I’d like to understand what happens if I delete my user account on the TTN account server (http://account.thethingsnetwork.org)?
If I deleted my own user account on the account server, is it going to remove all applications and devices registered under my account and then remove my account? Or, is it going to delete my account info on the account server only without deleting all applications and devices registered under my account ?
Indeed. Good question. As for some devices that use the mac address as a device ID, such as the lopy, it would effectively leave the device bricked as you couldn’t use it anymore as device IDs have to be unique within the ttn.
I think in order to develop practical way of workings for deployment we need to have this fully defined. Also, are items that were keys under one account re-creatable in a new account. I asked the question earlier but didn’t get a satisfactory answer.
Kim
As TTN is a distributed network, the Account Server does not know anything about devices. You can have different applications and devices registered at different Handlers and NetworkServers, or even at private networks that use our account server. Deleting your account simply revokes the access, it is your own responsibility to unregister your devices.
So, if I wanted to delete my account on the TTN account server, I believe I need to do it in the following steps:
STEP 1. I have to unregister all my applications one by one using ‘ttnctl applications unregister [app-id]’. I believe this command will also unregister all devices registered under the ‘app-id’.
STEP 2. Delete all my gateways registered under my user name, using ‘ttnctl gateways delete [GatewayID]’ command for each gateway.
STEP 3. At this point, all my devices and gateways have been deleted from TTN server and TTN account server. Now,
I delete my account on the TTN account server.
Why isn’t the application and its devices simply deleted once the last user account with access has been removed? I mean it seems there’s no way it can even be useful anymore except from blocking the id?
@hoonppark: Correct. We might add a feature to the console that does all this in one step, but for now this is indeed the procedure.
@TijnOnlijn: Because the Account Server does not know anything about devices, and the Backend doesn’t know anything about users. The only link between the two parts is the fact that the Account server signs JWTs (certificates) for users, confirming that the holder of this JWT has access to the devices of a certain application ID.
For future readers:
It seems that might be dangerous: