internal server error on UI while accessing Managed services
    • To resolve this issue, check the logs of cam-portal-ui and if there are any errors, restart the pod.

  Check whether all the pods are up and running for namespace cp4aiops.