Hi Harm
The problem was most definately in the ESB code.
The ESB got into a state where it didn't handle any messages anymore
(deadlock).
It also didn't handle any other messages for _other_ backends anymore.
Only restarting the ESB worked.
Meanwhile the backend (running on a separate tomcat) the was still
working. (without any restart).
*workaround*
whenever I fooled the ESB that the backend was running on a different
machine the ESB ran fine.(it really was still running on the same
machine (but rather then accessing it via 'localhost' rerouting the
traffic really through an other machine, where apache routing the
traffic back)
(so: ESB configuration has an endpoint on http://someotherhost/...
on someother host apache proxies the traffic back to the ESB host.)
Note: maybe this ticket has nothing to do with the original question
on WSAS/ESB on the same machine. I'm not sure.
The NIO code has a slight probability to behave differently with clients
on the same physical host and others - and this depends mainly on the
OS. (e.g. I first saw this couple of years ago with a very old Sun
Solaris). I do not want to comment on this specific issue without
further analysis, but like Ruwan said, quite a few issues have been
fixed on the trunk after the Synapse 1.2/ESB 1.7.x releases, with help
mainly from real users who use it in production
cheers
asankha
--
Asankha C. Perera
AdroitLogic, http://adroitlogic.org
http://esbmagic.blogspot.com
_______________________________________________
Esb-java-user mailing list
[email protected]
https://wso2.org/cgi-bin/mailman/listinfo/esb-java-user
