Description of problem: When /etc/init.d/vncserver is invoked with "restart" or "condrestart", it runs its "stop" procedure and then its "start" procedure. In most init scripts, this behavior is fine. However, in the case of vncserver, a directive to shut down the vnc server instances might not be instantaneous. I suggest a "sleep 10" between the "stop" and "start" lines as a quick fix. Better might be to explicitly wait for the servers to exit. How reproducible: Very Steps to Reproduce: 1. Configure vncserver init script to start a session 2. Connect to that session 3. From another login, invoke /etc/init.d/vncserver restart 4. Observe that although the vnc session goes away, it does not get restarted. Actual results: Since the VNC server has not yet completely exited, it appears to already exist when the "start" is invoked. Naturally, "start" fails to invoke another instance. When the "stopping" instance finally goes away, there is no VNC server for the user on that system, until either reboot or other invocation of "/etc/init.d/vncserver start". Expected results: Existing VNC servers go away. New instances of the configured VNC servers are started. Additional info: This was observed on a fairly fast P4 system. However, I suspect that the more things are running within the VNC session(s), the more likely this result will be.