I have written a Powershell script and it is saved in a .ps1 file. It only works with the 32-bit version of the Powershell that is located in
%SystemRoot%\syswow64\WindowsPowerShell\v1.0\powershell.exe
The script works when I run it manually but it is not run via Windows Task Scheduler. I am running the task as the same user that runs the script. In the Action part of my task I put the above address as Program/script and I write the full path of my .ps1 file as Add Arguments (optional). But it does not seem to work. I have also tried with putting the parent folder of my .ps1 file as Start in value to no avail.
How can I tell Task Scheduler to run my Powershell script using the 32-bit version?
UPDATE: I have to add here that my script actually opens an Excel file, refreshes it and then closes it. I know that using Excel in a non-interactive environment is a bad idea. But I still don't know if this is the reason my script is not run.
Highly suspect Excel is the reason this appears not to work. Have your script do something non-Excel (e.g. create file) and check if this part was executed fine
Two major gotchas I've come across when automating Excel:
Create empty folders if they don't exist (excel automation bug)
Ensure DCOM security settings are configured to allow Excel to run. This is still required if you are running task as same user who manually runs script.
When DCOM permissions are not set correctly and running the script as an automated task, you will get the below error. Saw this as session was transcribed, and transcription output to text file.
New-Object : Retrieving the COM class factory for component with CLSID
{00024500-0000-0000-C000-000000000046} failed due to the following
error: 80070005 Access is denied. (Exception from HRESULT: 0x80070005
(E_ACCESSDENIED)).
You could adopt your script to determine the current powershell version and invoke the same script with the 32bit version if necessary. Put these lines on the top of your script:
# ensure the script is running with the 32bit powershell
if ($env:Processor_Architecture -ne "x86")
{
$psx86 = Join-Path $env:SystemRoot '\syswow64\WindowsPowerShell\v1.0\powershell.exe'
& $psx86 -noprofile -file $myinvocation.Mycommand.path -executionpolicy bypass
exit
}
Note: You can append parameter to the powershell invoke in case your script requires them.
Related
I have a Powershell script assigned to the context menu and it has been running fine. Recently, after I've reset my win10 system and then manually reconfigured my whole system Powershell begins to behave weirdly. The script won't be run correctly anymore, after a lot of researches I find that the problem is caused by the Set-Clipboard cmdlet which being the PCSX version when launching through the context menu where it being the original version if I launch powershell manually. How comes it happen? My script has Set-Clipboard with -value parameter which is not present in PCSX version and it causes an error.
I'm working on a project where we are remotely deploying software using a configuration management, part of which delivers PowerShell scripts to Windows servers, which are then executed in order to perform parts of our setup and/or configuration.
The current deployment method writes the scripts to disk, but doesn't close the file handle. This was done to make the script more "secure" by preventing any other process from tampering with the file prior to it being executed. In order to run the script, PowerShell must read it from stdin, since it doesn't run a script if it cannot get exclusive access. The invocation looks something like this:
powershell.exe -Command - < C:\temp\some_name.ps1
This has a number of drawbacks, primarily that I can't pass parameters to the script. Additionally, reading large scripts from stdin gets funky with bad characters, line returns, etc.
I'd like to invoke the scripts in a more traditional method, e.g.
powershell.exe -File C:\temp\some_name.ps1 -Param value ...
but also keep in the spirit of making sure nothing can tamper with the script prior to it being executed. To that end I want to sign the powershell script and run powershell with the "AllSigned" execution policy.
The issue lies is that I can't really sign the script on the target server, as it has the same issue as PowerShell running the script...I have to release the exclusive lock to let PowerShell sign the file, but it could get tampered with.
I then decided that if I could sign the script on the server that is delivering it to the target machine that would work much better. However, our configuration management software servers are all Linux, and I have been stumped trying to find some way of signing PowerShell scripts on Linux. Mono has support for Authenticode, but out of the box it's only for exes and dlls. I've tried digging into the PowerShell .Net functions but found they use Cryptui.dll, which is Windows-specific.
