vmk0 and port group setup with routed based on physical NIC load
My vmk0 (management) which inherits the MAC of a host is attached to a VDS port group that is configured to routed based on physical NIC load. A number of our vSphere nodes, though not all within a cluster size of 17, are seeing intermittent connectivity/mac flapping between two uplinks.... regardless of the workload on the sever. It appears that vmk0 is impacted, I am not seeing the same issuers on VMs on the same port group/host, but we haven't run analysis beyond a few hours given that the issue in observation was not occurring. I changed the port group to failover by explicit order and the issue disappeared, however I slipped the active/standby NIC and at some scale the issues returned. Any suggestions as to the next steps to diagnose this problem? Could this be a known issuer with respect to vmk0 and a port group on a distributed vSwitch that is configured to route based on physical NIC load?
I have configured the port group where the management interfaces (vmk0) are attached to use explicit failover order and removed all but one interface. No issues so far, I will flip the active interface every 8-10 hours and resume monitoring. If vmk0 is taking the MAC of the first pNIC, could that be related to this behavior?