syncthing shows different Globald and Local States in Send Only Configuration

Dear syncthing Community,

I do have a two node configuration with syncthing running. One node acts as the “server” and shares a directory configured with the “Send Only” Option. I need this in my specific case since i just want to replicate data to the oder node. File versioning is also disabled in my case.

My problem is that i do have probably bogus file and directory counts for the “Global States” and “Local States” on both sides. They don´t make sense to me.

I rescanned the folders on both sides and restarted the nodes already.

Here are the screenshots:

Node 1 (“Server” - Send Only)

Ignore Patterns from the directory:

Node 2 (“Receiver”)

From my point of view those numbers clearly do not add up. Whats absolutely weird to me is that even the global state is differing.

Is there a way to reinitiate the sync without the need to delete all the files on the replica side? Is there something else that can do.

Or do you think the files are all in sync and everything is alright because it says “up to date”?!

But right now i am not confident that things are fine in this case.

The global state business is likely, which is fixed on master (will be in v0.14.50).

The discrepancy between the amount of ignored items on the filesystem and the difference between local and global state are to be expected: Syncthing does not add ignored files to the index unless they are already there. Meaning if you set up the same ignore patterns on all devices at folder creation time, ignored items will never make it into the global state.

@imsodin: I think that solved my problem! Thank you very much! I was really scared!

–> I added the ignored file pattern on the 2nd node too and now at least the Local states are matching! The Global state is still drifiting with those 1000 files apart. But thats truly likely to be related to your mentioned issue.

Server: image 2nd node: image

But besides. If you would do a complete resync of a folder how would you do it without loosing all the data on the 2nd node? Is there a way to reinitiate the folders?

thanks in advance!

What’s the purpose of that? If you just want the data to be in sync, you are fine if the folders on both sides state themselves as up-to-date. If you want to clean up the global/local state, you can start syncthing with the env var STRECHECKDBEVERY=0s set to prompt an immediate recalculation of those values. If you really want to resync, as in transfer all the data you need to remove the folder in the UI, then delete the directory on disk and then readd the folder - I don’t see the point of this though.

thanks thats what i wanted to achieve… I wanted to clean the global local state. But i should just trust the “Up to date” state that i´ve been given by syncthing already.

thanks for your advice!

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