The communication with vsmserver currently happens over a Unix socket, which means that tlctl only works on the master machine(s). It would be convenient if it worked from at least any machine in the cluster.
Right now, this is the error message you get when running `tlctl session list` on an agent instead of the master: > tlctl: error: cannot contact the master (vsmserver): [Errno 111] Connection refused New users are likely to get confused by this. Hence --- just adding a more descriptive error message here (for example by adding "Make sure you are running this command on a ThinLinc master node") would relieve the situation quite a bit.