I have a script that I've written that has a small program attached to it as a dependency and a batch file. I'd like to run the script on ~1600 systems at the same time but before I do that I was hoping to know whether or not I can run a script on that many systems at once without breaking anything.


I know that with the "Force Inventory Update" feature, you should run it on >50 systems, does that same rule apply to the "Run Now" feature?

0 Comments   [ - ] Hide Comments


Please log in to comment

Answer this question or Comment on this question for clarity



There is no run now limit, but caution is always warranted when targeting so many machines for a single action.  Ensure your K1000 Task Throughput setting is set properly for your environment.  A lower value improves K1000 responsiveness when it is running many agent tasks as one time.

Answered 05/07/2013 by: mdri
Blue Belt

Please log in to comment

I want to say the guidance for a physical K1100 is no more than 50 machines at once. Believe me, if you run 50 at once you probably won't have a lot of luck using the WebUI for a little bit. Generally I'll limit to 30 or so. It gets the most done at once while ensuring access to the WebUI for helpdesk tickets and other activities.

Answered 05/07/2013 by: GeekSoldier
Red Belt

  • The upper limit is 50 for most tasks, as documented on the built-in "Force Check-In"script.
Please log in to comment

I'll add my own experience just for informational purposes.  We deployed a script that copied down 300MB of files and then launched one of them, and we did this to about 200-300 machines at once in a location, about 80 separate times.  Each time we did this, the web gui became unresponsive for about 5 minutes.  We did not have any dependencies once the files were copied, so if you have tasks on the K1 that need to run sequentially, I would steer clear of doing this.

Answered 05/07/2013 by: nheyne
Red Belt

  • If you have a large byte size for your dependencies be careful.
  • Letting it trickle down with agent check in every couple hours during non-production time keeps your systems admin, your network admin, and your helpdesk happy!
    • That's pretty much how we handle everything else, we just had time requirements for that particular project to fit within a certain window. It worked, it just probably wasn't ideal.
  • yes , just like you said.
  • Hey all,

    Just a quick update, we did this last night and everything worked as expected. I just kept an eye on our task throughput (currently set to 2) and over the course of about 30 minutes all 1600 systems had run my batch file that makes use of small program we wrote that just verifies the value of an xml node.
Please log in to comment