Open
Description
Currently we run all jobs in step in parallel - we need to support running agents in sequential order. In the PR we consider cases where node can be configured to run
"ipmi reboot, apc1_off, apc2_off, apc1_on, apc2_on" in the power-management step.
if ipmi finished successfully we can stop the fence flow. if it failed, we need to make sure that all apc calls ran successfully.