Replies: 1 comment 1 reply
-
After following your suggestion to run multiple threads like so, I can push traffic a lot higher than before. I think I still see some loss though but I'm not sure what I'm actually looking at:
Am I correct in thinking that ens2f0's RX is going to ens2f1's TX and vice versa than the PPS values should line up? So that in this case there's 435kpps getting lost in one direction and 90kpps in the other? |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
This conversation is about to discuss and collect BNG Blaster scaling numbers, recommendations and tweaks.
Beta Was this translation helpful? Give feedback.
All reactions