I have written a command which run a powershell script as an elevated user of the 'Domain Admins' group.
I want this script to aim UAC elevated privileges it is requiring, since this script should be able to perform administrative tasks. I want it to ask then, possibly popping Windows to the currently logged user, which has launched my command.
How to mark a powershell script as requiring elevated priviliedges if UAC is activated.
I'm using standard output and error as log, it would be nice to keep them.
Cordially,
Related
I startet to look around but only found usage with a prompt. But as this is for the users I was wondering if I can make it run without a prompt still with elevated privileges behind.
Start-Process powercfg.cpl -Verb runas
this is a simple oneline but this prompts for creds can I somehow tell it just to run elevated without a prompt
Update1: We are thinking of creating a AD User which has only rights for the needed task than can be shared for this prompt.
Update2: WE figured out that pre installed Toshiba software was causing it to jump back after restart and after every edit. We deinstalled the software and since there was no need to run it as admin anymore.
I'm attempting to run a Powershell script as the SYSTEM account as a scheduled task, and the task runs OK, but the script runs incorrectly because it does not load some Citrix Powershell snap-ins that are needed when run as SYSTEM, but runs OK as my user account.
The script does a check to load the Citrix snap-ins, but those do not appear to work when run as SYSTEM.
if ((Get-PSSnapin "Citrix.Common.Commands" -EA silentlycontinue) -eq $null) {
try { Add-PSSnapin Citrix.* -ErrorAction Stop }
catch { write-error "Error Citrix.* Powershell snapin"; Return }
Is there anything special I need to do to get those Snap-ins loaded correctly? I'm calling the script like this, if it matters: powershell.exe -executionpolicy bypass -file C:\path\to\script.ps1.
EDIT: From running (Get-PSSnapin -registered).count as both SYSTEM and my user account, I can see that the snap-ins are loaded correctly, but still can't figure out why the script behaves differently.
OS is Server 2016, version 1607, this is the script: https://gist.github.com/4oo4/85cec464e123d7f2793745e662d6e7ab
This isn't the answer why your specific script doesn't work under the SYSTEM account but explains how you might troubleshoot your (or any other) PowerShell Script under the SYSTEM account.
Give a man a fish, and you feed him for a day. Teach a man to fish,
and you feed him for a lifetime.
The whole thing around this, is that you can actually open a interactive PowerShell command prompt under the SYSTEM account were you probably not aware of.
Run PowerShell as SYSTEM
There are a few ways to start a interactive PowerShell command prompt but probably the easiest one is using PsExec.
Using PsExec
Download
PsTools
Extract PSTools.zip and just copy PsExec into your executable path
Run PowerShell as Administrator (accept the User AccountControl prompt)
In the PowerShell administrator window, give the command: .\PsExec -i -s -d PowerShell
A new PowerShell command window will open:
(Type WhoAmI to confirm the current account)
From here you can troubleshoot the specific script:
Are there any errors when running the specific script?
Does it hang or crash at a specific command?
Are there any differences with running it under a user account?
If it appears that the script runs actually fine in the SYSTEM window, then I would check for any errors in the Task Scheduler:
Select Task Scheduler Local - Task Scheduler Library in the left pane
Select your task in the middle top pane
(Make sure you have Display All Task History in the right pane Enabled)
In the middle bottom pane check the events in the history tab
I am running powershell script through jenkins. It has one cmdlets which require elevated permission. so i am writing those cmdlet as below
start-process powershell.exe -verb runas -argumentlist "net localgroups administrators domain\user /add"
But this prompts a UAC where i have to manual click yes. then its moves further.
I want to elevate the cmdlet without giving UAC prompt and continue to go ahead....
The account used to run the script has admin permission on that machine.
Besides disabling UAC - which obviously should be the last resort - you may achieve your goal with creating a 'scheduled' task which is set up to run elevated and trigger that task from Jenkins.
The difficulty here will be probably about how to pass information to and retrieve information from the task - maybe you can achieve that via some files of well-known paths.
See here for how to set up such a task and here for how to trigger it.
As I do not have any Jenkins installation right now I could not test it though - sry.
The problem is the switich:
-verb runas
That instructs Windows that you need your code to run as an Administrator.
Remove that, and Windows will stop prompting the user for administrator privileges.
Your next question might be:
But i want a standard user to be able to do things that require administrative privileges.
Sorry, that is not allowed on secure operating systems.
if I'm a standard user
I simply can't just decide to be an administrator
I actually have to be granted those rights.
The 8 year old, or the corporate desktop user, can't just become an administrator because they wrote:
start-process explorer.exe -verb runas
They will need me, or someone from IT, to walk the 6 buildings over to type in my admin credentials - because i actually do have Administrator privileges.
Imagine Life Before UAC
Every developer complaining about UAC, who hates UAC, wants to go back to before UAC. Lets imagine that.
It's 2002, you're running Windows XP SP3
There's no UAC, so you're always a standard user
And you want to run some code as an Administrator.
You can't do that; you're a standard user.
