[hfluz]

My host.xml:

<?xml version='1.0' encoding='UTF-8'?>
<host name="master" xmlns="urn:jboss:domain:1.2">
    <management>
        <security-realms>
            <security-realm name="ManagementRealm">
                <authentication>
                    <properties path="mgmt-users.properties" relative-to="jboss.domain.config.dir"/>
                </authentication>
            </security-realm>
            <security-realm name="ApplicationRealm">
                <authentication>
                     <properties path="application-users.properties" relative-to="jboss.domain.config.dir" />
                </authentication>
            </security-realm>
        </security-realms>
        <management-interfaces>
            <native-interface security-realm="ManagementRealm">
                <socket interface="management" port="${jboss.management.native.port:9999}"/>
            </native-interface>
            <http-interface security-realm="ManagementRealm">
                <socket interface="management" port="${jboss.management.http.port:9990}"/>
            </http-interface>
        </management-interfaces>
    </management>


    <domain-controller>
       <local/>
    </domain-controller>


    <interfaces>
        <interface name="management">
            <inet-address value="${jboss.bind.address.management:189.90.70.58}"/>
        </interface>
        <interface name="public">
           <inet-address value="${jboss.bind.address:189.90.70.58}"/>
        </interface>
        <interface name="unsecure">
            <!-- Used for IIOP sockets in the standard configuration.
                 To secure JacORB you need to setup SSL -->
            <inet-address value="${jboss.bind.address.unsecure:189.90.70.58}"/>
        </interface>
    </interfaces>

    <jvms>
                <jvm name="default">
            <heap size="64m" max-size="256m"/>
            <permgen size="256m" max-size="256m"/>
              <jvm-options>
                  <option value="-server"/>
              </jvm-options>
         </jvm>
             </jvms>


          <servers>
        <server name="server-one" group="main-server-group">
                  <socket-bindings socket-binding-group="ha-sockets" port-offset="150"/>
        </server>
        <server name="server-two" group="main-server-group" auto-start="true">
            <socket-bindings socket-binding-group="ha-sockets" port-offset="250"/>
                              <jvm name="default"/>
        </server>
        <server name="server-three" group="other-server-group" auto-start="true">
            <socket-bindings socket-binding-group="ha-sockets" port-offset="350"/>
                              <jvm name="default"/>
        </server>
    </servers>
</host>

[hfluz]

I uploaded all the files that might be relevant.

[hfluz]

I'd like to debug using the JBoss-cli, but I just can't use it.

In the JBoss AS 7 Administration book  the command :list-proxies is executed inside the modcluster subsystem, but when I try to execute it, cli simply can't find it.

[domain@189.90.70.58:9999 /] cd subsystem=modcluster  
[domain@189.90.70.58:9999 subsystem=modcluster] :list-proxies            
{
    "outcome" => "failed",
    "failure-description" => "JBAS010850: No handler for operation list-proxies at address [(\"subsystem\" => \"modcluster\")]",
    "rolled-back" => true
}

The book is based on JBoss AS 7.0, something changed in the 7.1 version?

[kazaag]

If I read well your configuration on server-three is using a ha profile, so it is the only one that will use modcluster.

For cli the example used the standalone configuration as you are using the domain one it will be different.

The path structure is different in domain mode (https://docs.jboss.org/author/display/AS71/Admin+Guide#AdminGuide-Manageddomain)

You can get to /profile=full-ha/subsystem but this will only list the configuration point of view.  To see the discovered list you need to go to /master/server=server-three/subsystem=modcluster

[hfluz]

Actually I realized that a few hours ago and changed to:

<server-groups>
        <server-group name="main-server-group" profile="full-ha">
            <jvm name="default">
                <heap size="64m" max-size="512m"/>
            </jvm>
            <socket-binding-group ref="full-ha-sockets"/>
        </server-group>
        <server-group name="other-server-group" profile="full-ha">
            <jvm name="default">
                <heap size="64m" max-size="512m"/>
            </jvm>
            <socket-binding-group ref="full-ha-sockets"/>
        </server-group>
</server-groups>

But that unhappily didn't solve my issue.

And previously I sent the wrong host.xml, the right one is attached to this post.

About CLI, how I go into a server-group? "cd server-group=main-server-group"?

