My backup solution using Syncthing stopped working. It is backing up one folder between two Debian servers. It is all installed using the apt package manager from the stable repo. Screen shots and logs attached:
i have tried multiple configuration, double checked the firewalls, purged everything (database, config file, etc) yet still always stuck at the syncing.
Firewall and config should be all good, the UI and logs show that they are connected. What is not happening is the index exchange (backup global is 0 and only a few bytes were transferred), and that’s nothing you can misconfigure (lets see whether that statement will be coming back to bite me later).
Can you enable the model debug facility? To conveniently get the logs as text, you can use the journal (e.g. journalctl -u syncthing@*yourusername*).
I have no proxy settings on site. Both machines are in the same network as you can see from the log, no firewall rules. I am running ufw as firewall and I have already given permissions for Syncthing to work on both machines:
ufw allow syncthing
ufw allow syncthing-gui
Yes I have fixed the listing port on both machines. I have done that because I have already tried to fix it myself before creating the ticket. I have tried the default, fixed ports, fixed IP, using a different interface.
@Catfriend1, Aren’t they supposed to be connected via TCP? I don’t think UDP is an option to back up important data.
There are indeed non-standard ports (23000) involved, but to me it still looks like there is a connection mixup within Syncthing (at the very least if it’s falsely configured, this isn’t indicated in any meaningful way). The two devices connect on ipv4, then one tries to send a hello message on ipv6 on a connection that isn’t mentioned anywhere else in the log, and then a bit more than 5 min later the ipv4 connection is closed.
prime
[FXOY4] 2018/11/08 13:44:15.442285 service.go:276: INFO: Established secure connection to 7WNN76F-QGLMMBK-I5HMD36-YADCBCJ-ZMOODW2-S6OMT3O-FHFG4BY-CWZGGQX at 10.7.243.68:34711-10.7.243.24:22000/tcp-client (TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305)
[FXOY4] 2018/11/08 13:44:15.442349 model.go:1640: INFO: Device 7WNN76F-QGLMMBK-I5HMD36-YADCBCJ-ZMOODW2-S6OMT3O-FHFG4BY-CWZGGQX client is "syncthing v0.14.52" named "penelopebackup" at 10.7.243.68:34711-10.7.243.24:22000/tcp-client
[FXOY4] 2018/11/08 13:51:45.442629 model.go:1293: INFO: Connection to 7WNN76F-QGLMMBK-I5HMD36-YADCBCJ-ZMOODW2-S6OMT3O-FHFG4BY-CWZGGQX at 10.7.243.68:34711-10.7.243.24:22000/tcp-client closed: read timeout
backup:
[7WNN7] 2018/11/08 13:44:15.441981 service.go:276: INFO: Established secure connection to FXOY4QM-D62YMWF-KSGLJNM-Q7O7UUO-EMA445V-MMARWQN-T2WUQI7-WM3ZLQL at 10.7.243.24:22000-10.7.243.68:34711/tcp-server (TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305)
[7WNN7] 2018/11/08 13:44:15.442037 model.go:1640: INFO: Device FXOY4QM-D62YMWF-KSGLJNM-Q7O7UUO-EMA445V-MMARWQN-T2WUQI7-WM3ZLQL client is "syncthing v0.14.52" named "penelopeprime" at 10.7.243.24:22000-10.7.243.68:34711/tcp-server
[7WNN7] 2018/11/08 13:44:15.477184 service.go:212: INFO: Failed to exchange Hello messages with FXOY4QM-D62YMWF-KSGLJNM-Q7O7UUO-EMA445V-MMARWQN-T2WUQI7-WM3ZLQL at [fe80::250:56ff:fe87:cf4d%ens192]:22000-[fe80::250:56ff:fe87:9bd3%ens192]:38333/tcp-server: EOF
[7WNN7] 2018/11/08 13:51:45.442325 model.go:1293: INFO: Connection to FXOY4QM-D62YMWF-KSGLJNM-Q7O7UUO-EMA445V-MMARWQN-T2WUQI7-WM3ZLQL at 10.7.243.24:22000-10.7.243.68:34711/tcp-server closed: read timeout
Maybe enabling protocol and connections debug facilities can shed some light on what’s going on (@moalibi).
At first I used port 22000 for both machines. And I found out that I should change it. so I have chosen 23000/24000 for one and 22000 for the other. My current configuration screenshot is like this:
The logs are missing lines. Also, I think you should reset to the default config and try again, as it seems you’ve misconfigured something port wise, as the ports don’t line up across the different logs you provide.
@imsodin I have followed your request to select protocol and connections. Should I select everything in the GUI log menu?
@AudriusButkevicius Yes I think that is because I have been trying to do selective logs. When you say reset to default, do mean just put default and dynamic on all address fields?
So I have changed everything to default. and enabled all the log except for http because it is too much. if you want it on I can add it.
I didn’t mean additional facilities (model,connections,protocol is enough), I just want all lines of the log. Now the log is very verbose and it is highly cumbersome to look through it. Please disable everything except those facilities and post the logs again - thanks.