Powershell remoting - cannot execute an exe as another user - powershell

I've a commandline program (c#) that encrypts config files based on machine key.
A powershell script copies the build to a Target Server, modifies configs accordingly and installs windows services.
All the windows services run as local system account (standard user, non-admin) - let's call this account "locuser".
The Target Server is a Win 2012 R2 Server. All of the above is achieved by PS remoting from the Build Server to this Target server.
Now, I need to run the encrypt commandline program as "locuser", so that the program can use the account specific key to do the encryption.
I know that this can be easily achieved by calling Start-Process cmdlet with -Credentials parameter. Well, here's the catch, the above works fine, if I remote in (RDP) to the Target Server and then run the Start-Process .... -Credential $cred from a Powershell Console.
However, I need this to be working while I remote-in (using my scripts) to the TargetServer whilst deploying. When I remote-in to the TargetServer I use credentials that has Admin privileges.
I've tried the following
I've granted "locuser" both "Full Control" and "Invoke (Execute)" permissions by using the Set-PSSessionConfiguration -Name Microsoft.PowerShell -ShowSecurityDescriptorUI command. I've run this command for both Microsoft.Powershell and Microsoft.Powershell32 - Still get Access Denied
I've edited the "Local Security Policy"->"Local Policies"->"User Rights Assignment"->Impersonate a client after authentication - and added both the Admin account (that I login with) and the "locuser" account - Still get Access Denied
I've also granted locuser admin rights - Still get Access Denied
I'm pretty sure, there is some configuration on the PS Remoting Side of things that I'm missing out but can't figure out what - because all Powershell throws me is a Access Denied error (see screenshot) with little to no useful information to troubleshoot further.
Also, checked Event logs for any traces but to no avail.

You've fallen prey to the dreaded Double Hop. Basically you're authenticating from computer A to computer B, then trying to authenticate again from computer B to computer C (which also happens to be B in this case).
If at all possible, you would be better off ending the session and starting a new one with the locuser credentials, then just calling Start-Process. Another, more messy approach is to use schtasks.
I can tell you how to do it in the same session but it's a bit messy and very complicated, and should only be a last resort:
On the originating server (Build Server):
Run the command Enable-WSManCredSSP -Role Client -Delegate [name] where [name] is an IP or DNS address / range including any target servers (eg "192.168.1.*")
Open GPEdit.msc, navigate to Computer Configuration\Administrative Templates\System\Credentials Delegation and check that the rules Allow delegating fresh credentials and Allow delegating fresh credentials with NTLM... are enabled and include [name]
On the Target Server:
Run the command Enable-WSManCredSSP -Role Server
Running the command:
Invoke-Command [targetserver] [-Credential $cred] -Scriptblock {
## do stuff
Invoke-Command . -Credential $locusercred -Authentication Credssp -ScriptBlock {
Start-Process -FilePath $sc #etc
}
}
Some things to be aware of:
Firstly I used this setup to create a local session, then remote from there (so A-A-B instead of A-B-B) so the Group Policy stuff might be in the wrong place but pretty sure it's right.
Secondly I found that credentials are a pain to get working in sessions (in this case $locusercred). I did get it going natively but weirdly it suddenly couldn't decrypt the securestring. I ended up saving a securestring with a defined key to the registry so it can always be decrypted from any account, you may need to come up with your own solution there.
All this stuff is explained in the free eBook "The Secrets of PowerShell Remoting", if you go for the double-hop approach I recommend giving it a read.

Related

Double-Hop Errors when running Skype for Business Cmdlets

I am attempting to automate the Skype for Business Server installation process in Powershell, I have a script that remotes into specified machines and begins preparing them as Front-End servers. The problem lies when certain SfB cmdlets (SfB commands are all of the form "verb-Cs...", ex. Get-CsUser or Get-CsPool) are run in remote sessions, they throw the double-hop error:
Exception: Active Directory error "-2147016672" occurred while searching for domain controllers in domain...
This is after running Enable-CsComputer, which enables the computer's role-based off its definition in the topology (topology was published successfully). The user object is in all required groups (RTCUniversalServerAdmins, Schema Admins, CsAdministrators & Local Admin rights on all SfB Servers). Oddly enough, the command 'Import-CsConfiguration -localstore" does not throw errors, and it's in the same remote session. There may be other local or domain groups that I need to be in, but I cannot pinpoint exactly which and have not seen them documented in the Skype build guides. Skype commands that have parameters to specify targets or just pull data, such as Get-CsPool or Get-CsAdForest, do not have errors because they are run in the local scope. The Enable-CsComputer has no parameter for the computer name, it has to be executed from that machine itself.
Enabling CredSSP delegation on each server is not an option, and I'm not understanding why there is a "second hop" in this command! If the second hop was a resource on a file server or database, that would make sense, and be easy to solve, but in this case, I can't track it. Can anyone tell me what I may be missing?
Here's a code sample to try and illustrate. From the jumbox I get the pool data to create an array, and a session is opened to each machine:
$ServerArray =get-cspool -identity $poolName
$i=0
$SessionArray = #{}
foreach($server in $ServerArray.Computers){$SessionArray[$i] = new-PsSession -ComputerName $server}
foreach($session in $SessionArray.values){
invoke-Command -session $session -scriptBlock {
#remote commands:
import-csConfiguration -<config file path> -localstore; #no errors
enable-CsReplica; #no errors
enable-cscomputer; #double hop error here
}}
If I log into that machine and run the same command, it executes fine but the intention of the project is to automate it on an arbitrary number of machines.
It looks like it's just trying to authenticate to a domain controller, which is reasonable. You'll have to approach this like any other double-hop issue.
Microsoft has an article dedicated to the double hop issue, and has a few solutions other than CredSSP that you can look at: Making the second hop in PowerShell Remoting

