I'm converting an old .NET WCF service to Web API 2
To maintain backwards compatibility I have applied a RoutePrefixAttribute to my controller as shown:
All has gone pretty smoothly until I try to publish my service and access it via IIS
When I run my service via localhost (debugged from Visual Studio) and make a request via Postman all is well and I get the expected response:
However, after I publish the site to IIS, set a host entry and try to access the same endpoint:
I receive a 404 not found:
I did some playing around, and decided to remove the ".svc" from my RoutePrefixAttribute for my Controller. And voila, I can now hit my endpoint via IIS:
So my question is: Does Web API 2 not support the ".svc" or even perhaps periods in their routes? Has anyone encountered something similar and found a reasonable workaround?
Thanks
My issue was as Kiran Challa indicated. I had to add the following line to my system.webServer handlers:
<system.webServer>
<handlers>
<add name="ApiURIs-ISAPI-Integrated-4.0" path="*." verb="GET,HEAD,POST,DEBUG,PUT,DELETE,PATCH,OPTIONS" type="System.Web.Handlers.TransferRequestHandler" preCondition="integratedMode,runtimeVersionv4.0" />
</handlers>
</system.webServer>
More information here: Dots in URL causes 404 with ASP.NET mvc and IIS
Related
I'm using ASP.NET Core RC1 as server to host my Aurelia app. My app was working just fine but the last couple of weeks something changed so that the app does no longer load when hosted on Azure. I'm not sure if it is something I changed or if it's a change on the Azure side but I'm leaning towards the latter.
I've narrowed down the problem quite a bit. The app runs fine locally, with ASP.NET Core Kestrel server and also other servers (e.g. webpack-dev-server). I have continuous deployment setup from Visual Studio Team Services to an Azure Website. The app is published and a web.config is automatically created in my wwwroot:
<configuration>
<system.webServer>
<handlers>
<add name="httpplatformhandler" path="*" verb="*" modules="httpPlatformHandler" resourceType="Unspecified" />
</handlers>
<httpPlatform processPath="%home%\site\approot\web.cmd" arguments="" stdoutLogEnabled="true" stdoutLogFile="\\?\%home%\LogFiles\stdout.log"></httpPlatform>
</system.webServer>
</configuration>
Nothing happens when I navigate to my site, e.g. http://demo.azurewebsites.net/. When looking at the console I get a 404. Once I actually got this error but I can't seem to bring it back: 502 - Web server received an invalid response while acting as a gateway or proxy server
I have index.html set as default document but it is not loading. If I enter it explicitly, the app works: http://demo.azurewebsites.net/index.html
If I remove the httpplatformhandler from the web.config, then it works as expected (index.html is loaded automatically). The same happens when I remove the web.config entirely. In these cases the MVC 6 WebAPI behind the scenes does not work at all. I assume that's just logical since I remove the platform handler.
So, why is this httpplatformhandler added? Is it necessary? Why is it created? Is there some setting in the Azure portal that I can adjust to prevent this handler to be configured like this?
I also found this link that seems to suggest that things are changing and that this httpplatformhandler is about to be replaced: Closer Look: Hosting ASP.NET Core on Azure App Service
I'm out on deep water here and any and all help is appreciated.
To get default document support with the static file server middleware you need to use app.UseFileServer() instead of app.UseStaticFiles()
IIS Express is producing 403.14 Forbidden errors when a URL that would otherwise be handled through ASP.NET URL routing happens to correspond to a physical folder in my ASP.NET project. (The folder contains only code, and it's coincidental that the folder name happens to match the URL of a page; my URL structure is determined dynamically by a database, and users can edit that structure, so although I could just rename my project folder, in general I can't prevent this sort of collision occurring.)
This seems to be happening because the DirectoryListingModule steps in to handle the request, and then promptly fails it because directory browsing is disabled. I've tried removing this:
<system.webServer>
<handlers>
<remove name="StaticFile" />
<add name="StaticFile" path="*" verb="*"
modules="StaticFileModule" resourceType="Either" requireAccess="Read" />
</handlers>
</system.webServer>
That removes the default StaticFile handler configuration, which has modules="StaticFileModule,DefaultDocumentModule,DirectoryListingModule", and replaces it with a configuration that provides just the feature I want. (I want static file serving, but I have no need for directory listing or default documents in this app.) But the effect seems to be that IIS then produces a completely empty (0 byte) response (with a 200 status) when I hit the offending page.
So next, I tried configuring the StaticFile handler to handle only the specific physical folders that I want to make available:
<system.webServer>
<handlers>
<remove name="StaticFile" />
<add name="StaticFileCss" path="style/*.css" verb="*"
modules="StaticFileModule" resourceType="Either" requireAccess="Read" />
<add name="StaticFileScripts" path="Scripts/*" verb="*"
modules="StaticFileModule" resourceType="Either" requireAccess="Read" />
</handlers>
</system.webServer>
But when I hit the offending URL, this then produces a 404.4 - Not found error, with a message of The resource you are looking for does not have a handler associated with it.. (The Detailed Error Information on the error page says that we're in the IIS Web Core module, during the MapRequestHandler notification, the handler is Not yet determined, and there's an Error Code of 0x80070002, which is a COM HRESULT that corresponds to the Win32 ERROR_FILE_NOT_FOUND error.)
