On 2011-7-25, at 下午11:27, BRUNO MELLONI wrote:
We have a simple service that handles emailing for our web
applications. It takes the request information and sends the email
using SMTP through Exchange. It usually works great, but Web
service call timeouts have been a nuisance from the beginning.
If for some reason the application server where this application
runs gets overloaded (CPU utilization near 100%) it is common for
our client apps to get timeouts calling the service. The issue is
not related to our code, we already tested and response is quite fast.
This happens even though we already raised the client timeout to 120
seconds (which seems huge) and yet, this still happens... especially
during server startup. The following is the code we used:
<http-conf:conduit name="*.http-conduit">
<http-conf:client AllowChunking="false"
ConnectionTimeout="120000"/>
</http-conf:conduit>
Hi,
You need configure ReceiveTimeout in this case, take a look at [1] to
get more details.
[1]http://cxf.apache.org/docs/client-http-transport-including-ssl-support.html
Freeman
Is this kind of long web service delay normal during high CPU
utilization? Is there something that we need to configure to make
the web service more responsive even during high utilization of the
application server? Should we increase the timeout to a
ridiculously high 5 minutes to prevent the problem?
---------------------------------------------
Freeman Fang
FuseSource
Email:[email protected]
Web: fusesource.com
Twitter: freemanfang
Blog: http://freemanfang.blogspot.com
