Syncthing sync process stalled at 100% on all hosts

Support
1

I configure one master server for “Send only” sync and some slave servers. Sync process is started and stalled at 100%. Here is screenshot of master server: http://i.imgur.com/U2iVE33.png Here is one of slave servers: http://i.imgur.com/amlPyLU.png

I try to restart Syncthing on all hosts, wait more than 24 hours, but it still at 100%. In Syncthing logs I see no new records with errors.

Also I try to press “Override changes”, but this isn’t help too.

How I can find reason for this stalled sync?

2

Click on the “Out of sync” text and look at the errors next to the file names.

3

I see file list only on master host: http://i.imgur.com/CKUiGYV.png On slave host shows thats all is up to date with empty file list: http://i.imgur.com/y7iouhK.png

4

The other (non send only side) has modified those 6 files, and send only is refusing to accept them which makes it out of sync.

5

Thanks for the info! How can I lookup hostnames where is this problem?

And why “Send only” mode not overwrite (restore) this files from master server by default?

How I can force restore (overwrite) those files from master server?

6

The docs explain how send only folders work. Override changes button overwrites the remote state with the state of the master. The devices that are not “up to date” from the send only devices perspective are the ones to blame.

7

I already try to press “Override changes” but this isn’t solve the problem, slave hosts are still at “Syncing (100%)”. And is there any option to force “Override changes” for automate overriding files without manual actions?

8

No there is no such option, you should make your files read only with permissions to make sure they are not modified.

Do you have ignores on any side setup?

9

No there is no such option This is bad, because the same problem is in Seafile server, and I move our infrastructure to Syncthing for solve this problem.

We are use software, that must have write access to it’s config files (it fails if see that files and folders are read-only), so I can’t mark all files as read-only. But we must always restore original config files from master server, to disable making changes in user side.

I’ll try to write script that emulate pressing “Override changes” via cron, hope it helps solve the problem.

10

You don’t need to emulate, just use the rest api: https://docs.syncthing.net/rest/db-override-post.html

11

Thanks, this is easy! But this isn’t solve the problem. After pressing “Override changes” and wait more hours - slave clients are still with “Syncing (100%)” status, and with no new log entries with problems. Here is on master server: http://i.imgur.com/Mf5KhKA.png Here is on slave server: http://i.imgur.com/T9tbIrY.png

12

There is one extra file on non-send only side that is causing this. Override changes won’t fix this as that file is not tracked/does not exist on the send only side.

13

I have found a workaround via deleting all files on slave server, after this - process begins from 0% and finish successfully.

So the problem seems is when Syncthing start one-way sync with already exists files - it can’t finish, because all files are already here.

14

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