Add special message for user root to restart slurmctld daemons as needed.

I keep finding the backup slurmctld down on production clusters and hope
this helps.