I have a customer with multiple clusters running hyper-V 2016. We are noticing dropped RX packets in the clusters. There was to an extent where there was a RX dropped packets alert generated in one of the clusters. So we proceeded to upgrade the Intel NIC firmware as recommended by support, after this, we notice the alert does not happen anymore. But we still see dropped packets.
Is this normal or okay for a cluster/NIC to have rx dropped packets?
Best answer by ShauryaView original
Having very high rates of packet drops is not a normal thing. There is an NCC check host_rx_packets_drop which will throw an alert if the packet drop rate is high.
You can take a look at this article which talks about the same check and possible solutions.