Parallel retrieval of user sessions from multiple Windows servers in PowerShell - windows

Question

Assume I have an array of Windows hostnames:
$server_hostnames = #("server1","server2")
For each server I need to find if there are any sessions for a specific user, and then ideally capture all the sessions in a new array.
$results = #()
$account = "JohnDoe"
foreach ($s in $server_hostnames) {
$r = (qwinsta /server:$($s) 2> $null | out-string -stream | sls $account)
$results += $r
}
In this example I have 2 servers but in production I would run it against 2-3000 servers, so parallel execution is extremely important.
I have made several attempts at rewriting the code using jobs, workflows, or Split-Pipeline, but with little success.
Usually, filtering with sls (Select-String) doesn't work, not even with findstr.
Split-Pipeline example:
PS C:\SplitPipeline> $server_hostnames | Split-Pipeline {process{ $_; qwinsta /server:$($_) | out-string -stream | sls $account }}
Split-Pipeline : Cannot bind argument to parameter 'Pattern' because it is null.
At line:1 char:21
+ $server_hostnames | Split-Pipeline {process{ $_; qwinsta /server:$($_) | out-str ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo : InvalidData: (:) [Split-Pipeline], ParameterBindingValidationException
+ FullyQualifiedErrorId : ParameterArgumentValidationErrorNullNotAllowed,SplitPipeline.SplitPipelineCommand

Answer 1

Using jobs you could do something like this:
$servers = 'server1', 'server2'
$account = 'JohnDoe'
$results = #()
$sb = {
Param($server, $name)
& qwinsta /server:$server |
Select-String $name -SimpleMatch |
Select-Object -Expand Line
}
$servers | ForEach-Object {
Start-Job -ScriptBlock $sb -ArgumentList $_, $account | Out-Null
}
$results += while (Get-Job -State Running) {
Get-Job -State Complete | ForEach-Object {
Receive-Job -Job $_
Remove-Job -Job $_
}
}
$results += Get-Job -State Complete | ForEach-Object {
Receive-Job -Job $_
Remove-Job -Job $_
}
However, since you want to run qwinsta against several thousand servers you'll probably want to use a job queue rather than running all jobs at the same time, otherwise the massive number of jobs might exhaust your local computer's resources.