Hmmm. What would a client connection timeout of just 10 do? A receive timeout of 10? It would be good to shrink those values from the ones you have below to confirm there indeed a bug, i.e., the client is indeed never timing out.
Also, are you thinking of doing an async[1] web service? Not that it's fully relevant to your question, but it may be good to have it designed the way it will finally be before doing this stress-testing. HTH, Glen [1] http://www.jroller.com/gmazza/entry/creating_service_side_asynchronous_web XyLus wrote: > > Hello All, > > I'd like to simulate latency in my upload service for test purpose ( I > send message using MTOM mechanism). > > I have just added Thread.sleep(latency) to service implementation method > and everything works till I send small request. However when I send 20 MB > message then inovcation does not times out, even when server simulate 5 > minutes latency and client has following configuration: > > <http:conduit > name="{http://x.y.z/uploader}UploadServicePort.http-conduit";> > > <http:client AutoRedirect="true" ConnectionTimeout="10000" > ReceiveTimeout="10000" AllowChunking="false" /> > > </http:conduit> > > > One more thing I've been testing it on my localbox ( client and server are > hosted on the same box) > > Please let me know what should be changed in order to get timeouts for BIG > message. > I hope it is not a bug in CXF itself. > > > Regards, > Maciej > -- View this message in context: http://www.nabble.com/How-to-simulate-service-latency--tp23175136p23175814.html Sent from the cxf-user mailing list archive at Nabble.com.
