Intermitent unstable connection

Hey! I’m having issues with my connection when using Tailscale, and I don’t have these issues if I open a port and use that instead. For example, when I’m copying a big file from a SMB share the transfer will work fine for a few seconds, and then stop for a few more. I also had similar issues with a ZFS replication. In both cases it was enough to crash the entire operation.

The machines I’m using are

  • Windows 10 with Tailscale 1.34.1
  • TrueNAS SCALE server with Tailscale 1.34.1 running in a container. The TrueNAS server itself is accessed via subnet router.

In the container log I get messages like these (filtered for port 445, there are other messages):

2022/12/29 16:52:33 proxy connection closed with error: write tcp 192.168.68.68:445: connection reset by peer
2022/12/29 16:52:40 Accept: TCP{100.70.xxx.xxx:53721 > 192.168.68.68:445} 76 tcp non-syn
2022/12/29 16:53:36 proxy connection closed with error: readfrom tcp 192.168.68.68:44204->192.168.68.68:445: read tcp 192.168.68.68:445: connection reset by peer

I get a ping between the machines, both normal and tailscale ping. When I run tailscale.exe status on Windows I get this for the other node: active; offers exit node; direct

What can I do to diagnose what’s the problem?

I am facing a similar issue here, running Tailscale on TrueNAS Core in iocage jail. Connection breaks if i try to copy large files over the network throu SMB.

tailscale status shows direct connetion is used.

Log:

root@tailscale:/var/db/tailscale # cat tailscaled.log1.txt
{"logtail": {"client_time": "2023-03-27T14:55:58.78957987Z","proc_id": 1563024017,"proc_seq": 209}, "text": "Accept: TCP{100.64.xx.xx:52125 > 192.168.1.71:22} 76 tcp non-syn\n"}
root@tailscale:/var/db/tailscale # cat tailscaled.log2txt
{"logtail": {"client_time": "2023-03-27T14:56:06.006621183Z","proc_id": 1563024017,"proc_seq": 210}, "metrics": "a_bunch_of_characters","text": "proxy connection closed with error: readfrom tcp 172.16.0.2:48298->192.168.1.71:445: read tcp 192.168.1.71:445: connection reset by peer\n"}
{"logtail": {"client_time": "2023-03-27T14:56:06.006733995Z","proc_id": 1563024017,"proc_seq": 211}, "v":2,"text": "netstack: forwarder connection to 192.168.1.71:445 closed\n"}
root@tailscale:/var/db/tailscale # tailscale --version
1.36.1
  go version: go1.19.6

I’m having a similar issue. Trying to transfer a single file >10GB from remote Synology to computer and they connection interrupts and drops download after about 500MB. Did you find a resolution for this?