Did anyone reply to this?  I see a few uses of 'tc.node-name" in the code
base:

 

-          RequestCountTest.java and RogueClientTestApp.java look like
they are setting it

-          the spring webflow and rife demos.

-          Some of the error messages in StatisticsGathererSubSystem and
StatisticsGathererSubSystem seem to make reference to it

-          The session configurator use you describe

 

I have no idea if it is being used for the same purpose in all places. To
make matters worse, I think we have a property called "tc.nodeName" in the
maven plugin which is probably something completely different :-) 

 

Definitely feel like some things might need to be cleaned up here

 

 

 

  _____  

From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Gary Keim
Sent: Friday, April 18, 2008 2:47 PM
To: [email protected]
Subject: [tc-dev] tc.node-name

 

Does anyone understand why the JVM system property tc.node-name is being
used in the codebase?  I originally used it in the Session Configurator
startup scripts as a way to tag a JVM as taking part in a Configurator
session.  But somehow it got used as a way to change the JMX object name
for the L1Info bean, replacing the DSO Channel's remote address.

 

Without setting tc.node-name the L1Info bean's name contains:

 

Node=client.host/2361

 

But with it set to "Tomcat-9081" it becomes

 

Node=Tomcat-9081

 

This is problematic as the L1Info bean is tunneled into the server and we
need a consistent way to discover it.  Tc.node-name is screwing that up.

 

I can't find anywhere in the code where this property is actually used.

 

_______________________________________________
tc-dev mailing list
[email protected]
http://lists.terracotta.org/mailman/listinfo/tc-dev

Reply via email to