The mystifying thing is that it's not even bothering to ask ASP.NET whether it has a handler for it. IIS seems to be deciding all by itself that there definitely isn't a handler.
This only happens when there's a folder that matches the URL. All other resources with dynamically-determined URLs work just fine - IIS asks ASP.NET for a handler, ASP.NET's routing mechanism runs as normal, and if the URL corresponds to one of my dynamically defined pages, it all works fine. It's just the presence of a physical folder that stops this all from working.
I can see it's IIS doing this because I get one of the IIS-style error pages for this 404, and they have a distinctive design that's very different from the 404s produced by ASP.NET. (If I try to navigate to a URL that neither corresponds to a physical folder, nor to a dynamic resource, I get a 404 page generated by ASP.NET. So normally, IIS is definitely handing requests over to ASP.NET, but IIS is definitely getting in the way for these problematic resources.)
I tried adding this inside my <system.WebServer>, in case the problem was that IIS has decided that requests corresponding to physical folders do not meet the managedHandler precondition:
<modules runAllManagedModulesForAllRequests="true">
But that doesn't appear to help - it still doesn't get ASP.NET routing involved for URLs that correspond to physical folders. In any case, it would be suboptimal - I would prefer not to have managed handlers run for the content that I definitely want to handle as static content. I effectively want ASP.NET URL routing to be used as a backstop - I only want it to come into play if the URL definitely doesn't refer to static content.
I don't understand why ASP.NET isn't even asking ASP.NET what it thinks in this scenario. Why is it not calling into ASP.NET during the MapRequestHandler phase if there's a physical folder that happens to correspond to the URL?
When a physical file or folder with the same URL as the route is found, routes will not handle the request and the physical file will be served.
Althrough you can change this behavior by setting the RouteExistingFiles Property from the RouteCollection object to true.
Take a look at the MSDN page Scenarios when routing is not applied
I have a ASP Web API project that is being hosted over SSL. I have another ASP MVC project that makes use of the API. While debugging, I am seeing behavior where the OPTIONS requests are hanging often (but not always) and preventing the other calls from proceeding.
In the Chrome debugger, these are just shown as 'Pending'. If launch Fiddler, everything works fine. I see no errors at all, things just hang. This fails before it hits the authorization code, so I can't even set any breakpoints.
Could this be a certificate problem? Firewall?
Could you try this: adding following node in web.config.
<system.webServer>
<handlers>
<!-- your other handlers -->
<remove name="OPTIONSVerbHandler" />
</handlers>
</system.webServer>
My Asp.net application is hosted over Azure Cloud,
In that application I do have a Java Chat control, which has its server on Linux,
now I have created a HTTPHandler to redirect that chat request to the Linux server, but some how it doesn't work over the Cloud environment (though it works very well on web environment)
it shows the below error
Microsoft Visual Studio
Windows Azure Tools for Microsoft Visual Studio
There was an error attaching the debugger to the IIS worker process for URL 'http://127.255.0.0:82/' for role instance 'deployment16(315).Cloud.AnotherHttpHandler_IN_0'. Unable to start debugging on the web server. See help for common configuration errors. Running the web page outside of the debugger may provide further information.
Make sure the server is operating correctly. Verify there are no syntax errors in web.config by doing a Debug.Start Without Debugging. You may also want to refer to the ASP.NET and ATL Server debugging topic in the online documentation.
I have even put the Handler under System.webServer as well in the web.config file, see below code
<system.web>
<httpHandlers>
<add verb="*" path="http-bind/*"
type="HelloWorldHandler"/>
</httpHandlers>
</system.web>
<system.webServer>
<handlers>
<add verb="*" path="http-bind/*" name="HelloWorldHandler" type="HelloWorldHandler"/>
</handlers>
<modules runAllManagedModulesForAllRequests="true" />
</system.webServer>
I don't know what restricts it over the cloud environment.
Please take a look at this question to see how to configure httpHandlers for running in Azure.
Most importantly - remove the httpHandlers section under the system.web and leave only handlers under system.webServer. Then add also the resourceType="Unspecified" attribute to the handler declaration. This should solve your issue.
I too had same problem.
I fixed it in following way .
The Problem was my machine is 32 bit and the azure server is 64 bit.
In Order to solve the problem i only changed Enable 32 bit option in
the advance setting of application pool. And the Original code worked
out fine.
i got this help from this Link
We have a site running on MOSS 2007 which makes calls to custom web service asmx methods on the same domain from the client.
