There are so many questions about this issue but none has been able to address my issue specifically and I have yet to find any valid explanation of the error itself:
The underlying connection was closed: The connection was closed unexpectedly
In our situation we are making a call to a 3rd Party API via SSL. On my local PC I can connect to that API make a request and get a response back, but on an IIS Production server I get this error. The API is using OAuth to authenticate.
What exactly does it mean. Is the request leaving our server and rejected by the remote server, or is it not even leaving our server and our system is preventing it from making the request.
Some more information incase anyone may know what the issue is:
No known changes to any networking, servers, routing, security (apparently) No code changes recently
According to our own internal logging, the issue started off as an ocassional 403 Error-Forbidden then we saw a number of Cannot Connect to Remote Server. Eventually it failed with The underlying connection was closed: The connection was closed unexpectedly.
Can someone please explain what the actual error means? If anyone has experienced this in a similar situation and can shed some light, that would be greatly appreciated.
The underlying connection was closed: The connection was closed unexpectedly
This just says, that (probably the remote end) closed the TCP connection which underlies the SSL connection. Usually an SSL alert should be sent back on SSL related errors, but some stacks instead close the connection. It might also be that the peer does not expect SSL at all and thus closes the connection because of invalid data.
On my local PC I can connect to that API make a request and get a response back, but on an IIS Production server I get this error.
It is hard to say what the problem might be, but if this is not only the same API but also the same server then the problem must be related to differences in the client. This can be support in ciphers, TLS versions, client certificates etc which can be different between machines. If this is not even the same server you should make sure that the problem is not server related by contacting the non-working server with the working client. A good idea is also to make a TCP dump (wireshark) and compare the handshakes.
More detailed problem analysis can only be done when you provide more details about the problem, see http://noxxi.de/howto/ssl-debugging.html#hdr2.2 on what might be useful information.