Receiver out of sync and missing a lot of files

I’m new to Syncthing, but have synced some folders with success.

I now find I have a folder I can not get to sync correctly.

I have setup a send only folder on a master server (with about 338000 files) the master has now scanned them and has status “Up to Date”.

The receiver has a receive only folder is state “out of sync” and its global state, indicates that it has just over 270000 files and 163 failed files.

I do not understand why it is not syncing the last 68000 files.

Is there a way to get the receiver to re-get the list of files from the master?

I have noticed that new changes are getting synced, just not the renaming 68000 files…

I have tried re-scanning both the master and the receiver, but this changed noting.

I have also tried pressing the “Revert local changes” on the receiver, this also does not seem to do anything.

I would prefer “fixing” the existing folder on the receiver, as it would take several days to re-transfer the data.

I’m using v0.14.51.

As I recall I started the sync before the master completed its scan - baby this has caused some problem?

Check the failed items first I suggest.

Thanks for the input.

As I’m new to syncthing, I’m unsure what to check in the failed files.

One has the status: “filename is invalid”

The rest has “peers who had this file went away, or the file has changed while syncing. will retry later”

As this sync is send only on master and receive only on the slave, I do not understand why files that “went away” would give an error. If they are no longer at the sender, the receiver should not care…

That being said, I’m not sure how to deal with these errors - input would be appreciated

Invalid file name implies that the local filesystem does not support the characters in the file name.

Peers who had this file went away usually means the file changed after sync was started, potentially because you atarted syncing before the scan was finished.

Please post screenshots. We can suggest and hypothesize, but often a look at the web UI clears up things or at least makes for more targeted advice.

Here are screenshots of the master and slave.

Note the difference in file count between the master and slave.

Also the master thinks the slave is at 82%, but slave thinks it’s at 99%.

Is there a trick to getting the slave to re-sync with the master?

Otherwise I’m considering removing the entire directory from the slave and starting from scratch - would be much nicer to be able to get it to sync the renaming files.

Did you check the loga on the receiving device. First I’d try to address the failed items by renaming them, and see where that takes you

Is there any change on the receiver (local/global state, folder status, …)? That the status is syncing not out-of-sync and cpu usage is high(ish) looks to me like it is still working on exchanging and comparing file metadata - if this is after initial setup it can take quite a while. If that’s not the case (well anyway), please check the logs.

Thanks for the input

The status of the slave changed between syncing and out of sync. New items were synchronised from the master to the salve, but not the missing items.

The clients on both master and slave have now auto updated to the new version (v0.14.52) and now it seems to have resolved it self :slight_smile:

Unfortunately it also seems to have started from scratch in some regards, as the slave is now also uploading everything it has locally to the master - this does not make a hole lot of sense as the master is send only, and the slave receive only. I guess all files received at the master will just be ignored.

I believe this will take several days to complete - not ideal in regards to bandwidth usage and time spent syncing.

Receive only does not upload anything, so again, without screenshots it’s unclear what you are referring to.

Sorry, I’m not use to uploading screenshots to a public forum…

I thought the remote status on the master meant the remote is uploading everything, but judging from its speed I no-longer think that is the case.

The slave syncing has slowed to a crawl after the client was updated, but it is still syncing. I guess that when the remote status on the master has reached 100% the sync speed will return to normal.

I will not waste more of you time for now, but keep an eye on it to see if it returns to normal.

Status does not indicate actual transport of data. On upgrade the stored metadata is reset, so what the device are doing right now is exchange and compare each others metadata - as long as the data is the same no file data is transferred. However as you are syncing quite a lot of data, the transfer of the metadata itself will be noticeable. If you also have big files, it would probably speed things a bit up to enable useLargeBlocks (

Thanks, that makes a lot of sense!

It’s now at 63%, so it will probably be finished exchanging meta data in a couple of hours :slight_smile:

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