Does it make sense to define a VNC server in an OpenStack controller node?

asked 2015-05-24 16:30:00 -0500

rarylson gravatar image

I have some doubts about the vncserver_listen and the vncserver_proxyclient_address options in the nova.conf config file.

In the OpenStack install guide, the session (Chapter 5. Add the Compute service - Install and configure controller node) suggests to add the configurations below to the OpenStack controller node:


vncserver_listen =
vncserver_proxyclient_address =

Similarly, the session (Install and configure a compute node) suggests to add the following config options to the OpenStack compute nodes:


vnc_enabled = True
vncserver_listen =
vncserver_proxyclient_address = MANAGEMENT_INTERFACE_IP_ADDRESS
novncproxy_base_url = http://controller:6080/vnc_auto.html

I read some parts of the Nova source code and I do not understood why the vncserver_listen and vncserver_proxyclient_address options are also defined in the controller. The first option, for example, is only used by the libvirt driver (it is, code executed in compute nodes). It seems that these options only makes sense in compute nodes.

The guide is wrong, or I have misunderstood any concept? Why someone whould define vncserver_listen in one controller node?

edit retag flag offensive close merge delete