immediately return from RMQ starting

This improves UX, with the slight risk
of having it fail immediately without
noticing right away.

CMK-21190

Change-Id: Iba891076fb7a2ad2493cbd9ef35d57889e731fd5

omd/packages/rabbitmq/skel/etc/init.d/rabbitmq

@@ -43,32 +43,15 @@ await_process_stop() {
     return 1
 }
 
-rabbitmq_accepting_connections() {
-    "${RABBITMQ_DIAGNOSTICS}" check_port_listener --node "${RABBITMQ_NODENAME}" "${RABBITMQ_PORT}"
-    # no idea why we can't use this anymore; it will fail for the management plugin
-    # "${RABBITMQ_DIAGNOSTICS}" check_port_connectivity --node "${RABBITMQ_NODENAME}"
-}
-
 await_process_start() {
-    # Only write output in the end, otherwise we clobber the logfile.
     max=$(("${1}" * 10))
     # If the PIDFILE gets filled, rabbitmq started
     for N in $(seq "${max}"); do
-        [ -s "${PIDFILE}" ] && rabbitmq_accepting_connections &>/dev/null && break
-        [ -e "${PIDFILE}" ] || {
-            date +"%F %T.%N RabbitMQ failed to start (pidfile vanished)"
-            return 1
-        }
+        [ $((N % 10)) -eq 0 ] && printf "."
+        [ -s "${PIDFILE}" ] && return 0
+        [ -e "${PIDFILE}" ] || return 1
         sleep 0.1
     done
-
-    # once again, mostly to get the output of the diagnostics call
-    [ -s "${PIDFILE}" ] && rabbitmq_accepting_connections && {
-        date +"%F %T.%N RabbitMQ considered ready"
-        return 0
-    }
-
-    date +"%F %T.%N RabbitMQ failed to start (giving up after timeout)"
     return 1
 }