Integration of IceBox with IceGrid & Glacier2 & IcePatch2 with Java Services

Hello!

I ended up with a more complex setup, than I set out to implement. I'd like to verify the concept before starting testing and everything falls apart.
The Glacier2, IcePatch2 is working and tested, therefore I omitted the configuration of those.
My question is more about integration of IceBox into a routed IceGrid environment.

I'm sorry for the many question, before testing any actual code, but I believe it just doesn't make sense to implement it, if it's conceptually wrong.

1. Is this configuration correct and the system can understand it?:

<icegrid>
   
    <application name=“IceGrid-Dev">
    
        <replica-group id="Backend" filter="filterByRegion">
     
            ...
        </replica-group>

    
        <!-- ServiceProvider Template -->
 
        <server-template id="Provider">
   
            ...
        </server-template>

     
        <server-template id="IcePatch2">
  
            ...
        </server-template>

   
        <server-template id="Glacier2">
  
            ... 
        </server-template>

     
        <service-template id="ServiceTemplate">
    
            <parameter name="name"/>
      
            <parameter name="entryPoint"/>
     
            <parameter name="objIdentity"/>
            <parameter name="objType"/>
            <service name="${name}" entry="${entryPoint}">
   
                <adapter name="${service}" endpoints="default">

                    <object identity="${objIdentity}" type="${objType}"/>
                </adapter>
                <property name="${service}.Identity" value="${server}-${service}"/>
   
                <!-- IceGrid connection -->
  
                <property name="Ice.Default.Locator" value="${application}/Locator:tcp -h localhost -p 5064"/>
    
            </service>

        </service-template>

   
        <node name="node1”>
            <server-instance template="Glacier2"
 client-endpoints="ws -p 8111 -h localhost -r /bws:tcp -p 8110 -h localhost”
 session-timeout="300"
 server-endpoints="tcp -h localhost" />

  
            <server-instance template="IcePatch2" directory=“/path/to/serverdata"/>
            
<icebox id="IceBoxServer" endpoints="default" exe="java" activation="always">
      
                <option>IceBox.Server</option>
   
                <service-instance template="ServiceTemplate" name=“MyService” entryPoint="::com::full::path::to::MyService” objIdentity="MyServiceObject" objType="::com::full::path::to::MyServiceObject"/>
                <!-- server distribution contains environment -->
                <distrib icepatch="${application}.IcePatch2/server">
  
                    <directory>services/dev</directory>
     
                </distrib>
            </icebox>
      
        </node>
   
    </application>

</icegrid>


2. Can a well-known object be registered from the service-template, the way shown in the above configuration?

3. How would one get a proxy in the client to the MyServiceObject?
   My assumption would be (code is in javascript):

function getMyServiceFooResponse(param1, param2) {
    const setup = new Ice.Promise();
    Ice.Promise.try(
        function () {
            return MyServiceObjectPrx.checkedCast(_communicator.stringToProxy('MyServiceObject@IceGrid-Dev-MyService"));
        }
    )
    .then(
        function (myServiceObjectProxy) {
            return myServiceObjPrx.foo(param1, param2);
        }
    )
    .then(
        function (fooResponse) {
            setup.succeed(fooResponse);
        }
    )
    .exception(
        function (ex) {
            setup.fail(ex);
        }
    );
    return setup;
}

4. Is communicator.stringToProxy(s) looking for the proxy through the Glacier2's Locator (is it also a routed request)?

5. Does the IcePatch2 distrib configuration within the <icebox> applied above ensure that the <server-instance> finds the deployed .jar file for the service? Where should the .jar files be deployed for icebox? Or do I need to change the entry to represent the deployment path as well, something along the way of: ./services/dev/::com::full::path::to::MyService?

6. In general, does the <adapter> and <service-template> support replica-group, the same way as the <adapter> within a <server-template> tag does? So having a config with:

<icegrid>
<application>
<replica-group id="Services" filter="filterByRegion">
    <description>Development Environment access to Backend Services</description>
    <load-balancing type="adaptive" load-sample="1" n-replicas="0"/>
    <object identity="MyServiceObject" type="::MyService::MyServiceObject"/>
</replica-group>
...
<service-template id="ReplicatedServiceTemplate">
    
    <parameter name="name"/>
      
    <parameter name="entryPoint"/>
     
    <parameter name="regions" default="EUW,EUC"/>
    <service name="${name}" entry="${entryPoint}">
   
    <adapter name="${service}" endpoints="default"/>

    <property name="${service}.Identity" value="${server}-${service}"/>
   
    <!-- IceGrid connection -->
  
    <property name="Ice.Default.Locator" value="${application}/Locator:tcp -h localhost -p 5064"/>
    
    </service>

</service-template>

   
...
<node>
<icebox>
    <service-instance template="ReplicatedServiceTemplate" name=“MyService” entryPoint="::com::full::path::to::MyService”/>
</icebox>
<node>
</application>
</icegrid>

would mean, that the proxy can be retrieved with _communicator.stringToProxy("MyServiceObject@Services");. Is this assumption correct?

Thanks a lot, I appreciate the time you took to read through!