Hyper-V Switch in WSL2 is suspiciously unreliable

Hey all,

I have been trying to grok this issue for months and can’t figure out what is going on by myself. I have a TrueNAS core server with a samba share located at on my LAN, and I would like to use it from WSL2. It used to work great with both sshfs and samba, but since late this summer I have had woes.

At first I was blaming protonVPN because it would get better on the Windows 10 host when I disabled the TAP service after closing the app, but that no longer appears to be the case.

I have a git repo where I keep some notes in markdown on my NAS, but when I try to run git status from the terminal with the samba share mounted, it seems to never complete. (If I try to do the same thing with sshfs it will fail with the error fatal: Unable to read current working directory: Transport endpoint is not connected)

To try and troubleshoot, I popped open wireshark and tried capturing the traffic between WSL2 and the NAS. There seems to be a loop that the traffic pattern follows. There is a burst of samba traffic, followed by some retransmits, followed by the hyper-v switch saying “hey lets do the arp thing again” followed by a bunch of resets by the NAS, ending with WSL and the NAS reauthenticating. I have attached a portion of the pcap as plaintext so you can see what I am talking about.(I have the pcapng but I can’t embed it directly here) Anyone have any ideas as to what this could be?

badsession.txt (45.9 KB)