The only solution is to:
Fast User Switch
and get an Administrator to login to the machine
have them run your script
they then logout
and you fast-user-switch back to your own account
UAC is much better; since they can just type their credentials into the UAC dialog:
But I Just Don't Want A UAC Prompt
You might be saying:
I don't care about any of that. I just don't want the UAC prompts. I want it to work like it did in Windows XP
If you don't want the UAC prompts, and you want it to behave like it did in Windows XP: then you absolutely can do that. You are perfectly free to turn off UAC.
Standard users will always be standard users, with no way to elevate
Administrators will always be administrative users, with no need to elevate
And that is your preference, and you can do that.
Many other users don't want to do their day-to-day work as an Administrator. But since you're only running your script on your computer: it's fine.
I need using powershell script to kill a process, however access denied. How can I get admin with powershell script? In addition, I do not want to input Admin account and password manually. The Get-Admin process needs to be done automatically. What am I suppose to do?
You would need to elivate your script, or console prompt, to use the "run as administrator" option. A good example script can be found here on how you might do this in your script.
The meat of the script provided in the link just takes the user running the script and verify if the current session is elavated. If it is not you have to open one up as that in order to kill a process. You would also deal with UAC if you are on Windows, that if the user running it does not have local admin rights you will be prompted to enter credentials.
Snippet of the code that verifies if the execution account is admin:
# Get the ID and security principal of the current user account
$myWindowsID=[System.Security.Principal.WindowsIdentity]::GetCurrent()
$myWindowsPrincipal=new-object System.Security.Principal.WindowsPrincipal($myWindowsID)
# Get the security principal for the Administrator role
$adminRole=[System.Security.Principal.WindowsBuiltInRole]::Administrator
$myWindowsPrincipal.IsInRole($adminRole)
You can find a few other options to get elevated permissions here.
Just do
start-process pwsh -Verb RunAs
or
start-process powershell -Verb RunAs
to get yourself an elevated shell. Then run the command you want
Alright, so, I've been searching online forever, and I can't find anything on this at all.
Basically, what I want to do is run a program from an elevated PowerShell script, but I want the program to run as the standard user.
I need to do this because the program that I need to run requires access to a mapped network drive that the domain administrator accounts don't have access to. So, I basically need a line of code that will take the script out of elevated mode, or some extension to the Start-Program command that will make it run as the logged on user rather than the administrator account that the script is running from.
you could use psexec
psexec -l powershell.exe -executionpolicy unrestricted -noexit -file c:\temp\checkelevated.ps1
-l : Run process as limited user (strips the Administrators group and allows only privileges assigned to the Users group). On Windows Vista
the process runs with Low Integrity.
One way that I have used extensively in the past is to create a scheduled task on the fly specifying the currently logged user as the account that will run the task. The task would run some other script, command, etc. and it would occur in the context of the logged on user. This is possible by using Start-Process to call the schtasks.exe program that will...
Create the task (schtasks /create /tn "MyTask" /tr "powershell -file...." /ru "domain\username")
Run the task (schtasks /run /tn "MyTask")
Delete the task (schtasks /delete /tn "MyTask")
You would just need your script to get the current user, which can be done in a number of different ways. I've also put a 2 second pause in between those calls to schtasks just to ensure they all run.
There are more ways to do it (probably some even better) I guess, but this should also work.
If you need to run an executable or script under currently logged in user from an elevated environemnt, you can use RunAs with USERNAME environment variable passed as user argument:
runas /user:%USERNAME% program.exe
USERNAME environment variable should contain currently logged in user even in an elevated environment.
The generally intended and accepted way to do this is to specify the network UNC path instead of the network drive. You can even re-map the drive in the elevated process if you need it. That's how you're supposed to do it. If you have an account running a process that needs access to a network location, the proper answer is to grant that account the access it needs to do it's job.
However....
Does this or this or this describe the problem you're actually having? It's very unclear what you're trying to do. You've eliminated all context from your question.
If you're trying to run a script that needs to run elevated and needs to access the user's network drive and you can't use a UNC path for whatever reason, then the above three links are what you probably want.
If you really, truly need to impersonate a logged on user -- and I really struggle to think of a situation where I'd need to do this from a script -- then read on.
The alternatives that don't require knowledge of user credentials are:
Use a user logon script instead of a computer startup script. If necessary, grant the local user the permissions they need to run the rest of the script. I can't imagine you haven't thought of this already.
Create a scheduled task which runs as "Domain Users" or some other group that represents the users in question and the "Only run when logged on" is checked. Again, you'd need to grant the user the permissions they need to run the rest of the script, but it wouldn't tie you down to logon only.
Write a program which calls ImpersonateLoggedOnUser, which requires SeImpersonatePrivilege (Administrators have this by default, IIRC). These are native Win32 calls, not .Net, so they will not be straightforward to use in PowerShell. It's been about a decade since I've looked at this, and it used to be a huge pain because it would sometimes still prompt for credentials. I have to think that the increased security in Vista and later (UAC, et al) would have made this even worse. I also have no idea if you have access to mapped drives (i.e., if the impersonation survives network hops). I would choose this method approximately never.
For anything else, I think you will require credentials of the current user. What you'd be doing is credential hijacking, and OS security is specifically designed not to allow that.