I have a SL navigation application, that currently runs on a shared hosting package with a 3rd party ISP. I can login, and register using the ASP.NET membership and role providers.
I have now setup a dedicated server, on which only my app will run. It does not yet have a domain name that points to it... I access it via an IP address.
I've copied the entire site (including the ClientBin and all the XAP's) to the new server, but the Authentication and Registration services don't work... they just return NotFound.
When I check Fiddler on the working site, this service is called :
www.myaddress.com/ClientBin/MyApp-Web-AuthenticationService.svc/binary/Login
which of course succeeds. However, on the other site, the fiddler trace looks the same (because I just copied the site) :
123.123.123.123/ClientBin/MyApp-Web-AuthenticationService.svc/binary/Login
but, the call fails with NotFound. Fiddler reports it as HTTP/1.1 500 Internal Server Error. When I open
http://localhost/ClientBin/MyApp-Web-AuthenticationService.svc/binary/Login
on the server, I get the HTTP/1.1 500, as well as this description :
Handler "svc-Integrated" has a bad module "ManagedPipelineHandler" in its module list
Which leads me to believe that there is something wrong with my IIS config, as the exact same code is working on another system.
What is a "bad module"? How do I fix it?
Normally this type of error is that ASP.Net is not activated or that a handler for svc is not registered or registered correctly.
In your case is looks a bit different. It could be that you have .net framework 4.0 code that you are trying to run in a .net framework 2.0 application pool.
IIS 500 errors often show up with more information in the Windows Event log - if you can somehow get access to that?
Also I've noticed that often the server will actually send debugging output back to the client that everything seems to ignore. Have you checked the entire raw response that is coming back from the server to see if there are any clues there?
Related
I am developing a .Net Core (3.1) web application hosted with IIS 10. When I am trying to use SignalR on localhost, I am able to connect totally fine and receive messages from the backend, but after I publish it to IIS, I get the error: "Server returned handshake error: Handshake was canceled." In the client logs on Chrome's console, I get the message: "Information: SSE connected to http://myserver.com/MyApplication/output" before I get the error described above (which you would think means that I connected successfully)...
I have seen other posts where people have suggested that I have to enable webSockets on IIS, and I have already checked the my site has this enabled.
I have also seen people suggest to try using the longPolling argument in the withUrl function when creating the connection, and this has not helped either.
I have also added SignalR Event Log Trace Listeners to my web.config file, and the only warning I get that might be related to this issue is that I get a warning from the .NET runtime that reads "Failed to determine the https port for redirect." I am not sure what this means or if it would be related to SignalR, but as this also appears to deal with Middleware, it might be significant.
One thing that I thought might be related is that my site is just one of multiple hosted on this server, so that when I am running locally, I can use the following to connect to SignalR:
hubConnection = new signalR.HubConnectionBuilder().withUrl("/output").build();
But because of the way this site is deployed on my server, I have to use this:
hubConnection = new signalR.HubConnectionBuilder().withUrl("/MyApplication/output").build();
when I deploy the site or else I get a 404 error...
Would there be any issue there? I noticed when inspecting the network requests that on localhost, the URL to connect with the socket is "wss://localhost:44315/output?id=..." and I get the "Status Code: 101 Switching Protocols", but after I deploy, the url that is used is "http://myserver.com/MyApplication/output?id=..." and I get a "200 OK" just before receiving the error described at the top. Why does wss get used on localhost but http is requested from my client when I deploy?
FINALLY discovered the issue after days of debugging... The reason I was getting a "101" response on localhost and not when published is because 101 is basically the server (IIS in my case) saying "I support web sockets". I realized that because I wasn't getting this when published, web sockets were probably not enabled in IIS. It was weird, however, because I went into IIS manager and it said that web sockets were enabled, but after much more research, I found that you also have to go into "Server Manager" and enable it as well (see here https://learn.microsoft.com/en-us/iis/get-started/whats-new-in-iis-8/iis-80-websocket-protocol-support and follow step by step instructions). I did not realize both were required.
I'm having problems to find enough information to figure out what's missing on my published Silverlight application.
Running the applicaton locally (but with the db connectionstring to the "live" database), everything is running fine.
However, when I try to run at the published remote site it fails. It seems to be related to the db connection, but I don't get any detailed error from the server.
I've added a "debug" code at RIA service completed method
if (myLoadOperation.HasError)
{
MessageBox.Show(myLoadOperation.Error.Message);
MessageBox.Show(myLoadOperation.Error.InnerException.ToString());
}
The messagebox displays:
Load operation fail for the query 'GetData'. The remote server returned an error : NotFound
What is not found??!?
Additional information I get from JIT Debugger (VS2010)
NotFound. at System.ServiceModel.DomainServices.Client.OperationBase.Complete(Exception error)
I suspect that it might be a dll related to EF or RIA which may be missing on the remote server, but I'm not quite sure how to identify what's wrong. It's a shared hosting environment.
