Good Evening Ladies and Gentlemen,
first this is my first question here - so if i forgot anything important please let me know i will supply all necessary information asap. Second i already posted this question elsewhere in a different Linux forum - i hope this is in line with the general policy (the desperation level is high).
I am currently trying to set up a server for scientific calculations. The general idea is that all users can log onto the machine via VNC and upon connecting see a normal Gnome login interface where they can log in with their individual accounts having individual home folders etc. So far this already works well.
The problem is now that whenever the VNC session is closed the individual x-servers are also closed - including all the running software. The combination of multi day calculations and a forced windows update (from hell) on the client machines is just one scenario that underlines the absurdity of this situation.
Now the solution for this problem obviously lies in the configuration of the
xinetd super server that needs to create
XDMCP sessions that dont terminate upon loosing the tcp connection from the
VNC session. Also this seems to be straight forward. Xinetd has a general config file /etc/xinetd.conf and individual configurations for the services that are stored in /etc/xinetd.d/
My xinetd.conf looks like this:
Code:
# Simple configuration file for xinetd
#
# Some defaults, and include /etc/xinetd.d/
defaults
{
log_type = SYSLOG daemon info
log_on_success = HOST PID USERID
log_on_failure = HOST USERID
cps = 200 5
}
includedir /etc/xinetd.d
And my service configuration file called "vnc" looks like this
Code:
service vnc-1
{
disable = no
socket_type = stream
protocol = tcp
wait = yes
flags = NOLIBWRAP
user = standarduser
server = /usr/bin/Xvnc
server_args = -inetd -query localhost -geometry 1024x768 -depth 16 -rfbauth /home/standarduser/.vnc/passwd
type = UNLISTED
port = 5900
}
The curx of the bisquit is there the entry "wait = yes". Without this entry everything works just fine - the sessions are just not resumable. Anyhow with it the server is not connectable anymore - but if working it would be resumable - and i would be happy.
The content of my /var/log/demon.log for starting the xinetd daemon:
Code:
Oct 22 23:26:13 atto-calc xinetd[2913]: Reading included configuration file: /etc/xinetd.d/chargen [file=/etc/xinetd.conf] [line=20]
Oct 22 23:26:13 atto-calc xinetd[2913]: Reading included configuration file: /etc/xinetd.d/daytime [file=/etc/xinetd.d/daytime] [line=28]
Oct 22 23:26:13 atto-calc xinetd[2913]: Reading included configuration file: /etc/xinetd.d/discard [file=/etc/xinetd.d/discard] [line=26]
Oct 22 23:26:13 atto-calc xinetd[2913]: Reading included configuration file: /etc/xinetd.d/echo [file=/etc/xinetd.d/echo] [line=25]
Oct 22 23:26:13 atto-calc xinetd[2913]: Reading included configuration file: /etc/xinetd.d/time [file=/etc/xinetd.d/time] [line=26]
Oct 22 23:26:13 atto-calc xinetd[2913]: Reading included configuration file: /etc/xinetd.d/vnc [file=/etc/xinetd.d/vnc] [line=28]
Oct 22 23:26:13 atto-calc xinetd[2913]: added service swat [file=/etc/inetd.conf] [line=39]
Oct 22 23:26:13 atto-calc xinetd[2913]: removing chargen
Oct 22 23:26:13 atto-calc xinetd[2913]: removing chargen
Oct 22 23:26:13 atto-calc xinetd[2913]: removing daytime
Oct 22 23:26:13 atto-calc xinetd[2913]: removing daytime
Oct 22 23:26:13 atto-calc xinetd[2913]: removing discard
Oct 22 23:26:13 atto-calc xinetd[2913]: removing discard
Oct 22 23:26:13 atto-calc xinetd[2913]: removing echo
Oct 22 23:26:13 atto-calc xinetd[2913]: removing echo
Oct 22 23:26:13 atto-calc xinetd[2913]: removing time
Oct 22 23:26:13 atto-calc xinetd[2913]: removing time
Oct 22 23:26:13 atto-calc xinetd[2913]: xinetd Version 2.3.14 started with libwrap loadavg options compiled in.
Oct 22 23:26:13 atto-calc xinetd[2913]: Started working: 7 available services
and the content of the same log as soon as i try to connect to the server with "wait = yes" via "vncconnect localhost:5900"
Code:
Oct 22 23:27:08 atto-calc xinetd[2913]: START: vnc-1 pid=4352 from=<no address>
Oct 22 23:27:08 atto-calc xinetd[2913]: START: vnc-1 pid=4353 from=<no address>
Oct 22 23:27:08 atto-calc xinetd[2913]: START: vnc-1 pid=4354 from=<no address>
(...to infinity and beyond ...)
Some further information:
The relevant parts of my
/etc/gdm3/daemon.conf looks like this
Code:
# GDM configuration storage
[security]
DisallowTCP = false
[xdmcp]
Enable = 1
MaxSessions = 30
MaxPending = 60
DisplaysPerHost = 4
MaxWaitIndirect = 40
MaxWait = 40
[greeter]
[chooser]
Further i have explicitly added vnc-1 with 5900/tcp to
/etc/services and made corresponding entries in
/etc/hosts and
/etc/hosts.allow
I also tried several different VNC servers - like tightVNC, turboVNC, x11vnc and vnc4server but it workes with none of them (in combination with xinetd). As of now it seems like the complication lies somewhre with either Xinetd or XDMCP.
There might have been a few things i fixed on the way that i forgot to mention but. Thats the overall picture.
I have no idea anymore what to do. Hopefully you guys can help me out here ...
kind regards
Aaron
Debian 6 with Xinetd + XDMCP + tightVNC
