[jboss-as7-dev] purpose of jboss.node.name and how to set.

[Jason T. Greene]

Can someone look into this? I'd like to release 7.1.1 today, but this 
bug looks both severe and annoying.

On 3/8/12 9:17 AM, David M. Lloyd wrote:
> On 03/08/2012 06:21 AM, Wolf-Dieter Fink wrote:
>> I'm a bit confused about the purpose of 'node-name' and how to set it.
>
> The purpose is to have a unique identifier for a server.  There are
> several libraries and frameworks which expect this.
>
>> I found out that the nodes name is required for UserTransaction and also
>> a notice:
>>
>> https://docs.jboss.org/author/display/AS71/Clustered+EJBs
>> If it's on the same machine, the two things you have to make sure is
>> that you pass the port offset for the second instance and also make sure
>> that each of the server instances have a unique jboss.node.name system
>> property
>
> Right.
>
>> ==== standalone ====
>> In this mode the only way to set is to use '-Djboss.host.name='.
>> I try to use
>> <server name="MyNode" xmlns="urn:jboss:domain:1.1">
>> OR
>> <system-properties>
>> <property name="jboss.node.name" value="MyNode"/>
>> </system-properties>
>>
>> but both won't work (I check with EJBClient.getUserTransaction(
>> node-name ) ) but get the error "No EJBReceiver available for node name
>> MyNode".
>
> Then this is a bug.  You should be able to specify the node name in the
> name attribute of<server>  and that's it.  If that isn't working then
> there's a bug.
>
>> ==== domain ====
>> I try to check whether the<server name="">  work and use also the
>> system-properties element inside each server.
>> But both attempts won't work (with EJBClient.getUtx).
>> The server name from the configuration is shown always by mgnmt console
>> and not reflect the system-properties setting, but the property inside
>> the application is set by the system-properties element.
>
> Same bug, probably.
>
>> ==================
>>
>> I'm confused because I expect that I can see the node name (as server
>> name) in the console and use the<server name=>  also for
>> getUserTransaction.
>>
>>
>>
>> My Questions about:
>> - what is the purpose of node-name (except user identifier for console),
>> only a identifier for clustering?
>
> As above.  It's an identifier for Remoting, clustering, and other
> subsystems.
>
>> - Should the node-name in mngmt console and the property jboss.node.name
>> get by an application getProperty the same?
>
> Yes.
>
>> - am I right that it a bug that the node-name in standalone.xml or
>> host.xml won't work for getUserTransaction and I should raise a JIRA?
>
> Yes.
>
>> - is it possible that other functions are affected by this?
>
> Yes.
>


-- 
Jason T. Greene
JBoss AS Lead / EAP Platform Architect
JBoss, a division of Red Hat