I'm working on the application that uses REST web service with SSL. The web service is already in place, identifying itself with the SSL certificate. Even though the certificate seems to be valid, it is still being rejected by iOS. Below are the details:
Web service presents the certificate chain consisting of 3 certificates, the root certificate that is present on this list, the intermediate certificate and finally the web service certificate
The method - (BOOL)connection:(NSURLConnection *)connection canAuthenticateAgainstProtectionSpace:(NSURLProtectionSpace *)space is called each time I send requests to the web service. The comment in RestKit code says "server is using an SSL certificate that the OS can't validate", so I guess I should never reach that method, right?
I tried to access the web service from Safari on the test device, and the certificate is automatically accepted. The same goes for the desktop browsers.
Sslshopper says the certificate is fine.
I can connect to the web service by adding the certificate to the application bundle and handling it the same way I'd do with self-signed certificates, but I want to solve this the proper way. Any help greatly appreciated, especially if someone can point out why Safari on iOS accepts the cert, and my application does not.
Edit: it turns out that I've been using old RestKit version. I created the proof of concept that allowed me to dispatch some requests to my server, and with the RestKit master from today (3rd May 2012) the certificate is accepted. However, I get two calls to connection:canAuthenticateAgainstProtectionSpace:, in both of them I return NO, but it manages to connect properly anyway. So, the new question is, what exactly has changed in RestKit that I can connect properly now. I need to know this because I will have to patch the old version that I'm going to use for release. I will update RestKit afterwards.
Edit: I need to give that back to the RestKit guys, it was not their fault. The web service that I'm trying to connect to is using some kind of gateway which is the source of the problem. There were some fixes from the gateway crew earlier in the day, that's why it worked correctly by the time I managed to perform the tests with the latest RestKit version.
Related
I was studying Websockets (wss) and I am a little confused about their use of SSL. I hope some of you will be able to clarify some things that cause the most confusion to me.
Use scenario
Let's say I have an HTTPS secured website with a certificate signed by CA. I would like to communicate with a custom app in real-time (using the WebSockets) running on an external server that is not using the same domain (the site is for example www.mysite.com and the server is www.notmysite.com).
The confusion part
When the WebSocket client (wss) is connecting to the remote server, what SSL certificate does it use? Is it of the origin website, or of the server it is trying to connect to?
Given the scenario, can a connection be established with a single-domain SSL certificate - only for the website (www.mysite.com) where we know the private key and therefore we should be able to decrypt the encrypted data?
If question no. 2 is false, would multi-domain certificate containing both urls fix the problem?
The background story
I was trying to make a WebSocket server in .NET C#. I managed to get it working with HTTP however when I tried the code on HTTPS secured website I was presented with an encrypted text I could not really work with. That is why I bought an SSL certificate and put it on the website thinking that if I have the private key I can decrypt the data and communicate with the client (basically question no.2). Though I was only presented with different errors regarding cryptography. I was trying to figure this out for over a week and my progress and frustration can be seen in my other questions such as:
C# RSA decrypt parameter is incorrect and
AuthenticateAsServer() error
Then I figured out I may need to look at it from a different point and that's when I came up with this theory. I hope some of you will be able to end my suffering with an explanation of why what I did was wrong. I appreciate any help. Thank you very much.
I am not 100% sure if the multi-domain SSL certificate solved this issue since I combined it with using a premade library (WebsocketsSharp) which along with the new multi-domain certificate works perfectly. In my opinion, the multi-domain certificate is needed if you want to communicate between two different servers (either website and server or two servers).
I hope I will save some future headaches.
I have the following setup:
A self signed certificate for development purposes
An OWIN hosted Web API, deployed on a local Azure Service Fabric Cluster as a ASF service. The Web API uses HTTPS facilitated with the dev certificate in question.