Could it be located to the MySQL provider?
connectionString="metadata=res://*/Model.Model1.csdl|res://*/Model.Model1.ssdl|res://*/Model.Model1.msl;provider=MySql.Data.MySqlClient;provider connection string="server=xxxxxxxxx;User Id=xxxxxxx;password=xxxxxxxxx;Persist Security Info=True;Allow Zero Datetime=True;Convert Zero Datetime=true;database=xxxxxxxxxx;" providerName="System.Data.EntityClient"
I've uploaded the following dll's to the server's bin folder:
System.ServiceModel.DomainServices.EntityFramework.dll
System.ServiceModel.DomainServices.Hosting.dll
System.ServiceModel.DomainServices.Hosting.OData.dll
System.ServiceModel.DomainServices.Hosting.Server.dll
I found the blog post WCF RIA Services - “Not Found” Error Message to offer several methods of debugging this very generic error message. Using tools like Fiddler and WCF Binary-encoded Message Inspector for Fiddler should give you a better understanding of what the real error message is.
I have a client who installed our silverlight app recently. it works fine from their server itself. but when they try to run it from a client, they can log in through the aspx login page, but on the main page, which hosts the tag and the .xap file, they see absolutely nothing!
I cannot see their screens, just get occasional screenshots via email and cross my fingers that they are typing the URL I tell them to. Even Shareview is not working for them - they can see my screen but I cannot see theirs.
So I am pleading - help! Please throw out some wacky ideas. I just learned an hour or so ago that they did not even have silverlight installed, so the mornings debugging effort was a waste of time. So who knows what the next fascinating source of problems is?
Here is the user-agent info. Oor app is .net 4.0 could that be the problem? It does noty look to my untrained eye that the client supports 4.0 (from the web server log):
Mozilla/4.0+(compatible;+MSIE+7.0;+Windows+NT+5.1;+.NET+CLR+1.1.4322;+.NET+CLR+2.0.50727;+.NET+CLR+3.0.04506.30;+.NET+CLR+3.0.04506.648;+.NET+CLR+3.0.4506.2152;+.NET+CLR+3.5.30729) 401 3 5 0
Getting closer - I see that the GET request for the XAP is returning 401 - not found! What would do that?
It's very possible that they are blocking XAP files either directly or indirectly at the firewall. XAP files are simply ZIP files and inspection-based firewalls tend to look at these as a security risk. You may want to see if they can setup an exception in their rules.
Hmm..
1) check if your client access the good url. Request their IP address, then check the web server logs. ie: are they requesting the good url, are they downloading the xap? (check if they use a proxy, too?)
2) check that they restarted their browser after installing the Silverlight runtime.
3) Do you handle the UnhandledException event for your Application class? If no, use it to send detailed exception logs to your server on application crash.
4) Use javascriptto initialize your Silverlight application. This way, you can be notified if the runtime fail to start, (for example if it failed to load the xap file) You can use ajax to report the issue to the webserver.
for days now, I try to get the following scenario running:
I have a Silverlight 4 Application that calls a WCF-Service via HTTPS. The WCF-Service is located on one of our servers, the Silverlight 4 app is also served from this server. After several hours of trial and error (well actually more error than try) I get it running when starting it from my VS2010 development environment on my local machine (see also SecurityError when calling a HTTPS-WCF Service from Silverlight 4).
But ...
The call to the https-WCF ONLY works when I start the SL-App from my VS2010 AND when fiddler is running and after I tell fiddler to ignore the Certificate Error (I use a self-signed certificate). When fiddler is not running, I get a security error. That applies, when my browser is the IE. When I copy the localhost address from the IE and paste it in my opera, it works without fiddler.
When I deploy the Silverlight app to the server and call it from the server, the WCF-call doesn't work, whether fiddler is running or not. I just get the standard Message, that an exception has occurred. This happens in both Opera and IE. Fiddler doesn't ask me to ignore the Certificate Error like it does in the VS2010 scenario, so maybe there is somewhere the key to my problem. Unfortunately, I have not much experiences with this stuff, so I have no idea, where to look or what to do.
I already found something about the fiddler issue here on stackoverflow ( Silverlight Requests, Failures & Fiddler), but it doesn't help me solving my problem.
I hope that some of you can shed some light in this issue, before all my hair is turned to gray.
Thanks in advance,
Frank
Edith has just installed the Certificate as a trusted Certificate in IE - but I still get the error, when calling the WCF when starting the deployed silverlight app from the server. At least I can call the Service when I start the SL app from VS2010 in the IE now (even when that doesn't help me, when some one else has to work with the app.
Several thoughts:
What is the exact error message you're seeing? Any inner exceptions?