Powershell invoke-command multihopping

I have a question regarding multihopping in a windows environment.
Let's say I have a schedule running on Server A (Central Scheduler) which executes a command on Server B. This script contains a call to save files on a remote filer (UNC path, Server C). Hop 1 (from A to B) works well, hop 2 (from B to C) fails.
I already tested to save the files locally on server B, that works flawlessly.
I think there's a problem with the second hop. I remember reading something like this on a forum a while ago, but can't remember a solution.
In detail, the command looks like this:
$session = New-PSSession -computer ComputerName
$templatepath = "\\filerpath\"
Invoke-Command -Session $session -Scriptblock { powershell ovpmutil cfg pol dnl $Using:templatepath /p \BSH }
To clarify: Powershell gives me an "Access denied" when performing the second hop. I already enabled Credential delegation as described here:
Enabling Multihop Remoting
Any help is appreciated. Thanks in advance
The solution is a real pain in the backside if you ask me but here it is...
On the originating server (A):
Set-Item WSMAN:\localhost\client\auth\credssp -value $true
On the intermediate server (B):
Set-Item WSMAN:\localhost\client\auth\credssp -value $true
Open Group Policy editor on server A, navigate to:
Computer Configuration > Administrative Templates > System > Credentials Delegation
Enable these options:
Allow delegating fresh credentials
Allow delegating fresh credentials with NTLM-only server authentication
Both policies need to have server B added to the allowed list, wildcards are allowed. Note that if you use RDP from server A you'll also need to add TERMSRV/*
When running Invoke-Command from server A, include the -Authentication CredSSP param.
Note that if saving SecureStrings somewhere for the credential to connect to server C, you'll want to either use a fixed encryption (specify byte array) or plain text and convert it.

How to remotely register static ETW manifests as part of a website deployment?

I'm doing a pilot effort to use the new EventSource (Microsoft.Diagnostics.Tracing.EventSource from nuget) and its new support for ETW channels in order to write to the windows event log. The code is in place, and it writes properly to my workstations event log. I'm thrilled.
Now comes the difficult part. The application that's taking advantage of this capability is a web service, and we deploy it with webdeploy as part of a build-deploy-test system. Because usage of ETW channels requires static registration of provider manifests via wevtutil.exe. The EventSource documentation states that this is best done as part of an installer, but this seems a bit out of webdeploy's capabilities.
Our aim is that we would be able to automatically uninstall the manifest resident on the target server immediately before executing the webdeploy package, and then to import the new manifest after the webdeploy sync has completed. We're not set on this, but it seems like the most sensible way.
For that reason, it seems like maybe this is something that powershell remoting might be able to solve, but it's not an area I know much about.
Has anyone done something like this? Is there a better or simpler way?
There are only a few requirements here. A) the remote machine must have PowerShell remoting enable which also means it must have PowerShell 2.0 or higher B) the script running on the local machine must be able to run as administrator and the credentials used must have admin privileges on the remote machine. If you can meet those requirements then this should be cake.
On the remote machine you need to execute two commands to enable remoting:
Set-ExecutionPolicy RemoteSigned
Enable-PSRemoting -Force
Then on the local machine from an elevated prompt you should be able to execute something like this from a script:
# these two paths assume these files have been copied to the remote computer and to a directory
# in which the service account has privileges to read i.e. not under a userprofile dir.
$etwDllPath = c:\somepath\myassembly.mysourcename.etwManifest.dll
$etwManPath = c:\somepath\myassembly.mysourcename.etwManifest.man
$s = New-PSSession -ComputerName <remoteComputerName>
Invoke-Command -Session $s {param($man) wevtutil.exe um $man} -arg $etwManPath
Invoke-Command -Session $s {param($man,$dll) wevutil.exe im $man /rf:$dll /mf:$dll} -arg $etwManPath, $etwDllPath
Remove-PSSession $s
If you can avoid a remote path with spaces, try to. It will make this easier. :-)

