Re: Random port usage in distributed CC

Cool. Have you looked at the both the logs on the remote agent and the 
CC master? (Also, you be sure you run the CC master with the -debug 
argument).

Usually if you register with the LUS, the remaining issues are things 
like name/ip resolution. You may have seen the "DNS Hack" notes in the 
config files and CC Distributed docs.

Dan


Date: Thu, 30 Jun 2011 17:58:45 +0000
From: B?ns?ghi, Csaba <[email protected]>
Subject: Re: [Cruisecontrol-user] Random port usage in distributed CC
To: "[email protected]"
	<[email protected]>
	
Content-Type: text/plain; charset="iso-8859-1"

I may or may not have found the solution to this. In 
conf/transient-reggie.config, I changed the parameter of 
TcpServerEndpoint.getInstance(0) to the number of an open port. And yes, 
I need to use unicast lookup, since my agent machine is halfway across 
the internet (on Amazon EC2).

Now the agent seems to register with the LUS fine (it is issued a 
serviceID and no errors/exceptions are shown), but for some reason 
neither CC nor the agent-util can find it (I tracked the error back to 
MulticastDiscovery.java, where in findAvailableBuildAgentService(), 
ServiceDiscoveryManager.lookup() returns with null for whatever reason). 
Indeed, after running a network monitoring tool on the master machine, 
it seems that no frames are being sent to the agent at all, even though 
there's plenty of traffic during the agent's registration with the LUS 
(all on the right ports).

So yeah, I'm not sure if changing TcpServerEndpoint.getInstance(0) is a 
solution. On one hand, all network traffic seems to occur on the proper 
ports, and the agent appears to register with the LUS. On the other 
hand, CC/agent-util is then unable to reach this agent, though this 
might be because of some other, unrelated problem.

Any advice would be greatly appreciated.
Csaba

------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security 
threats, fraudulent activity, and more. Splunk takes this data and makes 
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2