Forums

JDK1.6.0_07 and _03: Non-blocking socket "oddity"
I have the following situation (on both Linux and MacOS) -- A non-blocking connection established to a server for a long-term persistent message oriented exchange (request-response). I'm doing negative testing and see something very odd (at least to my somewhat experienced "C" socket programming eyes). When I pull the cable, the select does not wake up (this is good, and what what I expect). When I write to the socket while the cable is pulled, I expect to see an IOException, but I don't. When I plug the cable back in at the server, the select in the java application "breaks", and the recovery processing is initiated. So, is this normal expected behavior in Java? Shouldn't the SocketChannel.write( ByteBuffer src) call fail?  — 

Re: Disabling ?WSDL
If request path needs changing you have no alternative beyond trying to avoid metro wsdl (maybe filter, trying to get load balancer to redirect ?wsdl, or try and rewrite the WSDL coming though (though that would be hard)). Otherwise you can try and get metro to have the correct URL: The address in the wsdl is derived from the servlet request enviroment. Hostname and port can be statically overridden using tomcat connector attributes. Alternatively if the request path (as opposed to host/port/protocol) is not changed, you could avoid request rewriting (can still do request routing or course) at the proxy so the web service sees the true URL via either http or ajp). — 

Re: asadmin in v3: Requesting user feedback on functional spec ...
First of all let me just say that I love all the proposed enhancements and new features. Given that as developers we spend a great deal of time working with asadmin, it's great to see that the tool is getting attention in v3 release. [...] One thing that has always bugged me was an inability to do a domain restart with one command. I don't know why such a simple, yet commonly needed feature is missing. —