A simple .net client application that calls the Web API. In that application the ServicePointManager.ServerCertificateValidationCallback method is set so it always returns true (since the CA is not trusted)
Both the client application and the Web API are on the same local development machine. The certificate is installed in that same machine on the "machine store"
Now I am able to make calls to the web api using Fiddler by providing the required client certificate. However when I try to do the same via .net code (be it RestSharp or WebRequest) the client certificate is not present in the RequestContext object on the server side. This result in an Unauthorized response. I do not think the problem is with the client code, the certificate is loaded correctly and assigned to the http client or request. Fiddler shows encrypted tunneling to the server. However the client certificate does not seem to be present at the server side.
I am at a complete loss regarding what am I missing that could cause this behavior. Any help will be deeply appreciated.
Install the certificate into the 'Local Machine/Trusted People'.
The server needs to already trust the certificate before it asks. The CertificateRequest message that it sends lists the acceptable CAs that can sign the client certificate. If the client certificate's signer isn't in that message, it can't send it.
How you accomplish that in your environment is left as an exercise for the reader. In general now that SSL certificates are available free there is little reason to indulge in the time costs and administrative inconveniences of self-signed certificates. IMHO there wasn't even if you were paying for a CA signature.
I had a problem when a service called another service over HTTPS and it couldn't setup a secure connection. My problem was that since the service is running as NETWORK SERVICE, it couldn't find the certificate, because it was looking in the localmachine/my certificate store.
When I was running from my web browser it was working fine because then, my browser found the certificate in the currectuser/my certificate store.
Add the certificate to the machine/my store and see if it helps.
We receive a SHA-2 certificate from a vendor to use on our site when making requests to them. This is unlike anything i have done before. Typically in IIS we would create the request, receive the certs and install.
In this scenario we were just handed one for requests to them. Anyone know how or what is supposed to be done with them?
Did use MMC and installed the cert that way, but not sure how or if that is to link to website.
It's only for server side certificates that you use the IIS management console. It sounds like this is a client certificate used to identify your app to someone else.
It is up to the language you are using to call the webservice to determine where the certificate should be placed. If you are using C# then there are lots of questions that talk about using client certificates. I would imagine that most languages would let you reference the certificate file directly.
I have a WCF service running on IIS6 which is configured with a WSHTTPBinding using the "Transport with Message Credentials" security setting (with the transport client credentials set to 'None' and the Message set to 'Certificate'. This is being consumed by a .NET client that was created through the auto generated code by consuming a service in VS 2010.
In my development environment, I created a certificate to act as both the service cert and the IIS SSL certificate, and since I was lazy, I exported the certificate for use with my client application. Everything works great in that configuration, but that is not the target environment. The idea is to have users utilize their smart cards to access this service. The server will not be on the same domain and LDAP is not allowed, for now.
The problem is that now I got this in a test lab and set up an environment where the SSL certificate was issued by the same CA as the user's certs, and I have created a custom x509 Cert validator and service credentials, but after the initial call to my validator, something is happening in between the hand off and kicks out an error on the service trace log saying, "Error verifying message security" with an inner exception of "Cannot verify the signature". Since is is coming through my custom validator and I got the message logging up, everything looks fine on the SOAP side and I am getting the certificate through on the other side. I have made sure that every cert that I am trying to use is in the "Trusted Users" store and even tried exporting the certs and putting them in the local computer "Personal" store, with no luck.
I have a feeling that I do not fully understand the use of the service certificate and the client certificate in the web.config file. Can someone please help me out here?
Ok so I have managed to answer my own question with some help from my coworker. After we sat down together and started to run a bunch of questions by me and doing some random google searches, we decided to change something in the configuration that surprised me.
So to finish my question up top, I had the message credentials set to "Certificate" with the encryption algorithm suite set to "TripleDesSha256Rsa15". Turns out that the certificates on the smart card, or the card itself, would only support Sha1 (which was not an option with everything else). When we changed the setting to "Basic256", everything started working. I eventually was able to set it to "TripleDesRsa15" and it continued to work.
