I have a situation where one of three different NFS clients will break after a period of time (some number of days). I find that the broken host and server are continually sending 'ack' packets (to the tune of 100+mb/s). Each side is sending the same packet, and neither side ever retransmits any actual data. 1 Eventually the server will reset the connections and NFS will function correctly for a short time (2-4 minutes). A server restart will fix everything until it breaks some days later.
The environment is a duplicate of other environment, and other NFS clients never have this problem (same OS and app versions).
While troubleshooting, I see wireshark reporting a packet as 'TCP ACKed unseen segment' but I don't understand why. Packet #28517 has Ack=2194563093 and packet #28513 has Seq=2194563093... that's the packet being acknowledged, right? I did another clean capture that looks the same so I don't understand why it's showing up like that. 2
I'm primarily looking for hints debugging the NFS issue (SLES on the client side, EMC VNX on the server side) and confirmation that the wireshark output is wrong (or else that is an actual clue I should dig into).
EDIT: Server and client are on the same switch, same subnet.