I want to log Outgoing (SOAP) messages generated by Web service in response of request sent by client (on server side).
Using Service model clients are able to log the response into the database by applying settings into the web.config of the web application e.g.
< system.servicemodel >
< extensions>
< behaviors>
< bindings> ...
Please guide me how to achieve logging and configuration on server side.
You can enable diagnostic tracing :
Just need to add a section in web.config -
<system.diagnostics>
<sources>
<source name="System.ServiceModel"
switchValue="Information, ActivityTracing"
propagateActivity="true">
<listeners>
<add name="traceListener"
type="System.Diagnostics.XmlWriterTraceListener"
initializeData= "F:\Sony\Traces.svclog" />
</listeners>
</source>
</sources>
</system.diagnostics>
http://sonyarouje.com/2011/11/10/diagnostic-trace-svclog-configuration-for-wcf/
http://msdn.microsoft.com/en-us/library/ms732023.aspx
Also check this related question How to turn on WCF tracing?
To see the actual SOAP envelope that is passed around
http://litemedia.info/debug-soap-request-and-response-in-wcf
Also found one more similar question How can I enable WCF logging so that it writes to a Database?
I have a Windows forms app which uses WCF services. Our application sends messages using one of our WCF services to specific users running our client, so our callback “http:” string is dynamically constructed each time a message is sent to a user. It includes the server IP address and port (126.221.97.105:701) onto which the current user is logged, the user’s id (56281), and the client GUID (7392d27a-e4a0-42e2-89a3-adc332e28934). So, a typical callback “http:” string looks like this:
http://xxx.xxx.xx.xxx:701/CmesCns/CALLBACK/56281/7392d27a-e4a0-42e2-89a3-adc332e28934
We have an http namespace (http://+:701/) on our client and the group “Everyone” is tied to this namespace with all of the access permissions checked (GenericAll, GenericExecute, GenericRead, and GenericWrite). We use “http namespace” to create our namespaces.
Our application has been in production (on Windows Server 2003) for a few years and everything is working fine.
We have recently converted our application to run in the Windows 2008 server environment. The “Target Framework” in each of our projects is set to the “.NET Framework 4.0”. Our application works fine on my Windows 7 developer workstation. That is, I am able to receive messages from our WCF service, but when I place our application onto our Windows 2008 server and I attempt to run the application, I receive the following error message:
"There was no endpoint listening at http://xxx.xxx.xx.xxx:701/CmesCns/CALLBACK/56281/7392d27a-e4a0-42e2-89a3-adc332e28934
that could accept the message. This is often caused by an incorrect address or SOAP action. See InnerException, if present, for more details.”
The http namespace (http://+:701/) exists on my developer workstation and on my Windows 2008 server. The group “Everyone” is tied the namespace on my Windows 7 box and on my Windows 2008 server, and all of the access permissions are checked (GenericAll, GenericExecute, GenericRead, and GenericWrite).
We have been searching the web for an answer but have not discovered anything. Would anybody have any ideas on why this would work on our Windows 7 workstations, but not on our Windows 2008 servers?
Any help is greatly appreciated!
Kevin
When you host a WCF service in IIS you don't specify an absolute url in the address. You should use a relative url to the .svc file. The base url will be determined by the web site where it is hosted.
<service name="WebService.Receptor">
<endpoint
address="/WS.svc"
binding="wsHttpBinding"
contract="IMyContract"
/>
and on the client, depending on how your IIS is configured you should obviously specify the full address:
<client>
<endpoint
name="Receptor"
address="http://MyServer:8000/WS.svc"
binding="wsHttpBinding"
contract="IMyContract"
/>
This assumes that you have configured a site in IIS that listens on the 8000 port and that you have hosted your WCF application inside this site.
if it does not help please follow these links, hope it would be useful.
Stack overflow link
Multiple Endpoint
Typically, this error is because there is no endpoint on the server that matches what the client is requesting (the address, the service, or the authentication is different).
However, in my case, I had the exact same error, and it was not due to any of these things.
