Detect ALL patches on All Machines?
I have over 1200 machines chacking in to my K1200S, and I want to be certain that I can detect all patches on all computers safely. Can this have an impact on network bandwidth? Will the end user see slowdowns? This is the very first time for these machines to be detected.
1 Comment
[ + ] Show comment
-
You're not using patch labels? I haven't tested it, but I'm pretty sure detecting all patches would slow down everything. - dugullett 11 years ago
-
I haven't created ones yet. I have been going through all of the KKE videos and the presenter suggested Detect ALL on ALL. Given that our numbers are so high, i guess it would impact speed. But it is only checking the signature files, right? With those being so small, could it work? Or should I detect specific machine labels, or patch labels? Also, does detect use the replication shares for signature files or is that direct to the K1000? - easterdaymatt 11 years ago
-
I would use patch labels. I only download patches that are associated to my patch labels to save on space. If you were going to do "all and all" I would only do it on a couple of machines. - dugullett 11 years ago
Answers (2)
Answer Summary:
"With 1200 machines, I would recommend breaking it down by labels." - ShawnCarson "[...]instead of having one monolithic job of 1200 machines, it would be 4-5 different jobs, defined by your machine labels. Ideal target for a job would be in the neighborhood of 250 machines." - ShawnCarson
"With 1200 machines, I would recommend breaking it down by labels." - ShawnCarson "[...]instead of having one monolithic job of 1200 machines, it would be 4-5 different jobs, defined by your machine labels. Ideal target for a job would be in the neighborhood of 250 machines." - ShawnCarson
Please log in to answer
Posted by:
Nico_K
11 years ago
usually you don't see relevant slowdown, except you have already a slow network.
he detect may run long on the clients but usually there is no problem
Comments:
-
if you see an impact, simply run the detects in different runs for smaller groups. - Nico_K 11 years ago
-
easterdaymatt,
Nico is correct here. Detections are generally not resource-intensive on the K1000, your network, or the machines processing them. You should, easily, be able to run one Detect job on 1200 machines simultaneously. If you really need confirmation, you could always try a selective sample while running a network sniffer, and perfmon on the machines.
Ron Colson
KACE Koach - ronco 11 years ago
Posted by:
ShawnCarson
11 years ago
With 1200 machines, I would recommend breaking it down by labels. If you do this, you can still patch all of your machines but instead of having one monolithic job of 1200 machines, it would be 4-5 different jobs, defined by your machine labels. Ideal target for a job would be in the neighborhood of 250 machines.