[hfluz]

I managed to add the advertise multicast address to the proxy-list through jboss cli gui (a very helpful tool). Using the same tool I can view the mod_cluster configuration in JBoss:

But 189.90.70.58:8888/mod_cluster-manager still doesn't display any connection between httpd and jboss.

[hfluz]

And that's mod_cluster info in full-ha-socket.

[kazaag]

So advertising is working, JBoss nodes are aware of the proxy.

So it seems that there is an issue when the JBoss try to register them-self to the proxy.  Can you provide the JBoss log, there should be some error related to mod_cluster.

Have you any web application deployed?  If it isn't the casse, JBoss may wait before registering to the proxy as there is no real need to.

[hfluz]

Yes, I have an enterprise app deployed with JPA, Hibernate, EJB, JSF and Primefaces. I can access the app directly through a node (189.90.790.58:8230/appname-web). I guess I'll try with a simple web app.

About the jboss.log, I didn't find it inside the domain/log dir, I can find only host-controller.log and process-controller.log.

And thanks a lot for your help and patience.

[hfluz]

I tried to deploy the cluster-demo app (recommended here) without success.

I also copied the full log from terminal (it is attached to this post). This line caught my attention:

[Server:server-one] 15:54:18,006 INFO  [org.jboss.modcluster.advertise.impl.AdvertiseListenerImpl] (MSC service thread 1-6) Listening to proxy advertisements on 224.0.1.105:23364

[Server:server-one] 15:54:27,840 INFO  [org.jboss.modcluster.mcmp.impl.DefaultMCMPHandler] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) IO error sending command INFO to proxy ntds1r.mcast.net/224.0.1.105:23364: java.net.ConnectException: Network is unreachable

The same happens with server-two and server-three.

[kazaag]

The proxy ip:port is set to the multicast adress.

In the proxy list you should put the ip:port of the virtual host.

[hfluz]

Great! You're right, I changed the value to 189.90.70.58:8888 and that error is gone. Now I guess I'm really close to get everything working ok.

That's displayed in 189.90.70.58:8888/mod_cluster-manager

Do you have idea why my nodes are as undefined and removed? I still cannot access 189.90.70.58/cluster-demo.

Edit: Complementing with jboss log:

[Server:server-one] 17:27:37,976 ERROR [org.jboss.modcluster.mcmp.impl.DefaultMCMPHandler] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) Error [MEM: MEM: Old node still exist: {4}] sending command CONFIG to proxy u217064.ati.uel.br/189.90.70.58:8888, configuration will be reset
[Server:server-one] 17:27:47,982 ERROR [org.jboss.modcluster.mcmp.impl.DefaultMCMPHandler] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) Error [MEM: MEM: Old node still exist: {4}] sending command CONFIG to proxy u217064.ati.uel.br/189.90.70.58:8888, configuration will be reset
[Server:server-three] 17:27:48,165 ERROR [org.jboss.modcluster.mcmp.impl.DefaultMCMPHandler] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) Error [MEM: MEM: Can't read node: {4}] sending command STATUS to proxy u217064.ati.uel.br/189.90.70.58:8888, configuration will be reset
[Server:server-two] 17:27:57,958 ERROR [org.jboss.modcluster.mcmp.impl.DefaultMCMPHandler] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) Error [MEM: MEM: Can't read node: {4}] sending command STATUS to proxy u217064.ati.uel.br/189.90.70.58:8888, configuration will be reset

[hfluz]

I managed to solve the issue above with the command below:

/profile=full-ha/subsystem=web/:write-attribute(name=instance-id,value=${jboss.node.name})

I got the solution from https://community.jboss.org/thread/196776

Now the nodes are found correctly, but I still cannot access 189.90.70.58/cluster-demo

The info below is displayed in 189.90.70.58:8888/mod_cluster-manager:

Contexts:
/cluster-demo, Status: ENABLED Request: 0 Disable 
Aliases:
default-host localhost example.com 

[kazaag]

On mod_cluster point of view it is ok.

The log didn't show any deployement error.  You should be able to acccess /cluter-demo.

[hfluz]

Thanks François, you helped me a lot. I'll try to solve this last issue in the mod_cluster forum.