We have been having quite a few issues automating our OnApp setup due to the fact that zombie tasks sometimes will not fail on their own and must be manually cancelled. If we suspend a customer and that customer has a zombie task running, the suspended VM stays online until the zombie task is cancelled. Even though the API responds in the affirmative and the VM shows as suspended, there is no way of knowing that this happened without checking the VM log and seeing that the StopVirtualServer command has not run. In other cases, clients will perform a task and then order additional services, then the task will become zombie tasks and we get no notification until the client complains that their additional services never arrived.
To combat this we have had to get our support team to manually verify every order, suspension, and termination. This may be fine for some, but it is nowhere near the level of automation that was promised when OnApp was pitched to our company.
Please implement another "miscellaneous" option similar to "zombie transaction time" which will specify how long to wait before automatically cancelling a zombie process. Adding a "set to 0 to never auto-cancel" option would be great for those who don't need this option.
Please sign in to leave a comment.