After a scaleout failure, automatic cleanup of virtual machines leads to a wait time.

Problem

When the VM_CLEANUP_AFTER_UPDATE_FAILURE value is set to true, and the scaleout operation fails, deprovisioning of virtual machines is done to ensure cleanup before the next scaleout update. During deprovisioning, if you query for the state of deployment, it might result in a failed task and the DEPLOYMENT_WITH _ISSUES state appears even as the update process continues. When this state appears, you must wait for sometime for deprovisioning to complete before trying another update process.

If you start another update process, the Cannot create update because update is still in progress. error appears.

Cause

You might be trying to schedule an update operation while the previous task has failed and the backend cleanup is still in progress.

Results

  • When you see the Cannot create update because update is still in progress error, retry the update operation with a delay of at least 120 seconds.

  • If you do not want to encounter this error and delay the update process, turn off the cleanup flag using CLI or REST API commands:

    update-global-prop --name VM_CLEAN_UP_AFTER_UPDATE_FAILURE --value false

    After you turn off the flag, you must delete the new virtual machines manually.

    For example, if a new virtual machine, appserver_3, was created in a scaleout update that failed, you must manually search for appserver_3 in the cloud provider virtual machines list and delete it.