well the Pauses and Discards need to be investigated, it is probably an issue with the ESXi configuration(this is a wild guess, I would need troubleshooting in order to really help) - for instance, it could be the shotgun mode because of beacon probing, or just the HA pinging the hosts.
But Pause and Discards are _not_ CRC errors - thats a different thing. Are you using fiber cables with the SFP+ for the 10GbE connections of the Nutanix nodes?
Our issue was fixed by our network engineer, who has since left the company, unfortunately.
We have few things underway at the moment:
I'm working on learning to use the Foundation tool to image new nodes. Prism images new nodes using the default memory on the CVM and adds the node to the cluster. SInce our current configuration requires more memory on the CVMs than the default, adding these lower memory CVMs to the cluster causes us problems.
I use Revello to have my own Nutanix box to play with, and then we have nutanix in labs and in production. I want to play with AHV and openvswitch more.
Going to start a POC on Data Protection groups. Before we make any more investment into nutanix, we want to see the replication and failover work. I will also will sneak in an AWS backup and load migration by using Ravello.