users@glassfish.java.net

Re: NPE caused by broker cluster manager not initialising

From: Amy Kang <Amy.Kang_at_Oracle.com>
Date: Fri, 02 Sep 2011 13:02:23 -0700

The NPE is caused by the following bug - embedded broker startup errors
do not throw an exception
http://java.net/jira/browse/MQ-106

For the "Loopback IP address" exception, to clarify what Nigel
mentioned below for the likely cause of it:
Some Linux installers automatically set the mahine hostname to loopback
network address, e.g. in /etc/hosts, something like
127.0.1.1 myhostname

Please see GlassFish MQ Administration Guide for what to do to correct
the system configuration
http://download.oracle.com/docs/cd/E18930_01/html/821-2438/aeohz.html

Amy

On 11-09-02 09:36 AM, forums_at_java.net wrote:
> The lock file problem is probably caused by a previous failure.
>
> I suspect the cause of your NPE is this error:
>
> [30/Aug/2011:16:31:05 BST] ERROR [B3168]: Invalid broker address for this
> broker to run in cluster: Loopback IP address is not allowed in broker
> address
> mq://127.0.1.1:27676/?instName=clusteri1&brokerSessionUID=1190975719361934592
>
> [1] for cluster<br />
> At a guess, this might be because you're using "localhost", and
> "localhost"
> resolves to the loopback address. I think that's common on Linix systems.
>
> Nigel
>
>
>
>
> [1]
> http://www.java.net//127.0.1.1:27676/?instName=clusteri1&amp;brokerSessionUID=1190975719361934592
>
> --
>
> [Message sent by forum member 'nigeldeakin']
>
> View Post: http://forums.java.net/node/839351
>
>