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.

src/scontrol/scontrol.c

@@ -339,6 +339,7 @@ _ping_slurmctld(char *control_machine, char *backup_controller)
 {
 	static char *state[2] = { "UP", "DOWN" };
 	int primary = 1, secondary = 1;
+	int down_msg = 0;
 
 	if (slurm_ping(1) == SLURM_SUCCESS)
 		primary = 0;
@@ -347,17 +348,27 @@ _ping_slurmctld(char *control_machine, char *backup_controller)
 	fprintf(stdout, "Slurmctld(primary/backup) ");
 	if (control_machine || backup_controller) {
 		fprintf(stdout, "at ");
-		if (control_machine)
+		if (control_machine) {
 			fprintf(stdout, "%s/", control_machine);
-		else
+			if (primary)
+				down_msg = 1;
+		} else
 			fprintf(stdout, "(NULL)/");
-		if (backup_controller)
+		if (backup_controller) {
 			fprintf(stdout, "%s ", backup_controller);
-		else
+			if (secondary)
+				down_msg = 1;
+		} else
 			fprintf(stdout, "(NULL) ");
 	}
 	fprintf(stdout, "are %s/%s\n", 
 		state[primary], state[secondary]);
+
+	if (down_msg && ((getuid() == 0) || (geteuid() == 0))) {
+		fprintf(stdout, "*****************************************\n");
+		fprintf(stdout, "** RESTORE SLURMCTLD DAEMON TO SERVICE **\n");
+		fprintf(stdout, "*****************************************\n");
+	}
 }
 
 /*