UPDATE: Modified the script to work within the bounds of PS1 as required by SQLPS.
Changed:
IF($property.Value -match $regex){
$currentBadLine = (ConvertTo-Csv $_ -NoTypeInformation -Delimiter $delimiter);
$badLines += $currentBadLine[1,2]
};
To:
IF($property.Value -match $regex){
$badLines += $_ | Select-Object | ft -autoSize;
};
Prints a new header for each bad line, but it's not the end of the world and not worth the effort to prevent.
I have a Powershell script that pre-processes CSV files before they have a chance to screw up my data imports.
On two servers in a row now I have confirmed that the PS Major Version at least 2, and that the following code snippet runs fine in Powershell ISE. The purpose of the code is to read in each line of the CSV, and then loop through the columns looking for the regex pattern in the $regex variable. When it finds one I want it to keep track of the error before fixing it so I can write an error log later before outputting a cleaned up file ready for import.
%{
Foreach($Property in $_.PSObject.Properties){
IF($property.Value -match $regex){
$currentBadLine = (ConvertTo-Csv $_ -NoTypeInformation -Delimiter $delimiter);
$badLines += $currentBadLine[1,2]
};
$property.Value = $property.Value -replace $regex;
}
$_
}
But once I put that code into an agent job the agent complains:
'The term 'ConvertTo-Csv' is not recognized as the name of a cmdlet,
function, script file, or operable program. Check the spelling of the
name, or if a path was included, verify that the path is correct and
try again. '
The question then is this: Is the Agents Powershell subsystem using a different version of Powershell than the rest of the system? If so, how do I find out which version the Subsystem is using and if possible upgrade it so I can fix this.
The server is running:
Windows Server 2008 R2 Enterprise
PS Major version 2, Minor 0 Build -1 revision -1
SQL Server 2008 R2 SP1 10.5.2500.0 64 bit
Thanks!
Yes, proper PowerShell support isn't really implemented until SQL Server 2012 (an even that is a bit flakey as to what cmdlets it supports)
In 2008 and R2 the agent's powershell implementation is actually a minishell, created by the now (thankfully) deprecated make-shell.exe utility, which only allows v1 cmdlets to run, and disallows the use of Add-PSSnapin so you can't add anymore cmdlets.
To get proper powershell support, you either need to shell out and call an external script, or run the job as a windows scheduled task rather than an agent job.
The following article explains a lot about why powershell support in 2008 R2 doesn't work like you think it should:
The Truth about SQLPS and PowerShell V2
One work-around: Export-CSV to a file, then Get-Content from the file.
$rows = ,(New-Object PSObject -prop #{a=7; b='stuff';});
$rows +=,(New-Object PSObject -prop #{a=77; b='more';});
#To run from SQL Job, change from this:
$csvrows = $rows | ConvertTo-CSV -NoType | % {$_.Replace('"','')};
write-output $csvrows;
#to this:
$rows | Export-CSV -NoType "C:\Temp\T.csv"
$csvrows = (Get-Content "C:\Temp\T.csv") | % {$_.Replace('"','')};
write-output $csvrows;
Related
I've setup a SQL Server 2016 agent job that has 2 steps that is executed by the same service account, Step 1 executes an SSIS package that writes 2 files out to a share. Step 2 executes embedded powershell code that combines these 2 files into a single file and creates a file on that same share where the 2 files were created by SSIS. When I execute the job, Step 1, the SSIS package completes successfully and creates the 2 files on this share. But when Step 2 executes I get the error below. Just a point of info, when I run this script locally from my windows 10 box, I have no issues.
I guess my question is, if SSIS can execute successfully with the same service account can this be a permissions issue for powershell scripts? I'm not a server admin so if this is a stupid questions please forgive my ignorance. Any help/direction anyone can provide would be appreciated.
Message
Executed as user: XXX\sqlsvc02. The job script encountered the following errors. These errors did not stop the script: A job step received an error at line 12 in a PowerShell script. The corresponding line is 'Get-Content $pathCounts, $pathDetails | Set-Content $outFile'. Correct the script and reschedule the job. The error information returned by PowerShell is: 'Cannot find path '\xxx\shared\MonthlyReports\BoundPoliciesAddedAtFaultEndorsementWithin10Days_Counts.csv' because it does not exist. ' A job step received an error at line 12 in a PowerShell script. The corresponding line is 'Get-Content $pathCounts, $pathDetails | Set-Content $outFile'. Correct the script and reschedule the job. The error information returned by PowerShell is: 'Cannot find path '\xxx\shared\MonthlyReports\BoundPoliciesAddedAtFaultEndorsementWithin10Days_Details.csv' because it does not exist. '. Process Exit Code 0. The step succeeded.
Here is my powershell script:
# Define variables
$path = '\\xxx\shared\MonthlyReports'
$pathCounts = Join-Path -Path $path -ChildPath 'BoundPoliciesAddedAtFaultEndorsementWithin10Days_Counts.csv'
$pathDetails = Join-Path -Path $path -ChildPath 'BoundPoliciesAddedAtFaultEndorsementWithin10Days_Details.csv'
$date = Get-Date
$dateStr = $date.ToString("yyyyMM")
# Define combined file name
$outFile = Join-Path $path -ChildPath("BoundPoliciesAddedAtFaultEndorsementsWithin10Days_" + $dateStr + ".csv")
# Execute combine files code
Get-Content $pathCounts, $pathDetails | Set-Content $outFile
Here is an image of what Step #2 looks like:
I have 2x mini scripts which are fairly similar.
