How would one go about tuning the K1 for optimal performance.....and know that they have succeeded?
Been experienceing some "performance issues" with the K1. I have gone in and deleted unused files and the "usual" tweaks. Currently agent task through put is sitting around .1 to .7. I have seen it as high as 2.2. I have the agent interval set at 2hrs with 5 tasks. When looking on the dashboard, the tasks in progress shows Krash upload around 95%. I would welcome any advice and or links. I am running 5.4 and have about 3000+ PC's.
Jim
Answers (4)
I actually would increase your 2 hour to at least 3 or 4. How many machiens do you have checking in per minute. You shouldn't have more then 7 or 8.
What is your scritping update interval. Is it still at one hour? if so you have up to 3000 machiens checking in to see if you have updated any scripts and download the updates. If you don't update scripts that often, I would start my chaning it to a 1 day check and go from there.
Comments:
-
OK so here are my current settings and it seems to be running pretty good after some tweakage.
Agent Run = 4
Inv interval = 0
Scripting = 12hrs
Task throughput = 7
current load avg is .16
Agent logs retained.
Looking at all the associated graphs leads me to believe all is well. - gargone 11 years ago
The Konductor section of this slideshow helped me get started.
http://www.kace.com/~/media/Files/Support/KACE-Konference/2011/US/K1000-Advanced-Topics.ashx
I think two hour check ins would be too much for almost anyone.
Comments:
-
Also check your scripts. I had to go back through an clean up all my scheduled scripts that were no longer needed, but still running. - dugullett 11 years ago
-
That document was exactly what I was looking for. It helped clear most everything up. This kind of info is extremely valuable. Thanx - gargone 11 years ago
-
There's an updated one that was shown at this years conference. It explained in more detail. I'm not sure if those are available yet. I haven't really looked. - dugullett 11 years ago
-
I remember hearing about it. But can't remember where. It was a great konference. - gargone 11 years ago
-
It was the advanced topics class. I got copies of a lot of the powerpoints. Just couldn't find that one. It when into some more detail. Very useful. - dugullett 11 years ago
-
I wasn't able to get into that one. Aren't those online somewhere? If not are you willing to share? - gargone 11 years ago
-
Found it.
https://www.kace.com/support/resources/kace-konference-presentations/KACE-Konference-2012 - gargone 11 years ago-
Good. I've been too busy to look. I knew they would post them eventually. That class was standing room only. That powerpoint explains more than the first one I sent. Still both good. - dugullett 11 years ago
-
By still running do you mean enabled? - gargone 11 years ago
-
The 2 hr was recommended in training as well as from the on-site engineer. I am looking at that article now. - gargone 11 years ago
Just to add some info, we have 25,000+ clients and our Load Average is currently 2.65 with a 2 day interval. 2 hours is a very short interval but for the size of your environment I doubt that is your issue. Not sure if this helps you, but I just wanted to chime in that a 2+ Load Average isn't bad.
Comments:
-
I had the task throughput set to 7 and the check in interval set to 2hrs when I noticed it was 2.2. 0 to10 is "optimal" so they say. Made some changes and will see how it goes. - gargone 11 years ago
