I am having a really bizarre routing issue between two ESXi server, that 'appear' to have the EXACT same configuration (apart from having non-conflicting IP addressing, obviously).
Each of my ESXi servers (call them ESXi-1 and ESXi-2) has six interfaces configured.
2 x 10GbE Intel SFP+ for the 'user facing' VM network
2 x 10GbE Intel SFP+ interfaces for the iSCSI network (doing MPIO to my storage array)
2 x 1GbE Intel copper interfaces for the vMotion network
There are 3 VMkernel Port groups configured on each ESXi server (vswitch0: user facing; vswitch1: iSCSI; vswitch2: vmotion)
The vMotion VMkernel Port interfaces are configured for 'Route Based in IP Hash' and the corresponding HP switchports are configured in trunk mode.
EVERYTHING is configured exactly the same, except that when I attempt to ping the IP addresses of the vswitch2 (vmotion) kernels on each server,from the default "user-facing" network, I only get a ping response back from the second ESXi server's vMotion IP address. The first ESXi server doesn't properly route back to my ping source host. The second ESXi server pings back just fine, and I've scoured the esxi configs and my switchport configs and they both look exactly the same (just with non-conflicting IP addresses).
Again, both servers AND their corresponding HP switchports are configured identically.
The sole difference is that ESXi-1 is version 5.5 U2 and ESXi-2 is 6.0 U1.
Something's missing, but I am beside myself as to what it might be. Any ideas?