1. Take node offline and node goes to draining status because of running job.
2. Later on running job finished and core in use is 0. However node stay on draining status and cancel take node offline action doesn't work.
3. Reboot hpc agent service, node and headnode don't help.