Upgraded to 13 and now my two nodes wont connect

I have syncthing 13 running on Windows 2012 and a Linux vps.

I upgraded to 13 today and since then my two nodes wont connect to each other… I’ve deleted both my folders and the ndoes from each other and re-added them but still no go.

Any advice on where to start checking?

As a note, even though i’ve removed and re-added the nodes to each other, the last seen times are still present from before i did that…

Edit : I ended up having to reinstall the linux end (though im sure either one would of been enough)… Now they see each other as new clients and are working.

1 Like

Did you happen to save the config from the device you ended up having to reinstall?

1 Like

Also, did you actually upgrade both sides?

I’m afraid i didn’t

I did yes.

Please post configs from both sides

I can post the current configs… but i do not have them from when i had this issue…

So there is no issue?

From the first post;

But I wish that someone who ran into this saved their config so we could figure out what the problem is and fix it for others.

I somewhat think it was different versions of syncthing, given a reinstall magically fixed it (or the discovery server address issue)

1 Like

Some dude on Twitter had the same experience. I’ll probably release a v0.13.1 tonight to take care of the discovery thing, if nothing else comes up in the meantime.

Apologies for that…

I had the same problem and copied the config.xml , is that what you need ?

Yes, please.

Couldn’t figure out how to PM here, so I sent you an email at nym.se

+1 One device “Disconnected” in cluster. Both 0.13 and 0.13.1 - same issue.

All devices presently on 0.13.1. The lost device is a remote VPS. Worked perfectly until the 0.13.0 update.

I can access the GUI on the lost device, but it is disconnected from all other devices in the cluster. The other devices correctly discover each other.

Like all the devices, the lost VPS node shows: Listening 2/2 Discovery 3/3 (run my own discosrv)

“ufw status: inactive”, is it safe to assume this isn’t a firewall issue?

I removed devices on both ends and re-added them, but that was not helpful. I am considering a ID / certs reset, but hopefully there is a better option.

Check or provide logs with STTRACE=connections from both sides.

Which discosrv are you running? Since you’ve lost a remote device that presumably relied on your discovery service, this may be a factor. I’ve found that the 3/3 indicator in v0.13/v0.13.1 nodes is not always a true indicator of whether discovery services are in fact working as you expect.

Fwiw, updating the Syncthing package on archlinux (maybe also other linux distros) doesn’t seem to restart the service, so the pre-0.13 binary keeps running and causes a protocol-error. Some may just have to restart the service to make sure the new release is actually taking over.

I have the same problem. My devices don’t connect anymore after upgrade to v0.13/v0.13.1. I’m using my own discovery server (latest version v0.12.2). Before the upgrade, I used the v0.13-beta.3, no connection problems. When v0.13-beta.4 was released I had the same problem as now with the v0.13/v0.13.1 version. So I switched back to v0.13-beta.3 und waited for the v0.13 release. When I activate the local discovery or enter the device IPs directly under “remote device” they will connect.