At this point I'm grasping for anyway I can get the signature added to the scripts, or else I'll have to fallback to some other way of running the scripts that isn't really native. If possible I'd like to be able to compute the signature in-memory on a string representation of the script, but I'll take a file-based method if that is all I can get.
Not sure if this will work for you, but any time I am getting signature errors, I just add this to the bottom of the script:
# SIG # Begin signature block
# SIG # End signature block
I know it looks benign, but this works for me.
What about compiling your script to .exe ?
ps2exe, or even some builtin .net calls can do that for you.
I have a PS script that opens Excel (Com Object), processes a bunch of information, re-saves, and then sends some critical information via e-mail. The script runs great, and when I run it from the Run Console, it works great as well. However, when I schedule it as a task in the Task Scheduler it is not working properly. The task seems to "successfully run" every single time, but I do not get the output e-mail that I am supposed to get. I have run many other PS Scripts without a problem using the same configuration in the task scheduler. Could this have something to do with opening Excel as part of a script scheduled? Any thoughts are welcome.
I solved this issue using the answer from #briantist in PowerShell script won't execute as a Windows scheduled task, but I wanted to isolate exactly which switch was solving the problem.
It had nothing to do with -ExecutionPolicy, -Noninteractive, -NoLogo, -NoProfile or any other system privilege, user account running the script, etc.
Just needed to add -File in front of the script path in the Task Scheduler > Actions > Arguments field. Without this switch PowerShell was launching and the task history was showing Action Completed, but the script was not executing.
I'm trying to run a task on the Worker Role. I added a file named Startup.cmd with the following content:
icacls "PDFs" /grant Users:(OI)(CI)F
EXIT /B 0
All this does is give the Users group full control to the PDFs folder. The task does run because the Microsoft Azure Activity Log window shows the message Details: Starting role... Application startup tasks are running. [2015-02-04T23:39:28Z] when the Worker Role is deployed. I then log into the machine using Remote Desktop and the security on the folder is not as expected. I then try to run the command manually and I see this:
The file has some weird characters which is making the command fail: icacls.
Can Azure Publishing be causing this?
It looks like your script is being saved with a byte order marker. Try saving the script as ascii or as UTF-8 without a BOM. That might work out better for you.
I'm currently working on automating a powershell script to make a dump of a small DB.
The database is accessible through a company Sharepoint and I am leveraging an IQY file exported from said Sharepoint to read the database contents.
When I run the script manually (i.e. calling it from the command line), it works with no issues, but when I try and start the script from Task Scheduler, the following exception is raised:
System.Reflection.TargetInvocationException
the line that is generating the exception is:
$iqy = $xl.Workbooks.Open($query, 2, $true)
Where $xl is a new Excel.Application ComObject, created correctly and $query is a string containing the full file path of the iqy file.
As for my configuration, I'm trying to run my script on a Windows Server 2008R2, using a local administrator user (same user that is supposed to launch the scheduled script), which is also authorized to access the Sharepoint. I'm running Powershell 3.0. I'm not loading any Sharepoint-related snapins.
The only answers I found poking around hinted at granting "Trust" to the IQY file location and the Sharepoint itself in Excel. That did not help.
The Exception also seems to point at a sharepoint issue, but since the script runs fine when started manually, I cannot understand how it could be anything to do with Sharepoint.
Am I missing some intricacy of the Task Scheduler?
Any help would be greatly appreciated.
Never mind, I seem to have found an answer.
If anyone is encountering the same issue, it appear the Excel Com Object has a bug that does not allow it to run from the Task Scheduler if you set it to run regardless of whether the user is logged in.
To circumvent that create the following 2 folders on the machine where the script is supposed to run:
(32Bit, always)
C:\Windows\System32\config\systemprofile\Dektop
(64Bit)
C:\Windows\SysWOW64\config\systemprofile\Desktop
After creating the folders, it worked as expected.
Source: http://social.technet.microsoft.com/Forums/windowsserver/en-US/aede572b-4c1f-4729-bc9d-899fed5fad02/run-powershell-script-as-scheduled-task-that-uses-excel-com-object?forum=winserverpowershell