Using SSL causes mssing IP adresses in eventid 4625 and to get them back .. disable NTLM ? Nope. Not really an option
Today I took me a lab day to actually sit down and spend time with the NTLM settings and the RDWEB and try it out on various platforms and do some more or less scientific testing.
In short, I’äm not impressed by how Micropsoft has actually implemented parts of ther stuff..
I used Windows 10, RD Client for IOS and RD Client for Android. The server infrastructure was a Windows Server 2008 R2 with valid SSL certficates for all services.
The underlying problem is basiaclly that if you use an SSL certificate for your RDP connections , failed logins aren’t correctly dispalyd , i.e. your missing IP adresses in eventid 4625. (When not using an SSL certficate , it is recorded but then your users and customers get a lot of warnings when connecting to your servers and some things just donät work very well sucha as the Webfeed for RD Web)
Syspeace is a Host Intrusion Prevention Software that uses this inormation about the source IP address to block brute force attacks against Windows Servers.
One way around this is to disable incoming NTLM traffic and sure enought , all IP addresses are recorded.
The downside is .. only “full” RDP connections will work meaning that for instance connections to a server desktop works fine but if you’re really into RemoteAPP (and that’s the way I want to go and a lot of tekkies with me) you’ll be running into problems.
And, by th way.. frankly, full desktop session don’t work either from IOS (at least remote Desktop Client 8.1.13 and my iPad, they do from Android though, same server, same username and so on)
Not even Windows really working correctly when disabling NTLM ?
I also did some testing for fun by creating a .wcx file and oddly enough. In order to get that to actually work with Windows 10 (and I’m guessing it’s the same for Windows 7 and so on ) , It just refuses to connect to the RemoteApp service if incoming NTLM is disabled.
I can howerver start a normal Desktop Session against the server so, what I would claim is that the fault is actually within RD Web and the way it handles authentication, requiring some parts to be using NTLM.
The usual RD Web login interface works so far that I can login and see the resources but I can’t start any applications from it. No errors, nothing.
If enabling NTLM, I can start the applications just fine. Once again. NTLM has to be enabled in order for full functionality 🙁
So, basically, if I change the policy settings for the RD Server not to allow incoming NTLM traffic in order to be able to actually handle a bruteforce attack and also keep track of failed logins with informaion that’s actually useful for me as a sysadmin and CSO
These are by the way the settings I’m referring to
Computer Configuration\Windows Settings\Security Settings\Security Options
– Network security: LAN Manager authentication level — Send NTLMv2 response only. Refuse LM & NTLM
– Network security: Restrict NTLM: Audit Incoming NTLM Traffic — Enable auditing for all accounts
– Network security: Restrict NTLM: Incoming NTLM traffic — Deny all accounts
Regardless of how I try, I can’t get it to work to actually add remoteapp resources (or Remote Resource Feed) neither Windows 10, nor IOS, nor Android.
So, what are the implications of this ? Does it matter ? Do we need the source IP address in 4625?
First of all, the way this is handled within Windows Server is an absolut nightmare and frankly, just usesless and I can’t see any reason for Microsoft developers to leave the IP address out when using SSL certificates or at least have another entry in the eventlog for it containg useful information.
It’s not possible to handle brute force attacks natievly within Windows Server as I’ve written about many times earlier.
The biggest problem is of course that if someone tries to bruteforce your server, then how will you stop the attack ? How do you gather evidence ?
If your’e running a larger server environment and hosting customers and so on , you’ll have no way of knowing what attempts are legitimate customers and user and which ones aren’t really.
You can hardly shut down your services can you ?
At the moment , I don’t have a good solution to this problem. Syspeace catches lots and lots of bruteforce attacks for me but these ones it can’t since it doesn’t have any IP address to block.
I’m just hoping for Microsft to actually solve this on the server side since that would be the easiest fix for them I’d say.
Of course they also neeed to get the RDP clients working for all platforms but basically it should be working with NTLM2 at least and also to log the failed logon request correctly if using an SSL certficate. Anthing else is just pure madness and stupidity to be honest and someone should get fired for not thinking ahead.