When I enabled the tracing on IIS and reviewed the svclog trace with SvcTraceViewer.exe (included in Visual Studio), the actual internal error was "Maximum request length exceeded."
My client was uploading an image via the service. And I guess the image was too big.
To enable tracing I added this to the configuration section:
<system.diagnostics>
<sources>
<source name="System.ServiceModel"
switchValue="All"
propagateActivity="true">
<listeners>
<add name="traceListener"
type="System.Diagnostics.XmlWriterTraceListener"
initializeData= "c:\log\Traces.svclog" />
</listeners>
</source>
</sources>
To solve the error, I increased the message request length in the web config and the error went away.
To do this, in the system.websection in the web.config I added the line:
<httpRuntime maxRequestLength="32768" />
Then I added this section inside the configuration section
<system.webServer>
<security>
<requestFiltering>
<requestLimits maxAllowedContentLength="32000000" />
</requestFiltering>
</security>
</system.webServer>
So I recommend you enable tracing and then review the trace for the exact error.
I have a WCF service with 2 endpoints using BasicHttpBinding running on an instance in Azure. When the service is called for the first time it takes 13-16 seconds for a response (which is to be expected) and subsequent requests are adequate at between 150ms and 1000ms depending on the request type (I am using soapUI for testing).
But if I do not send any requests for a few minutes (less than 4 or 5 minutes) the service instance does not appear to be getting reused as it is taking as long as the first request of 13+ seconds.
I have tried "InstanceContextMode = InstanceContextMode.Single" in my ServiceBehavior(s) but that did not fix the problem.
In the trace I can see that the thread id changes whenever the service gives a slow response.
I have also logged the time from when Application_BeginRequest is hit to the start of the method being called by the client which seems to be the largest delay during these 'slow' calls.
I have some fairly large classes representing business interfaces and am using Entity Framework for a MySQL database running on an Azure VM. (Although I don't think this is necessarily an Azure issue).
The slow responses are the same whether I am calling a complex message type linked to DB operations or just a simple "Hello World" response.
Even if a new instance of the service is required should it take as long as an initial load after an IIS restart or application pool recycle?
Would doing some magic using wrappers for Static classes help? (I'm thinking not).
Built with C# 4.0 / VS2010 Professional
There can be lots of reasons why this is happening. I suggest you start with some tracing in your WCF service.
Add some logging in your application using TraceSource:
private TraceSource ts = new TraceSource("MyApp");
public string GetData(int value)
{
ts.TraceInformation("GetData called with {0}", value);
return string.Format("You entered: {0}", value);
}
Configure the TraceSource in your web.config together with the System.ServiceModel source.
<system.diagnostics>
<sources>
<source name="MyApp" switchValue="All">
<listeners>
<add name="xml" />
</listeners>
</source>
<source name="System.ServiceModel" switchValue="All" propagateActivity="true">
<listeners>
<add name="xml" />
</listeners>
</source>
</sources>
<sharedListeners>
<add initializeData="Logs.svclog" type="System.Diagnostics.XmlWriterTraceListener" name="xml" />
</sharedListeners>
<trace autoflush="true" />
</system.diagnostics>
Open the svclog file and you'll be able to see a detailed overview of everything that's happing in your service and how long each little step takes. This should help you to poinpoint the issue.
I'm having an interesting compatibility issue between a WCF client and a Java web service. In short I've found that the way the header is generated is causing the problem - the ActivityId and Action elements in the header as well as what WCF is doing with the namespace of the custom header is causing issues. I've successfully consumed the WSDL with wsdl.exe, but WCF seems to be manipulating the header in a way that the Java web service doesn't like. Is there any way I can set up the bindings for the WCF client to not send the ActivityId and Action elements?
Do you have tracing turned on in the client? I think that is what is adding the activity ID as its trying to flow the tracing activity to the service for end to end tracing. Turn off the activity tracing flag and it should go - see my comment for the action header
This issue commonly occurs when a WCF client attempts to connect to a non-WCF server, e.g. JAX-WS, Websphere etc.
