cluster-mgr needs a way to interrupt an ongoing ansible provisioning

[mapuri]

this will be useful to recover out of scnearios where ansible run get's occasionally stuck.

Note that a timeout will not help here as ansible running time may vary based on amount of configuration and a timeout based logic will result in false positives.

[mapuri]

A workaround for short term is to kill the ansible related PID that shows up in systemctl status clusterm command, which will error out the operation and let clusterm handle it gracefully.