I have a command line program that listen to a tcp port until user type Q to exist. It works fine in local powershell window. But when I try to run it on another machine using powershell remote session, it just starts and quit. Is there a way to keep it running?
The remote script runs in a PowerShell that never becomes visible so AFAICT it doesn't even got a console handle by which to handle reading keyboard input.
You can take a look at the SysInternals utility - psexec. From my testing, that utility works for what you are trying to do.
Ensure you have Powershell 3 or higher since it adds support for detached sessions/background jobs.
Use a Remote Disconnected Session, described on Technet
Related
I am trying to create a small script I can run to connect to ExhangeOnline but when I run it connects but immediately disconnects before I can use the tool to make changes to Exchange. Any way I can keep PowerShell connected until I am done with it?
If I run the following command in my session...
(Get-Process -Id $pid).CloseMainWindow()
I am able to gracefully shut down a process (no modal windows or other popups arise).
If, however, the pid is in another user's session on the same machine (running RDS), the process does not close, and CloseMainWindow() returns FALSE (it returns TRUE if it's running in my own session). It also works if I run the powershell from the other user's session.
I specifically need a way to gracefully shut down the program as the program has a few important cleanup actions required to keep its database in order. So stop-process or process.kill() will not work.
After lengthy research, it does not seem possible to do this. There is, however, a solution which met at least some of my requirements.
You can create a Windows Scheduled Task which is triggered on session disconnect. This allows you to run a cleanup job as the user, rather than as the administrator, which allows programs to exit gracefully.
It has two major drawbacks....
It is called even if the user just has a minor network interruption (so you have to build a wait() function in the script to sleep for a bit and then check if it is still disconnected - not a clean solution.
It isn't called during a log-off event. For that you need to use a logoff script triggered by GPO.
Hope this helps someone in the future.
I do know about the double hop issue. My scenario is: I have a script I want to run remotely that calls another script located on a network share that calls a third script located on a second network share in a different domain.
Currently what I am doing is using Credssp (I've read there can be security issues but this environment is not public facing) to pass credentials for the 1st network share that has script2. I do not have access to the computer with the second domain so I cannot setup credssp on it. In order to work around that, inside of the script2 I am using "net use" command on the third script in order for the script to be able to find the path. I am then using "Copy-Item" to copy the third script on to the machine running script2 (the remote machine).
Up to this step, everything is working when I run script1. I can see script3 is copied over onto the remote machine. When script3 is called, it should make a web request that sends text to stdout (which I pipe to Out-File in script2). However, whenever I try to run the copy of script3 (located on the remote machine) from script2 (running on the remote machine) it does not seem to do anything. If I run script2 locally on the remote machine then it works fine (file is generated from script3's output).
Any idea's on why this won't work? I've tried running script 3 using several variations of invoke-expression, invoke-command, start-process, and even trying to run with cmd. I'm also having trouble getting output on what exactly is causing the issue (stdout and stderr are many times empty when using the different commands). Am I missing some command or tool that may make this easier to troubleshoot? It almost seems like script3 is still running into a double hop issue despite it only making a web request? And if it was running into that, I thought it would have had an error returned.
There my be a better design for doing what I'm trying to do. I'm fairly new to PowerShell and may be over complicating this.
Edit: Rewrote my scripts in python and got it working.
I've got a batch file dmx2vlc which will play a random video file through VLC-Player when called.
It works well locally but I need this to happen on another machine on the network (will be adhoc) and the result (VLC-Player playing the video) must be visible on the remote screen.
I've tried SSH, Powershell and PsExec, but both seem to run the batch file and the player in the session of the command line, even when applying a patch to allow multiple logins.
So IF I get to run the batch file it is never visible on screen.
Using Teamviewer and the like is no option as I need to be able to call all this programmatically from my dmx program.
I'm not bound to being able to call the batch directly, it would be sufficient for me if I could somehow trigger it to run.
Sadly latency is a problem here as we are talking about a lighting (thus dmx) environment.
Any hints would be greatly appreciated!
You can use PSexec if the remote system is XP with the interactive parameter if you state the session to interact with, 0 would probably be the console (person physically in front of the machine).
This has issues with Windows Vista and newer as it pops up a prompt to ask the user to change their display mode first.
From memory, you could create a scheduled task on the remote system pretty easily though and as long as it's interactive the user should see it.
Good luck.
Try using web interface. It is rather easy: VLC is running http server, and accessing particular URL from remote machine will give full control over VLC. Documentation can be found here
I have the following batch script on a Windows 2008 R2 server:
#echo off
djoin.exe /provision /domain my.domain.com /machine test /savefile savefile.txt
echo %ERRORLEVEL%
If I run the script on the server itself, either through command prompt or PowerShell, it works perfectly fine and returns "0".
The problem is that I need to execute it from a remote computer, so I do the following (an example just for testing):
Invoke-Command -ComputerName remotehost -ScriptBlock {.\script.cmd}
The output is "-1073740940", which is probably error code C0000374, which could have something to do with heap corruption.
This seems to be a problem with the djoin command itself. I can comment out djoin and run other binaries, like ping, with no issues using the same Invoke-Command.
Keeping in mind that the script works perfectly fine when executed from PowerShell on the target computer, what issues could the act of remoting be introducing?
In both cases, the script is executed with the same privileges using my account, which is a member of Domain Admins. I doubt that it's a permissions issue and have no idea where else to look.
[edit]
Gave up on the whole thing. This is either a bug in djoin or some obscure problem in the interaction between djoin and PS remoting.
I managed to run djoin directly on the client, using 'runas /netonly ...' to provide domain credentials. It's a very messy solution (and I have yet to figure out how to get the exit status of a process started by runas), but gets the job done.
This is almost certainly a classic "double-hop" authentication issue. Remember that when you use PowerShell Remoting you're using up one of those hops. Anything you execute on that remote machine that accesses a third remote machine is unlikely to work if it requires authentication.
To get around that, you can use an authentication method which allows you to Delegate Credentials such as CredSSP. It's a bit more involved than simply changing your authentication type as you have to make changes on the client side and the server side of the transaction. Refer to this blog post on MSDN, PowerShell Remoting and the “Double-Hop” Problem and this "Hey, Scripting Guy!" post, Enable PowerShell "Second-Hop" Functionality with CredSSP.