Just to add to Richard's lifesaver answer and address #irperez's comment, the actual settings which need to be disabled to prevent WCF diagnostics from adding ActivityId during WCF Diagnostic Tracing are to remove:
Remove ActivityTracing from switchvalue
Set propagateActivity to false
i.e. Change
<system.diagnostics>
<sources>
<source name="System.ServiceModel" switchValue="Information, ActivityTracing"
propagateActivity="true">
<listeners>
<add name="xml"/>
</listeners>
</source>
...
To:
<source name="System.ServiceModel" switchValue="Information"
propagateActivity="false">
<listeners>
<add name="xml"/>
</listeners>
If the ActivityId is enabled, it injects the below into the SOAP headers, which can break unsuspecting servers:
<s:Envelope xmlns:s="http://schemas.xmlsoap.org/soap/envelope/">
<s:Header>
<ActivityId CorrelationId="5de75017-da08-4ac2-84f2-5374953cc2a1"
xmlns="http://schemas.microsoft.com/2004/09/ServiceModel/Diagnostics">
9f076849-e76e-4675-84c1-5026b1c2eb1a
</ActivityId>
</s:Header>
I have a WCF service that accepts a complex type and returns some data. I want to use Fiddler to see what the incoming requests to the service looks like. The client is .net console app which uses a Service reference proxy. Is this possible with Fiddler. I'm new to this tool and have only used it in the past to post data with the request builder.
You need to add this in your web.config
<system.net>
<defaultProxy>
<proxy bypassonlocal="False" usesystemdefault="True" proxyaddress="http://127.0.0.1:8888" />
</defaultProxy>
</system.net>
then Start Fiddler on the WEBSERVER machine.
Click Tools | Fiddler Options => Connections => adjust the port as 8888.(allow remote if you need that)
Ok, then from file menu, capture the traffic.
That's all, but don't forget to remove the web.config lines after closing the fiddler, because if you don't it will make an error.
Reference : http://fiddler2.com/documentation/Configure-Fiddler/Tasks/UseFiddlerAsReverseProxy
Fiddler listens to outbound requests rather than inbound requests so you're not going to be able to monitor all the requests coming in to your service by using Fiddler.
The best you're going to get with Fiddler is the ability to see all of the requests as they are generated by your Console App (assuming that the app generates web requests rather than using some other pipeline).
If you want a tool that is more powerful (but more difficult to use) that will allow you to monitor ALL incoming requests, you should check out WireShark.
Edit
I stand corrected. Thanks to Eric Law for posting the directions to configuring Fiddler to be a reverse proxy!
Just had this problem, what worked for me was to use localhost.fiddler:
<endpoint address="http://localhost.fiddler/test/test.svc"
binding="basicHttpBinding"
bindingConfiguration="customBinding"
contract="test"
name="customBinding"/>
Consolidating the caveats mentioned in comments/answers for several use cases.
Mostly, see http://docs.telerik.com/fiddler/Configure-Fiddler/Tasks/ConfigureDotNETApp
Start Fiddler before your app
In a console app, you might not need to specify the proxyaddress:
<proxy bypassonlocal="False" usesystemdefault="True" />
In a web application / something hosted in IIS, you need to add the proxyaddress:
<proxy bypassonlocal="False" usesystemdefault="True" proxyaddress="http://127.0.0.1:8888" />
When .NET makes a request (through a service client or HttpWebRequest, etc) it will always bypass the Fiddler proxy for URLs containing localhost, so you must use an alias like the machine name or make up something in your 'hosts' file (which is why something like localhost.fiddler or http://HOSTNAME works)
If you specify the proxyaddress, you must remove it from your config if Fiddler isn't on, or any requests your app makes will throw an exception like:
No connection could be made because the target machine actively refused it 127.0.0.1:8888
Don't forget to use config transformations to remove the proxy section in production
So simple, all you need is to change the address in the config client: instead of 'localhost' change to the machine name or IP
This is straightforward if you have control over the client that is sending the communications. All you need to do is set the HttpProxy on the client-side service class.
