Since upgrading to Octopus 3.11.5 we are encountering issues with cancelling releases. After we are running deployments we sometimes encounter our deployments hanging requiring us to cancel, when cancelling the release the UI changes to show the deployment as cancelled but we see the following error in the logs…
“This task was canceled before it could start. The other task is still running.”
Rerunning the deployment shows the cancelled task as still running and causes the deployment to join the queue.
I’ve attached screenshots.
Thank you for getting in touch. This is part of an issue I have been working on and am keen to get more information. This message indicates that another task has an exclusive lock on the Tentacle.
When this next occurs, could you check the machine deployment information (
Environments -> Click the Machine -> Deployments tab) to see if anything else is running, or if the last deployment was cancelled. Also check the Task list (
Tasks in the menu) to see if anything else is running.
If you can get me a process dump of the affected
Tentacle.exe process, that would be great. I’ve created a secure upload folder for any files: https://file.ac/ut6yiT6h2XY/. Once I get that, I can check to see why the Tentacle is reporting that something else is running. My suspicion is that it is due to a previous task that was cancelled that left an orphaned hung process on the Tentacle.
I’m seeing the same issues. Where a team canceled a deployment it just hangs.
While I was looking into that, I noticed that a health check was still running. The health check started on April 3rd.
I canceled the health check, but it just hung that specific task.
I’m running version 3.11.9.
Would you be able to send me a process dump of the affected tentacle and the octopus server service? Also the task log of the health check and your server logs since the healthcheck started. You can upload them to https://file.ac/ut6yiT6h2XY/
Once you have captured that, restarting the octopus server service should complete the cancellation.
I’ve uploaded what ever I have.
No process dumps. No task log as I could no longer find the log from the hung health check. It was in a hung state for 30 days, before I found it.
I provided all the server and tentacle logs that I had.
Thank you for that. Please let us know if it happens again and are able to capture those things. We have changed the health check slightly in 3.12 to update the machine status during instead of at the end of the process, which may help.