Task never leaves cancelling state and all other deployment tasks wait in a queue - even for deployments to other tentacles. It works again after restarting server service and rescheduling all pending tasks. There is no error logging when this happens. There is no reproducible scenario - this happens from time to time - has happend twice in the last two weeks. There was one task that was able to deploy while the other task was hanging - one that did not have a nuget step. Any help or directions would be most helpful.
Link to the issue on GitHub: https://github.com/OctopusDeploy/Issues/issues/1669
Hi Uldis,
Thanks for getting in touch! Thank you for reporting this issue. Unfortunately it is a known issue https://github.com/OctopusDeploy/Issues/issues/1431
It has been fixed for 3.0.
We have never been able to find a solution for the issue. One customer found that upgrading to the latest version did help. You do not mention what version you are running.
It appears to be a persistent error that only a very small number of customers found to happen frequently.
When a task is cancelled are you able to have the cancel button hit a second time immediately? Many of the customers who had the issue would have the task in a cancelled state for a long time (over an hr in most cases). But if you are able, hitting the cancel button a second time tells Octopus that you do not care if the task has running actors or not and closes it much faster than if it cannot gracefully cancel.
Sorry that it is not better news.
Vanessa