Parallel Powershell - Part 1

Speaking of making code go faster...

I learned at the PowerShell Summit this year from Joshua King that doing repetitive tasks or loops one way may be quite a bit slower than doing it one way vs another. I learned that functions get compiled by the PowerShell engine, and don't continue to get interpreted.  I do recommend his presentation to understand how he did his research and validation. Whip Your Scripts into Shape: Optimizing PowerShell for Speed

This I had done without realizing the performance improvement it provided. It was chosen to work better with the Workflow process.

It appears that Workflows get compiled by the interpreter, and get the same speed up.  So these options are Good methods to speed things up. Workflows give the capability to run things in parallel.  Using functions within the workflow the method I used to use the parallel functions of Workflow and made the code easier to write and less restrictive. 

But back to my original story. 

One of the things I was concerned about was, what would happen if the server I tasked to do part of the testing, didn't respond for some reason. Would my Health Check crash? return a big red error? or worse. just stop dead and hang the script with no return. 

Fortunately, I haven't had a script zombie yet doing this in the parallel version of the script. But it was a plague in the previous versions. (a zombie is a script that never stops, waiting indefinitely. )
I did make sure to incorporate timeout settings on the web calls to ensure that the computer would not wait forever for a return. There were other checks that would get hung, so I made sure to verify the server was functioning properly before I ran those.  This seems to be where my scripts get stuck. 

The problem became running tasks in parallel, was what would happen if the server was off or nonresponsive. Fortunately, the parallel requests didn't hang like the previous versions. They did the opposite, they returned nothing.

So I had to devise a method to discover which tests had not returned. So I winced and then tried for a Jimmy Neutron brain blast...  that didn't work.
I tried things and came up with a loop that would check to see if each item in the original list of servers to test was in the results.

One approach was to review the errors in the shell variable  $ERROR.  but that didn't work out in some cases.
So to get a full picture, I found that comparing the list of servers submitted with the resulting Responses.

$Missing = $Servers | ?{ $result.PSComputerName -notcontains $_ }

So this gives us the ability to create some additional entries of the response to mark the missing responses.

$MissingServers = $Missing | foreach {
    $MissingServer = @{
        "URI"               = "https://$($_):443/App/Service.svc?wsdl";
        "StatusCode"        = "Fail 00";
        "StatusDescription" = "Fail N/A";
        "Result1"           = "N/A";
        "Result2"           = "N/A s"
        }
    New-Object -TypeName PSObject -Property $MissingServer
    } #Close foreach

So you can combine you results and get a complete table with :

$Result += $MissingServers

Where $Result are the results of the sucessful test.