VNC's XDMCP requests now fail?
For a while now I've been tunneling VNC over SSH and everythings worked
fine. I'd configured inetd to kick off a VNC session on the remote
system and restricted to only allow connections from the loopback
address. However, I noticed today that this was now broken on several
(most) of my systems. The VNC session now immediately terminates on all
but one system.
After bit of searching I found the following errors in syslog:
Sep 3 17:45:34 thor Xvnc[5787]: connect from localhost (127.0.0.1)
Sep 3 17:45:35 thor wdm: Cannot open config file. Using builtin defaults
Sep 3 17:45:35 thor wdm: X Error of failed request: BadAlloc (insufficient resources for operation)
Sep 3 17:45:35 thor wdm: Major opcode of failed request: 45 (X_OpenFont)
Sep 3 17:45:35 thor wdm: Serial number of failed request: 50
Sep 3 17:45:35 thor wdm: Current serial number in output stream: 51
Sep 3 17:45:35 thor wdm: Greet: guarenteed_read error, UNMANAGE DISPLAY
Sep 3 17:45:35 thor wdm: Greet: pipe read error with /usr/bin/X11/wdmLogin
These appear to kick off every time the VNC client attempted a
connection and inetd attempted to start the VNC server.
I can replicate the errors by running the following command on the
remote system:
Xvnc -once :1 -query localhost
If the -query option is dropped, and thus VNC's XDMCP request, VNC works
fine.
I've tried switching display managers from wdm to xdm with no change in
the error message. I've also tried both tight and real vnc on the
remote system without any change in the error.
Relevant portions of my configuration files:
/etc/X11/wdm/wdm-config:
# Don't listen for XDMCP
!DisplayManager.requestPort: 0
/etc/inetd.conf:
vnc stream tcp nowait nobody /usr/sbin/tcpd /usr/bin/Xvnc -inetd -query localhost -once -geometry 1024x768 -depth 24
/etc/hosts.allow:
Xvnc: LOCAL
/etc/hosts.deny:
ALL: ALL
Any ideas?
--
Jamin W. Collins
Remember, root always has a loaded gun. Don't run around with it unless
you absolutely need it. -- Vineet Kumar
Reply to: