add possibility to run maintenance procedure on resources

cluster/init.sls

@@ -29,6 +29,7 @@ include:
 {% endif %}
 {% if host not in cluster.remove %}
   - .configure_resources
+  - .resources_maintenance
 {% endif %}
 {% if cluster.hacluster_password is defined %}
   - .hacluster_user

cluster/resources_maintenance.sls

@@ -0,0 +1,50 @@
+# This is e.g. needed to be run on the initial HANA cluster node to detect a running HANA correctly.
+# It follows the maintenance procedures from `man 7 SAPHanaSR_maintenance_examples`.
+# Without it, a "end action monitor_clone with rc=8 (0.155.0)" might happen.
+# This is somewhat a new behaviour in SAPHanaSR-0.155.0-4.17.1.
+#
+# It might also be used by any other cluster resource, this is why it is generally available.
+#
+{%- from "cluster/map.jinja" import cluster with context -%}
+{% set host = grains['host'] %}
+
+{% if cluster.init == host %}
+{% if cluster.configure is defined and cluster.configure.template is defined and cluster.configure.template.parameters is defined %}
+{% set resources = cluster.configure.template.parameters.resources_maintenance|default([]) %}
+
+{% for resource in resources %}
+
+resource_maintenance:
+  cmd.run:
+    - name: crm resource maintenance {{ resource }} on
+    # start after resources are defined
+    - require:
+      - configure-the-cluster
+
+wait_after_resource_maintenance:
+  cmd.run:
+    - name: cs_wait_for_idle --sleep 5
+    - require:
+      - resource_maintenance
+
+resource_refresh:
+  cmd.run:
+    - name: crm resource refresh {{ resource }}
+    - require:
+      - wait_after_resource_maintenance
+
+wait_after_resource_refresh:
+  cmd.run:
+    - name: cs_wait_for_idle --sleep 5
+    - require:
+      - resource_refresh
+
+resource_maintenance_off:
+  cmd.run:
+    - name: crm resource maintenance {{ resource }} off
+    - require:
+      - wait_after_resource_refresh
+
+{% endfor %}
+{% endif %}
+{% endif %}