Service suddenly stopped working: Failed to exchange Hello messages

[tunarium]

Hi, I’m getting connection problems from this morning, this is the error logged by my node:

[XXXXX] 2024/09/03 15:54:14.929502 service.go:397: INFO: Failed to exchange Hello messages with XXXXXXX-XXXXXXX-XXXXXXX-XXXXXXX-XXXXXXX-XXXXXXX-XXXXXXX-XXXXXXX at [::]:22000-94.135.162.130:22000/quic-client/TLS1.3-TLS_AES_128_GCM_SHA256/WAN-P40-2VOS0O52JO5E8S7EB3GJTDJOUU: EOF

The other node appears as Disconnected. If the other node is restarted I can see from my node that the connections start to be established, and arrives to 1 + 4 (I sat up multiple connections) but after a bit the connection are lost:

1 + 2

1 + 1

1

and then the other node shows as Disconnected again.

No setup change has been done before the system started to fail. Could anyone confirm if the issue is related to this status problems?:

https://status.syncthing.net/

Tried to roll back to Window v1.27.10 from v1.27.11 didn’t solve.

Thanks in advance

[Nummer378 (Max)]

@tunarium

Could anyone confirm if the issue is related to this status problems?:

No, the status is fine and it wouldn’t cause QUIC EOFs in any case.

[tunarium]

Thank you for your reply @Nummer378

After some hours investigating suddenly the service started to work again.

I still suspect that some discovery issue was the cause.

The self-signed certificate errors showed in the status page, and the 204 No Content is what makes me think that the two problems are related.

This sync is in place from more that 1 year with multiple nodes and millions of files.

Thanks

[Nummer378 (Max)]

@tunarium

The self-signed certificate errors showed in the status page, and the 204 No Content is what makes me think that the two problems are related.

No this is normal, the discovery servers have always used self-signed certs (they’re authenticated using preshared keys) and the 204 is normal operation as well.

[tunarium]

Ok, thanks, I didn’t know that.

Another thing that worth to be mentioned is that the problem raised after the v1.27.10 to v1.27.11 upgrade. Even if the roll back to v1.27.10 doesn’t seem to be related to the problem’s resolution, because it occurred many hours after.

Any clue of what could be the origin of the problem?

Thanks.

[tunarium]

UPDATE:

Actually This is being shown in the Status Page.