Skip to main content

migrate-aks-troubleshooting

AKS Migration Troubleshooting​

For a Failed migration status, a reason for the failure is written in the Ocean logs. Refer to the failure messages below and additional information to help you resolve the issue.

  • “The migration has failed due to connectivity issues. Make sure the Spotinst controller is installed and working properly.” - Refer to the troubleshooting topic for your version of the Ocean Controller.
  • “The migration has failed due to the inability to spin up enough new nodes. Make sure all required Virtual Node Groups are configured.” - Review all your workload's requirements and ensure that the required Virtual Node Groups exist to comply with those requirements.
  • “The migration has failed due to technical timeout.” Try running the migration with smaller node groups. Group the nodes according to labels or by similar configuration requirements.
  • “The migration has failed due to the failure to schedule pods for new nodes.” This occurs if unscheduled or unready pods are in the cluster after the draining time threshold has passed. Verify that no new workloads are manually created during the migration.
  • “The migration has failed due to failure to terminate old nodes.” - This failure occurs if Ocean cannot terminate some or all of the migration nodes. Ensure you have the correct permissions and role definitions for the Delete Nodes action in your Azure account.
  • “The migration has failed due to the failure to terminate old nodes during the migration Stop process.” - This failure occurs if Ocean cannot terminate some or all of the migration nodes. Ensure you have the correct permissions and role definitions for the Delete Nodes action in your Azure account.
  • “The migration has stopped due to failure to reschedule old nodes during the migration stop process.” - This failure occurs if Ocean cannot reschedule the old nodes for the migration after you have stopped the migration. Validate the connectivity and permissions for the Uncordon Node action in the Kubernetes cluster.