soap:timeout on soap webservice reference

[boanergesza11]

Hi Everybody;

We are very close to releasing our first production application using SwitchYard. We are wrapping up some small outstanding issues, one of them is setting a timeout on our soap webservice request we make.

The documentation https://docs.jboss.org/author/display/SWITCHYARD/SOAP says that we can just add <soap:timeout>, however when I do that I get the following at build time.

Failed to execute goal org.switchyard:switchyard-plugin:1.0.0.Final:configure (default) on project PaymentGateway: Model [org.switchyard.config.model.switchyard.v1.V1SwitchYardModel] is invalid: cvc-complex-type.2.4.a: Invalid content was found starting with element 'soap:timeout'. One of '{"urn:switchyard-component-soap:config:1.0":basic, "urn:switchyard-component-soap:config:1.0":ntlm, "urn:switchyard-component-soap:config:1.0":proxy, "urn:switchyard-component-soap:config:1.0":mtom, "urn:switchyard-component-soap:config:1.0":inInterceptors, "urn:switchyard-component-soap:config:1.0":outInterceptors}' is expected. -> [Help 1]

Here is our reference in the switchyard.xml :

    <sca:reference name="WsBICISOPortType" multiplicity="0..1" promote="Base24Bean/WsBICISOPortType">

      <sca:interface.wsdl interface="wsdl/BICISOWSDL4.wsdl#wsdl.porttype(wsBICISOPortType)"/>

      <soap:binding.soap>

        <soap:messageComposer unwrapped="true"/>

        <soap:wsdl>wsdl/BICISOWSDL4.wsdl</soap:wsdl>

        <soap:wsdlPort>wsBICISOPort</soap:wsdlPort>

        <soap:endpointAddress>https://172.18.191.200:31340/soap</soap:endpointAddress>

        <soap:timeout>30</soap:timeout>

      </soap:binding.soap>

    </sca:reference>

[kcbabo]

The documentation link you referenced is actually our workspace for docs on the current release.  We update that space during a release and then create a distinct space for each released version.  So for 1.0.0.Final, that would be:

Home - SwitchYard 1.0 - Project Documentation Editor

The timeout parameter was added in 1.1, which is why you see it in the workspace version of the docs, but not in the 1.0.0.Final version of the docs.  If the need for timeout support is critical for you, then you might want to wait for 1.1.0.Final to be released.  I'm actually staging the Maven artifacts right now, so I would expect the release to be announced/ready within a week.  Congrats on the move to production!

cheers,

keith

[boanergesza11]

Excellent, thanks Keith.

I now understand the documentation setup.

It is important, but we can definitely wait for 1.1. I'll keep an eye out for it.

[boanergesza11]

Hi Keith;

Happy new year to all the forumites!

We installed 1.1 and the tooling, set the timeout on the reference to 30 milliseconds. However what happens now is that it times out after 60 seconds?

Below examples of what I mean

03 Jan 2014 10:51:26,134 INFO  [za.co.fnb.transformers.Transformers.log ...

03 Jan 2014 10:52:28,064 WARNING [org.apache.cxf.phase.PhaseInterceptorChain.doDefaultLogging 364] (http-/0.0.0.0:8080-1) Interceptor for {http://tempuri.org/wsBICISO/definitions}wsBICISO#{http://cxf.apache.org/jaxws/dispatch}Invoke has thrown exception, unwinding now: org.apache.cxf.interceptor.Fault: Could not send Message.

03 Jan 2014 11:03:41,184 INFO  [za.co.fnb.transformers.Transformers.log ...

03 Jan 2014 11:04:42,262 WARNING [org.apache.cxf.phase.PhaseInterceptorChain.doDefaultLogging 364] (http-/0.0.0.0:8080-1) Interceptor for {http://tempuri.org/wsBICISO/definitions}wsBICISO#{http://cxf.apache.org/jaxws/dispatch}Invoke has thrown exception, unwinding now: org.apache.cxf.interceptor.Fault: Could not send Message.

[mageshbk]

Could you please enable trace logging and attach the log here? Please share the switchyard.xml too.

regards,

Magesh

[boanergesza11]

Thanks Magesh;

Please see attached

[mageshbk]

Hi Gerhard,

Forgive my manners. Wish you too a Happy New Year!

