Cannot get Out of Sync, how do we reset?

Hi! First, I love SyncThing. Ben using NextCloud, OwnCloud, Seafile (all self hosted), and wish to not switch again.

Now, I got a problem syncing a folder with 11,500 files to the server, from the server standpoint. Server is out of sync and I cannot fix that.

  • By server I mean the self-hosted syncthing docker-compose from lscr.io/linuxserver/syncthing in version v1.29.7
  • server runs with PUID=1000 and all files belong to that user in the VM and the host
  • clients are Windows v1.30.0 with SyncTrayzor tray
  • folder is shared from main client with unstrusted password, all I want is 2 other machines to keep a copy.

main client → server

main client → client 2

client 2 and server do not share, it’s not a 3-way

client 2 and Main client on the same LAN show 100% sync between each other:

And main client show itself 100% in sync with server:

So what I tried is unshare that folder from the server and from client 2. Then I deleted the folder from /data on the server. Then from main client, I share it with client 2, it does accept, and sync happens and works great. Then from main client, I share it with server, it does accept, and then sync fails because folder does not exist… So I create the /data/folder and also folder/.stfolder/ manually because syncthing cannot create it (?), and then sync starts over. But it fails eventually. Now from ther server, it shows 11,900 files changed, and 25,400 files out of sync ??

What am I doing wrong? Do I need to unshare from main client as well so it thinks it’s a new folder?

No, you love “Syncthing”, no capital T.

I think you deleted the folder without removing it from Syncthing, and the entire point of the .stfolder is that isn’t recreated in this case to avoid propagating the files you just deleted.

Have you changed anything on the server locally (which you shouldn’t do with Receive Encrypted)? Iff you haven’t, have you tried the red “Revert Local Changes” button?

hi, yes I tried this option. Maybe I did not wait long enough and restarted the server at some point, so I just retried. It never ends, and the client originator stays stuck at 36% and still showing 25,421 items out of sync (double the actual number)

Does removing the folder (via Edit → Remove), and readding it to Syncthing help?

If you delete a folder on disk, without removing it from Syncthing, when you manually create the .stfolder, the new folder is considered the same as the old folder.

when i do that, after a little while I get an invite from the main client to add the folder. so i accept and it shows “unexpected items” image now I select “revert local changes” on the server and it seems to work, as the log states to Reverting unexpected items in folder "xx-Documents" (xx-Documents) (receive-encrypted)

it will take some time but looks better than previous! thanks, i will confirm if it went through

so far it looks good, except a minor bug on the server side: no files were transfered to the server at all but it shows as up to date !?

  1. Both clients show themselves and the cloud server as healthy and 100% in sync.
  2. The server ui shows itself and the clients all in sync as well YAYYYY
  3. the minor bug is that this very folder on the server web ui shows as 0 bytes and 0 files. do you think another restart of the docker compose server would fix that?

the file system folder is indeed empty:

bash:root[@]box:/mount/syncthing/data # du -sm ./*
24506   ./Family
56681   ./MP3
1       ./xx-Documents

and it contains 20 folders named 0.syncthing-enc all empty

so I did remove it from the server, unshared it from main client then waited a bit, and all subfolders on the server file system stayed there. then I shared it again from the main client, and the server popped up this new shared folder so I accept it, and then sync fails again:

On the server file system, the .stfolder/ is indeed missing what am i doing wrong??

I restarted server with STTRACE: model and now the sync started correctly, however all files fail to transfer EBUG: receiveencrypted/xx-Documents@0xc0003ab008 new error for V.syncthing-enc/SJ/G4Q2VQP7NFFM502856CRF8RTB8P7AO2ULF83TA9VGIHDRBQ3GMTKPPIQDGQLCNUBM0UDKLUDNUPDS1TCRP2DFJEB38P0PPK7G: finishing: pull: connection closed

is it the swag proxy frontend that could be the culprit? Again, I use the syncthing docker compose from linuxserver.io and the swag nginx proxy from the same team, and the nginx definition is also the sample provided:

server {
    listen 443 ssl;
#    listen 443 quic;
    listen [::]:443 ssl;
#    listen [::]:443 quic;

    server_name syncthing.*;
...

Should I use the quic protocol instead??

enabled quic protocol in nginx for syncthing and i don’t see file receive errors anymore in the logs.

Did I misread the documentation or smth? Where does it say to use quic when available? And why does sync start working only when I enable debug mode?

restarted server again without STTRACE and it still works fine, sync is progressing. really I don’t understand anything at all.

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