Too bad, I had it in my guts. I first upgraded my winx64 client from 0.12.x to 0.13.x via the red button. Shut down recycled came back up showed its two local folders. So far so good.
Then I went to the three participating macosx 64bit ones still with 0.12.x as well. Clicked their red buttons as well.
They came back up and briefly showed something or I dont even remember properly but the windows station showed on its page that all three fellow macosx participants briefly connected and it even showed the remote macosx ones as 0.13.4 version and synced at 100%
But then the trouble started, the macosx pages showed that they are (two of them one macosx is fine!? weird! ) unknown machine id or unknown state and the page is filled with the basic dialogbox that one which is on the left where it shows about its own state. It shows little traffic and no fellow nodes connected at all and also no syncs or directories which would be on the left hand side.
The odd stuff is, the one mac which I remotely logged into showed on the command line when I start syncthing (0.13.4) up, that it does have an ID and connects to that winx64 machine and even shows those virtual folder names or tags and some stuff that it scans directories but eventually there is no progress and it sleeps for one minute.
Thats all I can say. Two macs died with this syncthing or in limbo state or pending or whatever, one mac made it to 0.13 properly and stayed there.
The windows box shows one mac fine at 0.13.4, the second mac it already knows with its name and shows it at 0.13.4 as well but no sometimes connect or up and down flickering or something weird and the third mac is disconnected completely but already has its version number.
What now? I knew it that this is gonna fail. I am always very good at gut feeling in the bad things :(((((
Where are the IDS actually stored and the local information on the mac? some config file i suppose? But not in the path of the binary files where the syncthing executable and readme is, right? maybe I can still find some config file on the damn macs and reactivate them or edit something that is stalling or disabling these macs?
Maybe worth noting:
Is it also possible that syncthing instances of 0.12.x that have trouble with their syncs, meaning no complete syncs, are not able to upgrade to 0.13.x? But then again that windows machine and the one macosx updated fine.
These sync troubles: I am amazed that syncthing can get itself into unsolvable states, the sync said there were five directories that werent empty and couldnt be deleted. I am wondering as the mac and windows users of these machines are mere users, how can the syncthing navitate itself into these situations that cannot be cleared up actually? If someone delted some folder (via finder or whatever) then syncthing ought to sync this to the other participating nodes, shouldnt it? I dont understand how mere normal simple users can mess up these kind of applications this thoroughly
So how do the participants actually clean up these un-syncable situations again? On the command line output of syncthing I was able to see some real path names and sync problem hints but didnt have time yet to sort it or look at it, but when a normal user can fuck up its path and file names and structure beyond that syncthing can deal with it, this tool is problably too cumbersome for the administrator to constantly deal with these exceptional situations isnt it?