I did this, for example, to trace a web service client running on a smartphone. I set the proxy on that client-side connection to the IP/port of Fiddler, which was running on a PC on the network. The smartphone app then sent all of its outgoing communication to the web service, through Fiddler.
This worked perfectly.
If your client is a WCF client, then see this Q&A for how to set the proxy.
Even if you don't have the ability to modify the code of the client-side app, you may be able to set the proxy administratively, depending on the webservices stack your client uses.
Standard WCF Tracing/Diagnostics
If for some reason you are unable to get Fiddler to work, or would rather log the requests another way, another option is to use the standard WCF tracing functionality. This will produce a file that has a nice viewer.
Docs
See https://learn.microsoft.com/en-us/dotnet/framework/wcf/samples/tracing-and-message-logging
Configuration
Add the following to your config, make sure c:\logs exists, rebuild, and make requests:
<system.serviceModel>
<diagnostics>
<!-- Enable Message Logging here. -->
<!-- log all messages received or sent at the transport or service model levels -->
<messageLogging logEntireMessage="true"
maxMessagesToLog="300"
logMessagesAtServiceLevel="true"
logMalformedMessages="true"
logMessagesAtTransportLevel="true" />
</diagnostics>
</system.serviceModel>
<system.diagnostics>
<sources>
<source name="System.ServiceModel" switchValue="Information,ActivityTracing"
propagateActivity="true">
<listeners>
<add name="xml" />
</listeners>
</source>
<source name="System.ServiceModel.MessageLogging">
<listeners>
<add name="xml" />
</listeners>
</source>
</sources>
<sharedListeners>
<add initializeData="C:\logs\TracingAndLogging-client.svclog" type="System.Diagnostics.XmlWriterTraceListener"
name="xml" />
</sharedListeners>
<trace autoflush="true" />
</system.diagnostics>
I have used wire shark tool for monitoring service calls from silver light app in browser to service. try the link gives clear info
It enables you to monitor the whole request and response contents.
I just tried the first answer from Brad Rem and came to this setting in the web.config under BasicHttpBinding:
<system.serviceModel>
<bindings>
<basicHttpBinding>
<binding bypassProxyOnLocal="False" useDefaultWebProxy="false" proxyAddress="http://127.0.0.1:8888" ...
...
</basicHttpBinding>
</bindings>
...
<system.serviceModel>
Hope this helps someone.
You can use the Free version of HTTP Debugger.
It is not a proxy and you needn't make any changes in web.config.
Also, it can show both; incoming and outgoing HTTP requests.
HTTP Debugger Free
Use fiddler a Reverse Proxy is the final solution for me.
First, configure fiddler as reverse proxy with REGDIT, like the doc said: https://docs.telerik.com/fiddler/configure-fiddler/tasks/usefiddlerasreverseproxy#configure-fiddler-as-reverse-proxy
1)Click Tools > Fiddler Options. Ensure Allow remote clients to connect is checked
2)Create a new DWORD named ReverseProxyForPort inside HKEY_CURRENT_USER\SOFTWARE\Microsoft\Fiddler2.
3)Set the DWORD to the local port where Fiddler will re-route inbound traffic.
4)Restart Fiddler.
Second, change the client to call service through proxy
for example , here is my client app.config:
<client>
<endpoint address="http://localhost:61236/WeatherForecastService.svc"
binding="basicHttpBinding" bindingConfiguration="BasicHttpBinding_IWeatherForecastService"
contract="ServiceReference1.IWeatherForecastService" name="BasicHttpBinding_IWeatherForecastService" />
</client>
change the client to use proxy endpoint address.
WeatherForecastServiceClient client = new WeatherForecastServiceClient("BasicHttpBinding_IWeatherForecastService", "http://localhost:8888/WeatherForecastService.svc");
var data = client.GetData(1000);
client.Close();
Change the localhost in the URL to localhost.fiddler, this small change worked for me.
Also if anyone testing the service from WCF Test Client don't forget to edit the URL in the config endpoint
Right click on the config file
Click Edit with Svc Config Editor
Click on Endpoints and edit the endpoint to localhost.fiddler
Check Start a new proxy while calling method