Unify /websockify path used for vnc connections

Description In continue to https://github.com/OpenNebula/one/issues/878

Sunstone uses two different addresses for establish websocket connections with novnc.

If you call vnc console from the vm page, sunstone will call:

https://example.org:<port>/?token=jkkg7cdyc7...

Unlike, if you click , sunsone will call:

https://example.org:<port>/websockify/?token=jkkg7cdyc7...

To Reproduce

1. Configure sunstone:

   :vnc_proxy_port: 29876
   :vnc_client_port: 443
   :vnc_proxy_support_wss: no
   

2. Configure nginx reverse-proxy, use config file from https://github.com/OpenNebula/one/issues/878

3. Login to sunstone, and try to connect via vnc

   vnc server disconnected (code: 1006)
   

4. Click , and sunstone will successful establish vnc connection

Expected behavior

All connections will use /websockify path.

Details

• Affected Component: Sunstone
• Hypervisor: KVM
• Version: 5.6.1

Additional context

I found working workaround, just specify missing path in vnc_client_port variable:

:vnc_client_port: 443/websockify

Progress Status

• Branch created
• Code committed to development branch
• Testing - QA
• Documentation
• Release notes - resolved issues, compatibility, known issues
• Code committed to upstream release/hotfix branches
• Documentation committed to upstream release/hotfix branches

[tinova]

noVNC to be fully deprecated in #5422

[baby-gnu]

Since it’s working, I figure out that I needed to tweak a little the nginx timeouts.

When you keep a fixed image (like a login screen), no packet pass and the noVNC client do not seems to send ping to maintain the connection, so nginx close the connection after a time (60s by default).

I increase that to 15 minutes:

        location /websockify {
                proxy_pass http://websocketproxy;
                proxy_http_version 1.1;
                proxy_set_header Upgrade $http_upgrade;
                proxy_set_header Connection "upgrade";
		proxy_send_timeout 15m;
		proxy_read_timeout 15m;
        }