Thanks that solved the “problem”.
But is would also mean, without action on pc1, pc2 is not updated. My idea was:
having a certain release of data on PC1. PC2 and others can, if they are informed that a new release is available, synchronise with PC1. It would than run in their responsabily. If i need to initiate the send, its my responsability. At the update might also come to the wroing time.
Only by pausing the folder or remote device on the receiving devices, or by not running syncthing, unless you want syncthing to sync. As you stated, you tell those people, that there is a new release (outside of syncthing), so they know when to resume the folder/device.
New rel is available, I presse on PC1 rescan. The folders on PC2 and e.g. PC3 are on pause. PC1 gets info about a new rel and decides to activate immidiately and presses resume. PC3 waits another 4 weeks until resume is pressed.
To me it is totally not clear, what you want to achieve. Not in terms of Syncthing, but in terms of what data you want to sync when and to whom.
If it’s just that you don’t want to sync continuously on the receiving sides, you can just not run syncthing on the receiving side unless you want to sync. Or alternatively if syncthing must run for other folders, but shouldn’t sync this particular folder, just pause that folder until you want to sync again. All on the receiving side. Scanning is totally irrelevant in this case (you probably want default settings on the sending side, potentially very few scans on the receiving side).
If I push (sender) I change data on the reiver side and I dicdate the time, when the new data (eg. a programm) is active. That may lead to discussions. If send is not working for a particular side, for what reason ever, I will have to take care.
If these responsabilies are on the receiver side, situation for me as the supplier and for the receiver as well, is more relaxed.
Of course there are posibilities to avoid real bad situations, e.g. never send directly to the production system…
Not running syncthing is not a solution. If one starts it per accident, new data is activated per accident.
For a proper solution of my problem, the remote sides should be able to be controled from the sender side (one focal point of control)