We use HAProxy to distribute traffic from our upstream servers to our backend servers. Our backend servers were a mix of HPE Gen9 and Gen10 servers. Load were distributed properly via round robin algorithm. Gen9 and Gen10 servers were running on RHEL7.9 OS
But as Gen9 were reaching EOS/EOL, we introduced new set of Gen11 servers running on RHEL8.9 OS. Now we notice HAProxy is not distributing load properly. Gen10 takes more load (almost 80%) compared to Gen11 (20%).
We did all similar performance settings for G10 and G11 boxes. Other than underlying OS all are similar for the boxes.
We also tried removing keep-alive settings at backend and frontend but no luck
Any help on will be greatly appreciated