syncthing "Receive Encrypted" folder repeatedly shows "local changes" for all folders on headless server.

Hi everyone,

I am using the receive-encrypted feature on a headless server (in a docker container). I only use the (untrusted) server as intermediate server for offline synchronization between my devices.

What happens very often is, that the server shows the Unexpected Items Flag and offers to revert local items. Whats funny about it is, that it only seems to recognize all folders as “locally changes items”. So when I press the revert button it reverts the changes usually very quickly… I don’t know what causes this and I would like to know how I can help debug this issue.

I am running the newest version of syncthing on all my machines, including headless server (v1.23.1, Linux (64-bit Intel/AMD Container)). Syncthing runs in a docker container using a shared folder on an “openmedia vault” system (:= Debian), maybe that could be the issue? something like changing permissions or something like that? I told omv to not modify any permissions in subfolders.

Here is my docker compose script:

---
version: "3"
services:
  syncthing:
    image: syncthing/syncthing
    container_name: syncthing
    hostname: myserver-syncthing
    environment:
      - PUID=1000
      - PGID=1000
    volumes:
      - /srv/dev-disk-by-uuid-1c2c36bb-7750-4c45-8924-7e35e4abbf3f/syncthing_enc:/var/syncthing
    network_mode: host
    restart: unless-stopped

The standard logs don’t show anything…

2023-03-01 23:24:26 Connection from XXXX at 192.168.188.21:53084 (tcp-server) rejected: already connected to this device
2023-03-02 02:44:25 Connection to XXXXX at 192.168.188.27:22000-192.168.188.21:22000/tcp-client/TLS1.3-TLS_AES_128_GCM_SHA256 closed: read timeout
2023-03-02 04:48:06 Established secure connection to XXXXXXXX at 192.168.188.27:22000-192.168.188.21:22000/tcp-server/TLS1.3-TLS_AES_128_GCM_SHA256
2023-03-02 04:48:06 Device XXXXXXXX client is "syncthing v1.23.1" named "linux-computer" at 192.168.188.27:22000-192.168.188.21:22000/tcp-server/TLS1.3-TLS_AES_128_GCM_SHA256

Any idea where to go from here and which logs to get?

What type of filesystem is the /srv/dev-disk-by-uuid-1c2c36bb-7750-4c45-8924-7e35e4abbf3f volume?

The number of locally changed items matches the number of directories so it might be a timestamp issue.

its an ext4 type file system. It is accessed through docker…

wondering if permissions get changed throughout? but I have added an “ignore” permissions" settings to all of my shared folders…

and yes, the number of changed items always exactly matches the number of folders… so I have no idea whats going on there.

What do the directory names look like that it thinks are unexpected? A full example or two would be useful

its always all directories that it recognizes as changed… see this screenshot:

image

and file names on my ext4 disk:

the unencrypted directory names are names like:

  • PhoneMedia
  • Camera
  • Pictures
  • programs
  • public …

Nothing noteworthy I guess. Thousands of folders ;). And I have 6 shared folders… same exact problem for every single one of them. Also. If I press the “Revert Local Changes” button it’ll be fine for a short time until all folders at some point revert to a local changed state again…

image

I don’t even know how I would debug this. Thanks for any help here.

1 Like

bump… I really hope anyone has an advice on how I could debug this ;)?

You could copy one of those paths and run syncthing cli debug file FOLDER-ID PATH on it and post the output here. Could give a hint as to what causes the difference.

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