The first backs up our SQL Jobs and runs fine:
Import-Module "SQLPS" Get-ChildItem -Path SQLSERVER:\SQL\MYSERVERNAME\Default\JobServer\Jobs\ | %{$_.script()} | out-file -Filepath "MYFILEPATH_$(get-date -f yyyyMMdd).sql"
The second is meant to back up our SQL Stored Procedures, and does not produce results in the output file:
Import-Module "SQLPS" Get-ChildItem -Path SQLSERVER:\SQL\MYSERVERNAME\Default\Databases\MYDBNAME\StoredProcedures\ | %{$_.script()} | out-file -Filepath "MYFILEPATH_$(get-date -f yyyyMMdd).sql"
This problematic second script gives the below error:
Get-ChildItem : Cannot find path 'SQLSERVER:\SQL\MYSERVERNAME\Default\Databases\MYDBNAME\StoredProcedures\' because it does not exist.
It used to run successfully until around the time I upgraded to Windows10. Any suggestions on how to correct this second script?
Here is the story. I have a fileshare that is replicated between 2 servers located in different places in the world. DFS will not replicate a file if it has only been viewed, but I wouldn't want to delete that file/folder because it was used within the time period I have set (7 days). So to make sure that I don't remove still used files I have to check both locations for their LastAccessTime.
I currently have this
Set-ExecutionPolicy RemoteSigned
$limit = (Get-Date).AddDays(-7)
$PathOne = "FirstPath"
$PathTwo = "SecondPath"
$ToBeDeletedPathOne = Get-ChildItem -Path $PathOne -Recurse -Force | Where-Object { !$_.PSIsContainer -and $_.LastAccessTime -lt $limit }
$TobeDeletedPathTwo = Get-ChildItem -Path $PathTwo -Recurse -Force | Where-Object { !$_.PSIsContainer -and $_.LastAccessTime -lt $limit }
$DiffObjects = Compare-Object -referenceobject $ToBeDeletedPathOne -differenceobject $ToBeDeletedPathTwo -IncludeEqual
$ToBeDeletedOverall = DiffObjects | where {$_.SideIndicator -eq "=="}
After this, I loop through and delete the files that are marked for deletion by both locations.
Part of the problem I have is that there are a tremendous amount of files and this can take a very long time. So I wanted to make it better/faster. My idea is to have this script run the scan as a different script on each FS server and wait for them to return the output. That way it can scan on the local machine easier than remotely and it would scan both locations simultaneously.
The other part of the problem comes in with the fact that I have no idea how to do this. I will continue to work on this and if I solve it, I will post here in case anyone in the future finds this useful.
You could run everything locally. Copy the script to the machines you want (make a script to copy them if you need to) then use something like PSTools to kick them off on the local machines. This should essentially run the script simultaneously on all machines.
So, I am fairly new to PowerShell and need to create a script to rename the computers in our office. That portion of the script works. The part I am having trouble with is the output.
I have this set in task scheduler, but when it runs I do not see if the rename was successful. Below is my script and below that is what goes into the text file.
start-transcript -path C:\Users\abhagwandin.SENECA\Desktop\RenameResults.txt
$CSV = Import-Csv "C:\Users\abhagwandin.SENECA\Desktop\Computer Desktop Names Test.csv" -Header OldName, NewName
Foreach ($name in $CSV)
{
write-output $name
netdom renamecomputer $name.OldName /newname: $name.NewName /userd: admin /passwordd: pass /usero: admin /passwordo: pass /reboot /force
}
stop-transcript
-------------------------------------------------------------------------------
**********************
Windows PowerShell Transcript Start
Start time: 20150520154216
Username :
Machine : (Microsoft Windows NT 6.1.7601 Service Pack 1)
**********************
Transcript started, output file is C:\Users\abhagwandin.SENECA\Desktop\RenameRe
sults.txt
OldName NewName
------- -------
JFLAHNYCD1 JFLAHERTY
**********************
Windows PowerShell Transcript End
End time: 20150520154218
**********************
You know that renaming a computer through a cmdline command in powershell instead of using the built in cmdlets can give problems with the output if you don't parse the output (and preferably create a new object for it)?
Why don't you use rename-computer or the rename() method of the win32_computersystem wmi class? Both can be used remotely so you don't even have to schedule tasks that way.
Just create an input file with the current name and the desired names and use a loop to process them.
Can somebody tell me the best way to delete a series of files between April 1, 2014 and April 17, 2014 on Windows server 2003 command line
This PowerShell oneliner will do the trick. You can remove the back-ticks and put the code all on one line if you want.
Get-ChildItem C:\data\*.* | `
Where-Object {$_.lastwritetime -gt '4/1/14' -AND $_.lastwritetime -lt '4/17/14'} | `
Remove-Item
I found a shorter version of the script here. I prefer to use the full Powershell cmdlet names though.