I could see the request is timing out after 60 seconds as indicated in your log

08 Jan 2014 09:57:51,080 FINE  [org.apache.cxf.transport.http.HTTPConduit.onFirstWrite 1367] (http-/0.0.0.0:8080-1) Sending POST Message with Headers to https://172.18.191.200:31340/soap Conduit :{http://tempuri.org/wsBICISO/definitions}wsBICISOPort.http-conduit
...
08 Jan 2014 09:58:51,645 WARNING [org.apache.cxf.phase.PhaseInterceptorChain.doDefaultLogging 364] (http-/0.0.0.0:8080-1) Interceptor for {http://tempuri.org/wsBICISO/definitions}wsBICISO#{http://cxf.apache.org/jaxws/dispatch}Invoke has thrown exception, unwinding now: org.apache.cxf.interceptor.Fault: Could not send Message.

But what I see is that there are many Quartz logs in between that is puzzling me. I requested you switchyard.xml file and not the standalone.xml. I need to see how you have set the value for timeout and what else is there in the reference binding. I do not suspect the timeout being an issue as you can see we do test it unit wise inside the SOAP component here:

components/soap/src/test/java/org/switchyard/component/soap/SOAPGatewayTest.java at master · jboss-switchyard/components…

regards,

Magesh

[boanergesza11]

Thanks Magesh;

Yes apologies, you did say switchyard.xml. The many Quartz logs is for a periodic check we do internally for another service (theres two!) Please find attached the switchyard.xml

[mageshbk]

Hi Gerhard,

To isolate the issue, could you disable the cron services and share the log?

[boanergesza11]

I disabled the cron, turned on full debugging and reran our test. The action starts at "10 Jan 2014 09:02:45,937" end then ends at "10 Jan 2014 09:03:48,177"

[mageshbk]

Gerhard,

I am unable to replicate this issue on my local machine. Are you attaching some kind of debugger or are you clicking on the console window briefly? Something definitely is triggered at that exact moment when the reference is called. I have attached a modified version of soap-binding-rpc QS and it's execution log on a Windows machine. I can see the timeout happens immediately as it is set to 10. I am afraid I can't help further unless I am having an access to your machine to test your application and analyze your environment and setup. Can you test only this attached QS and see if that still stalls fo 60 secs?

thanks and regards,

Magesh

[boanergesza11]

Hi Magesh;

I looked at the log file above, Im curious about something. Why is the timestamp on the exeception before the timestamp of the line above it?

12:03:26,103 FINE  [org.apache.cxf.phase.PhaseInterceptorChain] (http-/127.0.0.1:8080-2) Invoking handleMessage on interceptor org.apache.cxf.interceptor.StaxInEndingInterceptor@867e05

12:03:25,661 WARNING [org.apache.cxf.phase.PhaseInterceptorChain] (http-/127.0.0.1:8080-1) Interceptor for {urn:switchyard-quickstart:soap-binding-rpc:1.0}HelloWorldWSService#{http://cxf.apache.org/jaxws/dispatch}Invoke has thrown exception, unwinding now: org.apache.cxf.interceptor.Fault: Could not send Message.

[mageshbk]

I believe the log formatter takes a small fraction of time to format the stack traces in the WARNING and a few logs get *flushed* to the file during this fraction of time.

[boanergesza11]

HI Magesh;

Thanks for your help on this trivial but important problem we are facing. I think I might have detected a problem in SwitchYard. If we connect to the soap webservice using http the timeouts work as expected. However, under SSL it times out in 60 seconds no matter what we specified.

WhenI looked at the test you showed me, I found this :

URL serviceURL = new URL("http://" + host + ":" + port + "/HelloSOAP12Service");

The test is also using http and not https

[mageshbk]

Nice find Gerhard! I did not notice this subtle difference of SSL url being used. I agree this is a bug. I believe that it could be one in CXF. I have created a JIRA to chase it up with the Webservices team or fix it in SwitchYard. Thank you for your investigation and support.

[SWITCHYARD-1931] Timeout configuration does not work for https connections - JBoss Issue Tracker

regards,

Magesh