Hi,
we are currently seeing a very strange problem and have no clue on how to track it down: We are using the continous integration server hudson to continuously build and deploy our application to a test-server (running Glassfish V2UR2) whenever somebody checks in a change. Building and deploying is done using an ant-script that (after building) just calls the asadmin command line tool (specifically "asadmin deploy --interactive=false --upload=false ourApp.war") to deploy the application. Afterwards, the ant script goes on to run some unit-tests and some tests using HTTPUnit on the deployed application.
For the first time(s) after a server restart, everything is working fine, but after some runs of the ant script, the build starts to hang at the deploy command. There is no error reported in the server-logs, the command does not produce any error message, it just hangs forever, completely stopping the build.
Since we don't get any error message or log entry, we aren't really sure how to find out what is causing this. And that it's not really reproducible (it only happens after a varying amount of builds from within hudson) doesn't help either. Maybe somebody has seen something similar once? Or could in any other way help in shedding some more light on this, giving some clues on what we could try to track this down?
Some more information on our setup: Glassfish is running on a 64-bit linux system using Java 1.6.0_03, the only two webapps deployed are our app and hudson. We are currently using the built-in derby database ("JavaDB") of GF for our application (using hibernate as the JPA provider). Oh, and our app is using ICEfaces.
Any help on this would be very much appreciated!
Regards,
David
[Message sent by forum member 'dklotz' (dklotz)]
http://forums.java.net/jive/thread.jspa?messageID=306769