Session resumption doesnt seem to work

[nckmml]

I dont know if I'm doing something wrong but no matter what I do, restarting teleport terminates all active sessions.

I'm using the official teleport deb repository on debian 12 servers. Node, Proxy and Client are the latest 17.3.2 version.

When I restart teleport on the target node using systemctl restart, all ssh sessions get terminated with an "ERROR: Process exited with status..." message. According to what I found, session resumption should work just fine with no further configuration needed when tsh is being used, starting with Teleport 15. Am I overlooking something here?

[espadolini]

systemctl restart will terminate the main process of the service, kill the whole cgroup associated with the unit (including all shells), then start a brand new process for the service. The Teleport systemd service definition defines a ReloadExec command (which just sends SIGHUP to the main Teleport process) usable as systemctl reload teleport.service which will make Teleport spawn a new copy of itself from disk and then wait until existing shells are done before exiting.