I have been using SimpleDetectorClient, SimpleDetectorServer, and SimpleJNDIServer classes of the org.jboss.remoting.samples.detection.jndi package to help in discovering the problem you are seeing.
First off, looks like one issue is the way the AbstractDetector does its connection validation when it thinks a server has gone down. Currently it uses a client invoker to send a ping invocation. The problem is that there are no constraints on the timeout of the invocation, so could wait as long as 30 mintues (default) for socket timeout. Obviously this is too long to wait for connection validation in most cases. I have the changes locally on my machine and will commit to cvs (after open a jira issue to tie it to). I have fixed this by wrapping the ping invocation so will timeout external to the call itself. This has fixed the problem where detection of dead servers is delayed by the transport timeout. However, I have found another issue that still needs looking into. If using multiplex transport and a server is killed and then re-started, it looks like the detector sees the server restarting, then tries to ping it, but can not, so discards it as being dead, then repeats this process. This does not happen when using socket transport, so looks like there is some issue where multiplex client is having trouble making invocations on a server after it has gone down and then come back up. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3954192#3954192 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3954192 Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ JBoss-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jboss-user
