Recent Upgrade Broke All Connections

I recently upgraded to version 1.27.3 (I do not recall what version I was running previously). I have syncthing running in kubernetes with the LinuxServer image.

Up until the upgrade, everything was working perfectly for ages.

As of now, no connections succeed.

I have an android device trying to connect to a main server.

Using the address quic://syncthing.domain.com, I receive the error no recent network activity.

Using the address tcp://syncthing.domain.com:22000 gives the error no application protocol.

Looking at the logs, everything seems to be running as expected.

Do anyone have any idea what could be the issue here?

I’m pretty sure Syncthing can still connect in 1.27.3, so I think you need to troubleshoot your k8s setup; services, load balancers, etc.

My ingress is traefik, and nothing has changed there.

As far as services, this is what I have for my syncthing namespace:

NAME        TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)                                  AGE
syncthing   ClusterIP   10.43.230.170   <none>        8384/TCP,22000/TCP,22000/UDP,21027/UDP   19d

I have verified that I can telnet to the TLD on port 22000.

And when you do, and you type some garbage, do you get a handshake error in the Syncthing logs?

[I6KAH] 2024/02/13 09:03:03 INFO: Listen (BEP/tcp): TLS handshake: tls: first record does not look like a TLS handshake

Could you share the traefik configuration?

@calmh and @bt90 thank you so much for helping me debug!

@bt90 you tipped me off by asking about traefik :slight_smile:

I have syncthing running with a custom helm chart, and the template changed recently. During rewriting everything, it went from three services to just one, so my IngressRouteTCP/UDP had a mismatch on the service name.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.