Ever get this fixed?
A similar discussion done here :
If you have configured all port groups correctly apart from IP hash then should be some thing on physical network thats causing mismatch.
Please take a look at this article as well : VMware Knowledge Base
I have the exact same issue, DL380's G10 w/ HPE Ethernet 10Gb 2-port 562FLR-SFP+Adapters. The vDS PG's are all using Route based on physical NIC load w/ two active up-links. Cisco trunks ports, no LACP or aggregation. Health Status shows a Non-IP Hash Mismatch. Everything appears to be working fine.
HPE just released an updated driver for the HPE Intel i40en Adapters running under VMware vSphere 6.7. Might give that a shot.
My problem was fixed by replacing the Intel X710s with Broadcom NICs. Also fixed my vSAN performance issue.
I also know this error with a single server in a cluster where all hosts and switch ports are absolutely identical. In the end, the only thing that helped was removing the host from the dvSwitch and move it out of the cluster and then reconnected it. After that the error was gone and all health checks were green again.