by Klaus Graefensteiner
22. July 2008 07:17
Introduction
Every now and then I need to investigate bugs that get exposed because countable "things" are less or more than expected. In my special case I was dealing with an application that manages subscription handles to memory registers of programmable logic controllers (PLC). These programs are called Data Access Servers. Occasionally we get calls form customers reporting that the number of handles is less, or sometimes more than expected. The first step to debug this situation is to find out, which handles are missing or which ones are duplicates. This blog post describes how PowerShell's Compare-Object cmdlet makes this task, which used to be a pain in the "peep" now a piece of cake.
Figure 1: Topology analysis of planet Jupiter
Procedure
The PLC memory addresses that clients are subscribing to are maintained in a list data structure. This list can be displayed in an MMC snap-in and it can be exported to a text file.
Step1: Export the item list
Open the Wonderware SMC and navigate to your DA Server node and the to Diagnostics\Client Groups\YourGroupName. Right click and export items into a text file.
I
Figure 2: Exporting Client Group Items
Figure 3: Item list in Notepad
Step2: Fire up PowerShell
Script for missing items
This script requires two files. One file that got exported with the correct number items. I named it ClientItemsBefore.txt. The other file was exported when items were missing. This is called ClientItemsAfter.txt.
1: #Load strings from first file
2: $a = Get-Content -Path $home\desktop\ClientItemsBefore.txt
3: #Initialize arrays of item names that we are going to compare
4: $ins1 = ,0
5: $ins2 = ,0
6:
7: #Exclude headings and start with first data line
8: #Get only the name of the item
9: for($i = 1; $i -lt $a.Length; $i++)
10: { 11: $in = ($a[$i].split("`t"))[0] 12: $ins1 += $in
13: }
14:
15: #Load strings from second file
16: $a = Get-Content -Path $home\desktop\ClientItemsAfter.txt
17: #Exclude headings start with first data line
18: for($i = 1; $i -lt $a.Length; $i++)
19: { 20: $in = ($a[$i].split("`t"))[0] 21: $ins2 += $in
22: }
23: #Find the differences in the two lists
24: Compare-Object $ins1 $ins2
Script for duplicate items
In this case I exported the items into a file called ClientItemsWithDuplicates.txt.
1: cls
2: #Load strings from first file
3: $a = Get-Content -Path $home\desktop\ClientItemsWithDuplicates.txt
4: #Initialize arrays of item names that we are going to compare
5: $ins1 = ,0
6:
7: #Exclude headings start and with first data line
8: #Get only the name of the item
9: for($i = 1; $i -lt $a.Length; $i++)
10: { 11: $in = ($a[$i].split("`t"))[0] 12: $ins1 += $in
13: }
14:
15: #Get rid of duplicates
16: $ins2 = $ins1 | sort -Unique
17:
18: #Compare before and after
19: Compare-Object $ins1 $ins2
That's it
Before PowerShell I would, in the missing items case, combine the two item export files into one file. Then I would use a LogParser query with a Having clause to find the ones that are not duplicates. If I expected duplicates in one file, then I would use LogParser and search for them with an appropriate Having clause to filter them out. Is PowerShell going to replace the LogParser?
The script and item export files can be downloaded here: FindingMissingItems.zip
Subscribe