Could this be a cross-domain issue? (Are the Silverlight XAP and the WCF service hosted from exactly the same domain, including its HTTP vs HTTPS characteristic, port number, distinction between "localhost" and an actual machine name, etc? What is the exact address as it appears in ServiceReference.ClientConfig or in code, and does its domain name exactly match the XAP's domain name?)
Do you see any call attempting to go through in Fiddler at all when it fails? (If not, I would strongly suspect a cross-domain issue)
If you manually browse to the https://....../YourService.svc from IE/Opera (instead of invoking the service through code in Silverlight), can you successfully see the service information page, with no certificate warnings/errors?
Does your SSL certificate has the same name as your service host? If not, that causes an error which can lead to the described behavior
I have a silverlight business aplication that gets data from silverlight enabled webservice.
When I run the application in dev environment, it works fine.
when i deploy the application and the Asp.net web development server is working, then to the application works fine.
But when I stop the development server, the application can't access a service.
My questions are:
When I deploy a silverlight business application, doesn't the service get deployed and get started.
The endpoint address in my ServiceReferences.Clientconfig file is endpoint address="http://localhost:9702/MyWebservice.scv. Do I need to change this?
The enpoint address in the web.config is blank.
Appreciate your help
Because the WCF client code is declared as a "partial" class, what I've been doing to this point is creating another c# partial class file to host a GetClient() method on it. You'll notice that the code is taking into account the port that the service is on... in a few of the environments that I've posted or will be posting to, as well as the development environment, the application is not always on port 80.
Namespace Project.Service
{
public partial class ServiceClient
{
public static ServiceClient GetClient()
{
return new ServiceClient("CustomBinding_Service",
new System.ServiceModel.EndpointAddress(new Uri(string.Format("{0}://{1}:{2}/Services/Service.svc",
Application.Current.Host.Source.Scheme, Application.Current.Host.Source.Host, Application.Current.Host.Source.Port), UriKind.Absolute)));
}
}
}
Hope this helps someone!
Yes you are going to want to change your endpoint address. I recommend doing it in the silverlight code when creating the connection to the WCF service. The service itself lives on the web server, whereas the silverlight application lives on the clients computer. If the web server stops, the web service stops but the silverlight app can keep running.
edit:
To do this in code, as long as the path is always in the same domain as the app you can use do like so:
BasicHttpBinding binding = new BasicHttpBinding(BasicHttpSecurityMode.None) //Use whatever security you need here
{
MaxReceivedMessageSize = int.MaxValue,
MaxBufferSize = int.MaxValue
};
Client client = new Client(binding, new EndpointAddress(new Uri(Application.Current.Host.Source, "../MyService.svc")));
Thanks so much for your help. I tried your approach to create the client code but that didn't work. And that's because the problem seems to be somewhere else.
So I installed fiddler to see the traffic.
Fiddler showed that the service was being accessed but the http response code was 302 showing that there was some redirection involved.
The address of my application is like this http:///Silverlightapp/(S(oirppxrwzhlf2a2vbia1ui45))/Default.aspx#/Home and it is hosted on IIS 6.
So I had to employ a workaround by installing the service on machine with IIS7 (and there was no session id involved like in the above url).I still kept the silverlight application hosted on IIS 6.
Anyway, in summary, to anyone who reads the thread, I did the following to troubleshoot and solve issue(temporarily)
Changed the end point address in the ServiceReferences.ClientConfig file. When you add the service using discover option in VS, the endpoint address is of the localhost and this must be changed.
Registered the service model using ServiceModelReg -i command. (this solved my problem that my applicaiton was only working from development server and not IIS)
-Put the CrossDomain and ClientAccessPolicy files in c:]inetpub\wwwroot folders.
-Used fiddler to look at http response codes. I had to do no configuration in fiddler.
Changed the binarymessageEncoding to textMessageEncoding iin the web.config file of the silverlight web project that also hosted the ecf service. I did this becasue adding a silverlight enabled wcf service creates a custom binding configuration in the web.config file by default uses binary encoding. I needed text encoding to see errors in fidder. But this didn't help much becasue I only saw the name of the operation in the Inspector>xml tab in fiddler. This was the same even after my issues was resolved by workaround.
Thanks for the help
Don't do it in code. Otherwise you won't be able to change it later without recompiling the application (when the address will need to change, perhaps years down the road when you've lost the source code :)
Change the address in ServiceReferences.ClientConfig to where the service is actually hosted... e.g. http://example.com/myVdir/MyWebservice.svc
If later on you need to change the address without recompiling:
- Open the .xap file (it's just a zip file with a different extension)
- Find the .ClientConfig file and change the address
- Put it back together as a .zip file and rename to .xap
Also, I can't remember anymore whether the .ClientConfig supports relative addresses (e.g. just "MyWebService.svc"), but if it does it may be a good solution as well.