VNC reconnect to same VM fails
|Assignee:||Hector Sanjuan||% Done:|
|Target version:||Release 3.8|
|Affected Versions:||OpenNebula 3.2|
User report in user list
I seem to have a noVNC-related issue. From time-to-time, I get a "Server disconnected" error, when accessing my VM's through VNC. I've noticed that it usually (only) happens when I was previously connected to that specific VM (for which the error occurs) through VNC for an extended period of time (but not more than 10-15 minutes), closed the connection, and then tried to reconnect (this is where I get the error). It never happened (so far) in situations other than this, if the VNC-connections are short-termed in time, I can close and reconnect to VM's through VNC as many times as I want.
The only workaround I've found so far is rebooting both the front-end and the respective host, which, as you can imagine, is rather inconvenient to do each time this error occurs.
sunstone.log contains the following about the error:
Starting vnc proxy: /usr/share/one/noVNC/utils/wsproxy.py 35822 pcls:5946
/usr/lib/one/sunstone/models/SunstoneServer.rb:234: warning: Insecure world writable dir /usr/X11R6/bin in PATH, mode 040777
127.0.0.1 - - [31/Mar/2012 12:08:17] "POST /vm/46/startvnc HTTP/1.1" 200 30 0.1036
127.0.0.1 - - [31/Mar/2012 12:08:17] "GET /vendor/jQueryUI/images/ui-bg_glass_55_fbf9ee_1x400.png HTTP/1.1" 200 120 0.0116
Traceback (most recent call last):
File "/usr/share/one/noVNC/utils/wsproxy.py", line 282, in <module>
File "/usr/share/one/noVNC/utils/websocket.py", line 794, in start_server
lsock = self.socket(self.listen_host, self.listen_port)
File "/usr/share/one/noVNC/utils/websocket.py", line 185, in socket
File "/usr/lib64/python2.7/socket.py", line 222, in meth
socket.error: [Errno 98] Address already in use
127.0.0.1 - - [31/Mar/2012 12:08:32] "POST /vm/46/stopvnc HTTP/1.1" 200 - 0.0856
Since I'm not an expert in Python, I didn't want to dig myself into the above-named files so far.
Any ideas what should I check, or what should I do?