At first everything works fine, but after a bit of time has passed the service will start to fail with:
http://[domain]/_layouts/error.aspx?ErrorText=Request format is unrecognized for URL unexpectedly ending in %27%2FIsSuspectWaterLevel%27.
Interestingly enough
http://[Domain]/_vti_bin/Custom/CustomFunctionality.asmx?op=IsSuspectWaterLevel
is still available, but a call to
http://[Domain]/_vti_bin/Custom/CustomFunctionality.asmx/IsSuspectWaterLevel
will fail as described.
We've found that "touching" C:\Program Files\Common Files\microsoft shared\Web Server Extensions\12\ISAPI\ web.config will bring the webservice back to life.
The asmx file lives at
C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\12\ISAPI\ECan\MyECan_ComplianceWaterUsage.asmx
Any ideas of what might be going on here and how to resolve them?
Some extra detail:
App pool settings in case they're useful: http://i51.tinypic.com/x51qw.png
The following web.config settings are present in the root and sub directory hosting the asmx:
<system.web>
<webServices>
<protocols>
<add name="HttpSoap" />
<add name="HttpGet" />
<add name="HttpPost" />
</protocols>
</webServices>
...
</system.web>
We are calling the web service from javascript (jQuery). I've checked all the settings mentioned in this link and all match. I think calling from javascript may not be the culprit though as going directly to
[domain]/_vti_bin/Custom/CustomFunctionality.asmx/IsSuspectWaterLevel
with parameters supplied also fails with the same error - no javascript involved. Failing after a short period of time has passed, but works fine when web.config has just been "touched" again.
Thanks in advance for any help! Cheers, Gavin
I'm currently working on the same problem, and I think you barked the wrong tree.
The problem is, that in the ISAPI folder of SharePoint is a web.config with the following lines:
<webServices>
<protocols>
<remove name="HttpGet"/>
<remove name="HttpPost"/>
<remove name="HttpPostLocalhost"/>
<add name="Documentation"/>
</protocols>
</webServices>
The problem is, that the desired protocols POST and GET will be removed for the entire ISAPI folder and its subfolders. I also tried to reactivate the protocols via
<location path="[Path to my Web Service].asmx" allowOverride="false">
<webServices>
<protocols>
<add name="HttpGet"/>
<add name="HttpPost"/>
</protocols>
</webServices>
</loaction>
in different places (machine.config, web.config of root folder, web.config app.config, ...), but it didn't last.
The only thing that worked, was, to change the "remove" items in the web.config of the ISAPI folder to "add" items.
But this has the nasty side effect, that the built-in web services, like "Lists.asmx" throw errors if you try to request their documentation pages...
If you can live with that, this would be your solution. I can't, so I still try to figure out a way to make my
<add name="protocol">
items persistent.
By the way: Also adding lockItem="true" to the <add/> items didn't do the trick...
Chris
It has been awhile since I have touched Sharepoint so this is a shot in the dark. If I remember correctly modifying anything in the web.config will restart the website in IIS. So what you may be seeing is IIS restarting the website that hosts the webservice putting it back into a good state.
Do you have the following in the web.config for the web application?
<webServices>
<protocols>
<add name="HttpGet"/>
<add name="HttpPost"/>
</protocols>
</webServices>
This is a strange problem and hard to diagnose due to the number of occcurances of the 12 hives web.config protocols issue which would appear to resolve 99% of the cases of this issue.
There is another issue called URL rewriting that will cause this
problem.
Some reverse proxy devices can modify the path of a request (the
portion of the URL that comes after the hostname and port number) in
such a way that a request sent by the user to
http://www.contoso.com/sharepoint/default.aspx, for example, is
forwarded to the Web server as
http://sharepoint.perimeter.example.com/default.aspx.
This is referred to as an asymmetrical path. Windows SharePoint
Services 3.0 does not support asymmetrical paths. The path of the URL
must be symmetrical between the public URL and the internal URL. In
the preceding example, this means that the "/sharepoint/default.aspx"
portion of the URL must not be modified by the reverse proxy device.
Even more depressing is that microsoft knows about this and actively refuses to support it.
Ref: URL Rewrite + SharePoint = No Support
Also : SharePoint, url rewriter, WebServices
An inelegant workaround to this issue that works for us: We've swapped out the web service asmx end point for a web handler ashx endpoint. This doesn't suffer the same issue for some reason.
I'm guessing from this that there's some issue creeping in after a period of time which is causing urls to resolve incorrectly. I suspect that the / after the .asmx in the url is the curprit. The ashx endpoint implemented is working purely on url parameters and posted data.
Obviously this work around won't always be an option for others who might experience the same issue as we're loosing a lot of the rich web service functionality that's pre-baked in to an asmx endpoint.
Unfortunately I won't be able to test any other solutions that people might put forward from now on as we've moved away from the web service asmx approach. Sorry. Thanks for all the suggestions though - it's been very much appreciated!