Powershell: Discover default network credentials when launched from runas /netonly

I am looking for a way to capture the network credentials of the current session into a variable that I can pass later...
The point is to execute commands on a foreign domain that I have account access/privileges to, but there is not a trust between the source and target domains.
First, we run inside a powershell that was spawned using the runas command (runas /netonly /user:domian\account powershell
From here I can do pretty much everything I want to except create an event in the task scheduler without hardcoding the username/password into the command line
invoke-command -computer $destination -scriptblock {schtasks -ru domain\account -rp password}
What I am looking to do is something like
$username = Get Current Session Network Username ($(whoami) brings up the actual local longon account,not the runas account that spawned the powershell window)
$password = Get the Password that was entered when the RunAs command was executed
Once a security token has been created from credentials entered and validated against active directory, the password is no longer kept around. It is not available for retrieval and reuse elsewhere. Only the token remains. This is by design.
UPDATE:
I dug a little further to bolster my case, and it's not quite as above but the end result is the same. The password used with runas.exe does not appear to be available. The network credentials are not validated against AD, which makes sense in retrospect since you often use /netonly for working with a remote, untrusted domain: By definition, you cannot validate the remote credentials from the local system. From MSDN:
Here's information for the flag LOGON_NETCREDENTIALS_ONLY, used with CreateProcessWithLogonW.
Log on, but use the specified credentials on the network only. The new
process uses the same token as the caller, but the system creates a
new logon session within LSA, and the process uses the specified
credentials as the default credentials.
This value can be used to create a process that uses a different set
of credentials locally than it does remotely. This is useful in
inter-domain scenarios where there is no trust relationship.
The system does not validate the specified credentials. Therefore, the
process can start, but it may not have access to network resources.
Ok, so since it can't validate the credentials and get a token, then it must store the password somewhere in memory since it must pass them over the wire later for SSPI etc. So, can we get at them from the process launched from runas.exe ? Let's see:
PS> runas /netonly /user:foo\bar powershell.exe
Enter the password for foo\bar: ******
I literally used foo\bar for domain and user above. It is not validated, as mentioned already. I entered 12345 for a password. The above line will launch a new instance of powershell. So, from that new instance, let's look at the default network credentials:
PS> [System.Net.CredentialCache]::DefaultNetworkCredentials
UserName Domain
-------- ------
Oh well, out of luck: Nothing there. My guess is the credentials are guarded in some encrypted part of memory in the kernel, probably the LSA (local security authority) out of reach from prying processes.

Run Command as administrator in PowerShell script. UAC

OK here is my issue:
I am trying to run a script remotely on a server.
I am an administrator on both boxes, firewall exceptions are in place, remote admin is enabled, and everything else looks good that i can see.
invoke-command -ComputerName $ComputerName -ScriptBlock `
{
cd C:\Windows\System32\inetsrv\;
./appcmd.exe ADD vdir /app.name:<SiteName>/ /path:/<VDir Name> /physicalPath:<Path to files>
}
I keep getting the following error in return
ERROR ( hresult:80070005, message:Failed to commit configuration changes. Access is denied.
The server it is trying to run on is a server 2k8 R2 box and I am thinking the issue is a UAC problem. Is there anyway to get this to run as administrator without having to click yes on a UAC box?
This piece of code will eventually become a script that will have to be completely automated.
Any help would be greatly appreciated.
OK. After some research and testing I figured out the issue. After disabling UAC and the firewall and the script still not working I dug a little deeper and discovered that the main issue was the way invoke-command runs the commands. it uses the credentials of the person running the script to authenticate to the server then tries to use another account to run the permissions or lowers the privileges of the user so that certain commands cannot be run.
I added the -Credentials switch to the invoke command and everything is working great now. Corrected code sample below:
$user = New-Object Management.Automation.PSCredential("$UserName", $securePassword)
invoke-command -ComputerName $ComputerName -Credential $user -ScriptBlock `
{
cd C:\Windows\System32\inetsrv\;
./appcmd.exe ADD vdir /app.name:<SiteName>/ /path:/<VDir Name> /physicalPath:<Path to files>
}
This seems to indicate that you need to ensure you are a local admin on the remote machine (although admittedly this is for WMI specifically). According to this you can change a registry key to stop UAC applying to remote logons for administrators (search for LocalAccountTokenFilterPolicy). That shouldn't disable UAC just not filter the token if you use powershell/WMI remotely with an administrator account.
Set the option "EnableLUA" (DWORD value) found in HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\System to 0 and reboot.
This will disable UAC without a problem, I would do it for all your users, whether with or without permission is up to you.
This trick works in Windows Vista and Windows 7 too.
Is there anyway to get this to run as administrator without having to click yes on a UAC box?
If this were possible it would entirely defeat the point of UAC.
Thus, it would appear your only real solution is to disable UAC on the box.