AI Center Is Inaccessible Due To Kerberos Login Failure

News Knowledge Base
, , ,
1

AI Center is inaccessible due to Kerberos login failure.

Issue Description: A hard reboot of the AI Center server and post that AI Center is inaccessible.

While browsing the URL it states : 'AI Center is not reachable' and gives 503 error

Troubleshooting Steps:

  1. Check the pod status and see if aicenter pods ( ai-appmanager, ai-helper etc) are in CrashLoopBackOff
  2. Check the logs and it will point to following error:

Exception during pool initialization.","traceId":"","spanId":"","parent":"","logging.googleapis.com/trace":"projects/dave-225414/traces/","logging.googleapis.com/spanId":"","account-id":"","user-id":"","project-id":"","organization-id":"","tenant-id":"","authorized-party":"","thread":"main","exception":"javax.security.auth.login.LoginException: Cannot get any of properties: [user, USER] from con properties not available to garner authentication information from the user\n\tat

  1. Check the status of apps aicenter and uipath in ArgoCD - they would be Degraded


Resolution:

  1. Check the cronjob 'kerberos-tgt-update' - kubectl get cronjobs -A
  2. Create a new job from the above cronjob to re run the kerberos update:
  • kubectl create job --from=cronjob/ kerberos-tgt-man-update
  1. Post that restart the deployment in UiPath namespace:
  • kubectl rollout restart deploy -n uipath
  1. Once that was done, see all pods come up as Running
  2. Re sync the app in ArgoCD and all should be in Healthy state
  3. Try accessing AI Center again and it should be accessible.