[Apologies for the long story!] Anyone tried this?
We have successfully created the RMI protocol for JMX in our Sandbox environment (CMC>Applications>Monitoring Application>tick box for "Enable RMI protocol for JMX" and enter unused port). Then restarted the monitoring service and a valid RMI JMX agent end point URL comes up. (service:jmx:rmi://[hostname]:6500/jndi/rmi://[hostname]:6500/[FQDN]:6400)
We then go to IBM Tivoli Monitoring Agent builder (used version 6.2.2. and version 6.3.0, makes no difference) and create a new project and create an initial JMX data source, specifiy the connection URL from the step above and a BI4 user account that is a member of the administrators group.
Then we have two scenarios:
If I have the SIA running under the default clustername (in other words: fully qualified hostname with port 6400), the endpoint URL has that FQDN clustername, and I can connect fine.
However, if I stop the SIA and change the clustername to, for example, BI4, and restart the SIA everything runs fine. Of course now the endpoint URL has changed to something like service:jmx:rmi://[hostname]:6500/jndi/rmi://[hostname]:6500/[clustername].
But back in Tivoli, I cannot connect anymore after I alter my connection to use the new endpoint URL.
The error message, KQZ0262E, only complains that it cannot find my host anymore and then it says that the hostname is the actual IP address of the workstation where I run the Agent Builder software.
So it seems to me that the Agent Builder software cannot determine the actual machine it needs to connect to, and (eventually?) tries to use it's localhost address. And of course that will not work.
So the workaround is to remove the clusternames on my environments to the default FQDN one, but that has too many side effects.
So anyone any thoughts how I can make Tivoli work with a changed BI clustername?