Ask Your Question
0

novnc "Failed to connect to server" after update

asked 2013-04-17 10:02:48 -0500

anonymous user

Anonymous

Hello all,

After applying an update a month or so back to my controller and compute nodes running Essex, I have been unable to use the VNC feature in Horizon. Prior to this, I was able to use the VNC feature just fine and have made no changes to my nova.conf or system configuration (firewall, hostnames, IPs, etc). I am trying to troubleshoot the issue, but can't seem to find a good place to get any information on what is happening. I have tailed the /var/log/nova-* logs during a request, looked at tcpdumps, and ensured the firewalls are not causing the issue. I can ping by hostname and IP from my controller to my compute nodes. Furthermore, I can connect via vncviewer directly to the compute node VNC port for a given instance and see everything just fine.

I was getting errors with nova-consoleauth but the following article fixed it for me: http://ispire.me/fix-for-broken-vnc-console-on-openstack-essex-under-debianubuntu/.

I am using novnc and the relevant portion of my nova.conf is below:

xvpvncproxy_base_url=http://PUBLIC-IP:6081/console
novncproxy_base_url=http://PUBLIC-IP:6080/vnc_auto.html
vncserver_proxyclient_address=PRIVATE-IP
vncserver_listen=0.0.0.0

Any help is greatly appreciated. Thanks!

edit retag flag offensive close merge delete

Comments

Hi, when you try to connect to your instances via Horizon, nova-consoleauth will log out the host that it's trying to redirect to. Can you verify that the IP address (and port) is listed correctly and that the token is accepted?

rdo gravatar imagerdo ( 2013-04-19 15:16:53 -0500 )edit

Did you try to put: novnc_enable=true in your nova.conf?

Alen Komljen gravatar imageAlen Komljen ( 2013-05-04 02:38:32 -0500 )edit

1 answer

Sort by ยป oldest newest most voted
0

answered 2013-05-03 15:54:28 -0500

david-wittman gravatar image

updated 2013-05-03 16:23:35 -0500

Your issue sounds related to a bug which was recently introduced as part of a security fix for VNC connections. If you check your novncproxy logs you should see a traceback similar to the following:

2013-05-03 00:43:59 TRACE nova.rpc.amqp Traceback (most recent call last):
2013-05-03 00:43:59 TRACE nova.rpc.amqp   File "/usr/lib/python2.7/dist-packages/nova/rpc/amqp.py", line 253, in _process_data
2013-05-03 00:43:59 TRACE nova.rpc.amqp     rval = node_func(context=ctxt,     **node_args)
2013-05-03 00:43:59 TRACE nova.rpc.amqp   File "/usr/lib/python2.7/dist-packages/nova/consoleauth/manager.py", line 102, in check_token
2013-05-03 00:43:59 TRACE nova.rpc.amqp     if token_valid and self._validate_console(token):
2013-05-03 00:43:59 TRACE nova.rpc.amqp   File "/usr/lib/python2.7/dist-packages/nova/consoleauth/manager.py", line 91, in _validate_console
2013-05-03 00:43:59 TRACE nova.rpc.amqp     console_valid = self.compute_api.validate_vnc_console(context,
2013-05-03 00:43:59 TRACE nova.rpc.amqp NameError: global name 'context' is not defined
2013-05-03 00:43:59 TRACE nova.rpc.amqp

Though it has yet to be repackaged, there is a patch to fix the issue https://review.openstack.org/#/c/24608/ (here).

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

Stats

Asked: 2013-04-17 10:02:48 -0500

Seen: 1,079 times

Last updated: May 03 '13