Previously nodes in a MAINT reservation would NOT reboot on demand.

Conflicts:
	src/slurmctld/job_scheduler.c

job_step_cancel() appears to have been renames job_step_signal()
at some point in the past.

Signed-off-by: Danny Auble <da@schedmd.com>

to set it apart from the rest of the network_data's out there ;).

acct_gather_infiniband_g_node_init

all preemptable jobs on the midplane instead of just the ones it needed to.