Send-Only folder constantly getting out-of-sync

Hello again,

i have another problem with my syncs: I have 2 shared folders that are send only on one device (let’s call it MASTER). Two other devices have a read-write copy of these folders (SLAVES).

After a while the MASTER shows ALL FILES as out of sync.

The second one even has more items out of sync than there are files (maybe folders).

Some time ago: the number of out-of-sync-items also once was equal to files + folders for one of them.

Alright, now since i read the manual i understand what “Override Changes” does and i press it … Now all 3 clients are doing a little bit of “syncing” work … (but no relevant traffic happening) and then all is fine and in sync.

But then after a while (probably 6000s, which is my rescan interval) the two folders get out of sync again.

I just observed that the slave and master are talking to each other (i see the transfer-rates match in sending & receiving) while the slave is scanning and the master is getting out of sync again:

Slave scanning (talking to the master):

Master getting out of sync in batches of 1000 items (in progress):

Can anybody please help? What is going on here? Why is this happening?

As a desperate attempt i already checked if maybe the system clocks run differently (all fine!).

And i already rebuilt the DB on all clients.

Thanks in advance!


This is explained in the docs:

The intention is for this to be used on devices where a “master copy” of files are kept - where the files are not expected to be changed on other devices or where such changes would be undesirable.

In send-only mode, all changes from other devices in the cluster are ignored. Changes are still received so the folder may become “out of sync”, but no changes will be applied.

But there are no changes on the other devices! And it seems the override command does not work (at least not for long).

I just noticed, that the file timestamps on the slave are much newer than on the master. I would expect them to be the same as the master. Very at least after i press override.

Maybe the slave is unable to apply the timestamps to the files?

Where is the syncthing log on linux when running as a service (systemctl start syncthing@sync.service)? I could not find it anyhwere after doing a big search (including google, FAQ, etc.)!

I saw, when i run systemctl status syncthing@sync.service i get about 10 lines of the log. How can i get the rest?

The data might be the same, but metadata (permissions/mtimes etc) being different also causes out of sync.

Furthermore, if you have A and B as send only, and they have different data or metadata, it’s inevitable one of them will be always out of sync, as one will refuse changes of the other.

As far as i know syncthing does NOT sync permissions. But Timestamps yes.

I just remember there is an option something about FAT?

Ignore Permissions
File permission bits are ignored when looking for changes. Use on FAT file systems.

Maybe this helps? I am using FAT.

Can Syncthing not detect this by itself somehow? I mean a FAT file-system.

Solution: Turning on this option (Ignore Permissions) apparently helped.

Thanks for the hint!

1 Like

We might be able to probe for filesystems that don’t keep all the info we set. But on the other hand, silently working around this might cause unexpected data loss for people who are not aware.

I think it’s s till a good idea. You could display a little information marker on the folder telling the user about this. So he can be aware of it.

journalctl -fu syncthing@sync.service

1 Like

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