Hi,
I have noticed the EAP 6.x install with my enterprise application is throwing the IOException type a total 103 times when starting the server.
These exceptions do not cause the application to fail, they are merely ignored. The only reason I noticed was I had attached a debugger to detect exceptions being thrown.
You can see the exceptions thrown by turning on TRACE logging for this category here
<logger category="org.jboss.ws.common.deployment.ResourceResolverImpl">
<level name="TRACE"/>
</logger>
It seems the web services stack is throwing IOException as a technique to working it's way back up the call stack. The code attempts to load a static file from the VFS representing the web service deployment. When not present an IOException is thrown. Code like this
https://source.jboss.org/browse/~br=container-jboss71-jbossws-jboss713/JBossWS/container/jboss71/branches/jbossws-jboss713/server-integration/src/main/java/org/jboss/as/webservices/util/VirtualFileAdaptor.java?r=17012&r=17012#to58
I have two points to make:
Should code really be throwing exception like this to log a message and return up the call stack ?
Given the effort invested into reducing server boot time, throwing exceptions like these starts to undo that work. Large applications with many web services will see the same behaviour.
The server log is included showing the exception stack traces.
Regards,
Jeremy
Boot times on laptop, running Fedora16
EAP 6.0.1 Vanilla boot time: 3373ms
EAP 6.0.1 with ear application:10752ms
AS7.2 Brontes with ear application : 9063ms