i have a playbook job running over 67 hours and im unable to cancel it, so everything is blocked in awx. i already pushed the cancel button in the ui multiple times but no go. so how to i cancel or remove this job?
Christian,
Can you share some details about your AWX install? https://github.com/ansible/awx/issues/new?template=bug_report.md
oh, of course.
- AWX version:1.0.6.43
- AWX install method: local rpm install (https://github.com/MrMEEE/awx-build)
- Ansible version: 2.6.2
- Operating System: RedHat EL 7.4
- Web Browser: Firefox
Christian,
The local RPM install method you referenced isn’t really a supported install method for awx; we’ve seen odd issues like this in the past. If you install the latest awx using one of the methods described at https://github.com/ansible/awx/blob/devel/INSTALL.md, do you still encounter the problem? I also know that from 1.0.6.43 → 1.0.7 we resolved several issues that could cause unreliable cancellation like you’ve described - it’s possible that updating might fix things for you:
https://github.com/ansible/awx/commit/ffdd7f162bca9369b6e0c779ee39f0ab811893f3
https://github.com/ansible/awx/commit/41fe9e1caf0d171f2d2da9c5be84332654a2c2e1
hmm…ok, will try out the new version.
Depending what your job is hanging on you may be able to kill the process and have it move on (i.e. kill the ssh session, which will fail the job but get around the “uncancleable” nature), you may want to go back and set the default timeout in awx to something more reasonable than the eternity it is set to by default to avoid the problem in the future.