Basically, WCF sends the certificate up without a private key, because it cannot access it with that encryption algorithm. WCF DOES NOT throw an exception with this when trying to access the certificate, it will only throw it on the server side with the exception "Could not Verify Signature". Kind of wish it would say "There was no signature" because that would have been a bit easier to track down. I don't know if anyone will ever find this useful, but if you do and can find a reference to the "Proper" way of using Smart Cards with WCF, please post it with a ref.
Overall, check your encryption algorithm that you are trying to use, and make sure it works with the smart cards/certificates that you plan on using.
I'm currently working on an automated deployment process for a hosted service for Windows Azure. The creation of the .cspkg and .cscfg files works perfectly using a call to msbuild. Now I'm writing a small .NET console app that should deploy these files to Azure using the Management REST API.
There is no problem concerning the API itself. I can send a request to the API using one of my management certificates. I upload the .cspkg file to Azure BLOB Storage and then try to call Upgrade Deployment. But every time I try, I get a "400 Bad Request" response stating that the certificate with thumbprint xy was not found. This certificate is the SSL certificate (not a management certificate) I'm using for HTTPS for my custom domain (DNS CNAME).
And now, the whole thing gets interesting:
When I deploy the files using the "Publish" command in my Visual Studio, there is no problem. (I compared the .cscfg/.cspkg files from VS and from my msbuild output: apart from a few GUIDs, they're identical). And furthermore, using the Silverlight Management thingy in my browser, I can even upload my generated files that could not be uploaded using the API.
When I retrieve a list of all certificates using the List Certificates call, the certificate which is said to be missing is apparently there. I can also retrieve its data using the Get Certificate call.
So why does Azure keep telling me that the certificate was not found when using the Upgrade Deployment call? Did anyone experience something similar? Has anyone the hint for me? Thanks in advance.
P.S.: This is what Azure says when I use the API:
<Error xmlns="http://schemas.microsoft.com/windowsazure" xmlns:i="http://www.w3.org/2001/XMLSchema-instance">
<Code>BadRequest</Code>
<Message>The certitficate with thumbprint 7b232c4a2d6e3deadbeef120d5dbc1fe8049fbea was not found.</Message>
</Error>
P.P.S.: Yes, the word in the response is certitficate, not certificate.
OK, after using the List Subscription Operations API call to find out what Visual Studio calls to deploy apps, I found the solution.
Turns out that the URL I used for the API request was wrong, but: with all due respect, I blame Microsoft for lousily documenting its Azure Management API.
In their documentation, they write the URL to use is:
https://management.core.windows.net/<subscription-id>/services/hostedservices/<service-name>/deploymentslots/<deployment-slot>/?comp=upgrade
And the description is the following:
To generate the request URI, replace <subscription-id> with your subscription ID, <service-name> with the name of your service, <deployment-slot> with staging or production, and <deployment-name> with the unique name of your deployment.
What they forgot to mention is, that you have to use the DNS Name of your service, and not the Name! They could at least return an appropriate error message telling you that the service name is invalid, non-existent or doesn't belong to your subscription ID, instead of complaining about some certificate issue.
Thank you Microsoft, that cost me more than two days.
The error indicates that you have not uploaded that certificate into the hosted service's secret store. Visual Studio might be doing that automagically for you, but if you want to replicate it programmatically, then use the Add Certificate API call and upload the PFX into the deployment.
You can see '400 BadRequest - The certificate with thumbprint XYZ was not found.' appear in the CreateDeployment or UpgradeDeployment scenario for the following reason (which I just debugged):
You use the same certificate for subscription management as you do for e.g. SSL or Remote Desktop password encryption in your hosted service. You therefore will use the certificate with thumbprint XYZ to authenticate your service management REST call that creates the deployment.
When specifying your deployment parameters you pass in your CSCFG which references that same cert by its thumbprint, because it needs to configure Remote Desktop/SSL etc.
That cert is not yet added to your hosted service certs.
In this case the 400 Bad Request error really is telling you that you have a bad request, because the certificate in your CSCFG is not yet attached to your hosted service. The confusion arises (for me) because, since its a multi purpose cert, you misinterpret the error message as referring to the authentication of the request, even though you are not getting 401.