Hi Everyone
Just got a question around setting up Worker nodes in Physical Blades… The Current setup we have is Master Servers in VMs and Worker nodes in Physical Blades…
We have 2 NICs with one port each, and to maximize the throughput we have created a Bond in the card. The mode chosen is Mode 1 ( Active/Backup ) , my suggestion was to setup the Linux Bonding in Mode 5 ( adaptive transmit load balancing policy ) but I have done a research and there were some users they were experiencing Unicast Flooding in their networks…
Has anyone have setup a Worker Node in a Physical Blade with multiple NICs? if so, what was the Bonding Method chosen that works best for you?
I can go with Link Aggregation Bonding Method, but that is my last choice as I do not want to depend on the network team to setup the LAC in the switches ports…
Looking forward to hearing your advises
Cheers