Every time I try to sit down and write this script, I spend a day in front of a command prompt... then I realize its too hard there so I open up a powershell prompt, spend 2 more days... and then my head explodes.

I'm not looking for somebody to write it, but maybe just somebody to point me in the direction of something that is similar.

The environment:

• Multiple Server 2008
• Failover Clustering
• Many-person administration
• Running the script on a server that isn't one of the failover nodes
• 100 Resources spread out across 1 cluster, 4 nodes, with about 20 groups.
• About 50% of the resources are java.exe...

The problem:

My coding team doesn't have the time to fix a problem with their code. When failover manager fails over (or even gracefully turns off) the java.exe's they don't always turn off. Leaving us to find and kill erroneous javas.

What I'd like to do:

Get a list of the PIDs for all the JAVA.exe's on each node of the cluster. Then look on each node/server for any JAVA.exe's that aren't on that list and terminate the process.

It seems so simple. I don't know why I've been having so much trouble. I am generally just a script-kiddy, googling something similar and adjusting it to fit my own needs, but this is something a bit too specialized I guess. I could PROBABLY do it with a batch file, but our group and resource names are pretty dynamic and I'd hate to make something that breaks when somebody puts a space in a group name or something.

D:>cluster /cluster:CLUSTER RES /priv | findstr "CurrentPID"

D RESOURCE_NAME CurrentPID 21768 (0x5508)

Our script puts a /priv property for "CurrentPID" . There's about 50 or so on some clusters. No Javas are running on the servers that aren't managed by the cluster. The Get-Cluster powershell commands seem powerful, but I'm unsure how to do this...