Add restart of xinetd in case of changing /etc/xinetd.d/mysql_clustercheck.

Service xinetd is not restarted after changing /etc/xinetd.d/mysql_clustercheck
This is cause false-positive checks of availability of galera cluster
and connections to DB are droped.

Closes-bug: PROD-36092
Change-Id: I74ccfb592f0208aeefc2604cf21b15c9b5786f7c
diff --git a/galera/clustercheck.sls b/galera/clustercheck.sls
index a311a84..036fb47 100644
--- a/galera/clustercheck.sls
+++ b/galera/clustercheck.sls
@@ -25,10 +25,6 @@
         per_source: UNLIMITED
     - require:
       - file: /usr/local/bin/mysql_clustercheck
-{%- if not grains.get('noservices', False) %}
-    - watch_in:
-      - galera_xinetd_service
-{%- endif %}
 
 galera_xinetd_package:
   pkg.installed:
@@ -40,6 +36,8 @@
   - name: xinetd
   - require:
     - pkg: xinetd
+  - watch:
+    - file: /etc/xinetd.d/mysql_clustercheck
 {%- endif %}
 {%- endif %}