folder contents disappearing? when does something become interpreted as deleted? st store on a nas device

[usernamegoeshere]

Oddity here. And urgently need to find out what happened. Some syncthing user is sharing/using two folders on a windows 10 machine. The folders reside on a network drive/share on that windows machine, e.g. a nas network attached drive connected via some network drive e.g. z:\folder\path\to\share\

Anyhow, yesterday we discovered, that stuff was missing at other participating machines, or files not yet arrived.

The drive z:\ at this troublesome machine with the nas, was x-ed red and in a disconnected state. the syncthing localhost webpage showed weird status info for the two syncthing shares, something like folder path unavilable or something.

restarting syncthing from the localhost webpage gui didnt help. it stayed in faulty state.

only a real shutdown of the syncthing.exe process via gui, waiting a bit and restarting main syncthing.exe from scratch brought back the two syncthing objecs in green sate synced state.

meanwhile / before, the network drive z: on that machine had been clicked on and browsed/tested/reconnected.

some artists .mp4 created animation file successfully transfered yesterday after that syncthing issue had been sorted out, and happily arrived at other remote syncthing participating machines. so far so good.

today, some folks in this syncthing group, now claim, that some/lots of stuff from various folders is missing or empty.

is it possible, that syncthing, during such situations with this messed up nas drive being disconnected, and that one machine nas drive giving headache to syncthing process, those error state with folder missing or such, but actually causing then deletes on the other participating machines all of a sudden?

but the situation here is, that not all of those two syncthing shared locations are empty, ofcourse that would have been immediately discovered.

its only like selecte subfolders further down two or three directory levels that people claim they had content in, are now empty or such stuff.

I am trying to make sense of this all and trying to find out what actually had happened and what not.

short: can such drive losses cause deletions? or is this rather impossible or how would syncthing handle such complete(?) disappearance of the root path where stuff would actually be in?

dont even actually really know how to phrase the question.

maybe we have different problems though. is it possible to look at a log file from syncthing what exactly has happened within select directories or something? would help a lot to verfiy some claims here with participants.

thanks lots.

[wweich]

If the network drive went completely offline all of a sudden, Syncthing would stop the folder. Usually with the status “folder marker missing”. It means that the subdirectory .stfolder inside the root of the shared directory does not exist (anymore). This is a safety feature to prevent deletions if a drive does go away.

But if there were network problems or Windows did report the path as beeing accessible for some reasons although it was not, and then Syncthing tried to read some files but at that point Windows reported them as nonexistendt Syncthing will mark the file as deleted and propagate that change. But if Windows for doesn’t

[Andy (Andreas Kaelber)]

It seems to me that the connection to the network drive is not the only problem here. Sometimes I have to deal with the ACLs too, e.g. set in the subfolders of a peer uniformly to 0777 and root:root and the ACLs of the server inherit that, for example. Items which are not synchronized normally are cleaned up after that. Whatever helps again and again is switching off and on again “Ignore permissions” after the respective scanning processes. That also helps from time to time with minor problems.

But in all time on my Linux servers the “Ignore permissions” are activated, all the activities runs better with that.