Hello Alex,
I think, this is a good question for the axis2 mailing list.
AFAIK, Axis2 have a module called Sandesha2 which supports reliable
Messaging out of the box according to the WS-ReliableMessaging-spec
which was released in final version in the middle of 2007. The
developers from wso2.org have developed a persistence mode for it, so
that it must be possible to get all messages, even if ODE or the
external service get an error and will be restarted. If you start an
async invocation with Axis2 over a seperate channel, the internal
simpleHTTPServer of Axis2 will be started and wait for an incoming
message. I think, the problem is how to restore the listener after a
crash...
I have started a thread in the nabble forum:
http://www.nabble.com/question-to-restore-callback-listener-after-crash-td15496163.html
Thanks,
Jens
Alex Boisvert schrieb:
On 2/14/08, Thilina Gunarathne <[EMAIL PROTECTED]> wrote:
Hi Alex,
Thanks a lot for the crystal clear explanation. I'll try to hack this in
my
system over the weekend and let you guys know if there are any
complications.
After sending my email, I started thinking about reliability aspects and
what would happen if a failure happened during the [Any amount of time]
portion.
How would you reliably handle an async response in Axis2? (Assuming the
callback is gone after a failure)
alex
