F14 and qtnx client

Mon Jan 3 02:11:19 UTC 2011

On Wed, 2010-12-22 at 11:04 +0100, Joachim Backes wrote: 
> Hi all,
> 
> somebody has been successful in connecting the qtnx client to a server
> running the nxserver from nomachine?
> 
> Wondering about the fact that there are only *fc11* versions of the qtnx
> client software :-)
> 
> If starting qtnx after having configured it, I only get an error msg like
> 
> ------------------------------------------------------
> Process started
> stderr> Pseudo-terminal will not be allocated because stdin is not a
> terminal.
> -------------------------------------------------------

That error can be safely ignored.  That's just ssh refusing to
allocation a pseudo tty because stdin to is is not a tty.  But you don't
need one in this case, since it's all going to be done over X11
forwarding.

> I was not successful in finding helpful entries in the forums.

I seem to have gotten further than you but I'm connecting to a freenx
server on another F14 box.  Unfortunately, I have also been
unsuccessful.  I may just load up the NoMachine server on one of my
servers, just to test it out and see how it behaves.

In my case, it goes through a long number of steps and messages out the
console finally ending with this:

Connection to 127.0.0.1 6001 port [tcp/*] succeeded!
stdout> NX> 1000 NXNODE - Version 3.2.0-74-SVN OS (GPL, using backend: 3.3.0)
stdout> NX> 700 Session id: alexandria.XXXXXXXXXX.com-2004-11B427C0BF684A790366A2977B076A36
NX> 705 Session display: 2004
NX> 703 Session type: unix-application
NX> 701 Proxy cookie: 04970d2fcf149f5fee5af219dea4497c
NX> 702 Proxy IP: xx.xx.xx.xx
NX> 706 Agent cookie: 04970d2fcf149f5fee5af219dea4497c
stdout> NX> 704 Session cache: unix-application
stdout> NX> 707 SSL tunneling: 1
NX> 1009 Session status: starting
NX> 710 Session status: running
NX> 1002 Commit
NX> 105 
stdin>  bye
stderr> /usr/libexec/nx/nxserver: line 1585:  9402 Terminated              sleep $AGENT_STARTUP_TIMEOUT
NX> 1006 Session status: running
stdout> bye
stderr> Bye
NX> 999 Bye
Starting NX proxy
NX_COMMFD=24Process started
stderr> Connection to Alexandria.XXXXXXXXXX.com closed by remote host.
Process exited

There it dies and times out and I get the "closed by remote host
finally.  I set up the server to save the session data and found that
the final user connection seems to be failing.  The connection from qtnx
to the nx at server succeeds (even over IPv6!  Yes!!!) and the ssh
connection from nx at server to user at 127.0.0.1 also succeeds but then the X
session back over the forewarded X connection seems to fail with an
"unable to connect to display.

I saved an earlier session by setting up the appropriate parameters in
the node.conf file and ended up with this in the session file:

NXAGENT - Version 3.3.0

Copyright (C) 2001, 2007 NoMachine.
See http://www.nomachine.com/ for more information.

Info: Agent running with pid '7189'.
Session: Starting session at 'Sun Jan  2 15:13:44 2011'.
Info: Proxy running in server mode with pid '7189'.
Info: Waiting for connection from '127.0.0.1' on port '6004'.
Info: Accepted connection from '127.0.0.1'.
Error: The remote NX proxy closed the connection.
Error: Failure negotiating the session in stage '7'.
Error: Aborting session with 'Unable to open display 'nx/nx,options=/home/mhw/.nx/C-alexandria.XXXXXXXX.com-2004-F9AACA89CE1E5FEC50094A113491891A/options:2004''.
Session: Aborting session at 'Sun Jan  2 15:14:19 2011'.
Session: Session aborted at 'Sun Jan  2 15:14:19 2011'.

I've done the same thing with the NoMachine nxclient and it works fine
but that client sucks.  It can only establish one session at at time,
failing with a bizarre error that didn't make any sense that I finally
tracked down to the client trying to listen on local TCP port 12000 for
something about X11 fonts.  After the first session has it, subsequent
sessions catch fire and burn when they can't bind to it, complaining the
address is already in use.  :-P  It also fails miserably if the host is
an IPv6 only host.  It seems their client is IPv4 only even though ssh
is perfectly happy over IPv6 and the qtnx sessions to freenx seem quite
content over v6 and I know X11 forwarding works on ssh over IPv6.
Lamers need to get with the program.  But, it did work to establish a
connection into my freenx server setup (on IPv4) and give me an
X-session back.  So I know the server setup is correct.  Still qtnx
fails.

Seems like, in my case, it's some disconnect in setting up the ssh X11
forwarding and I can't seem to tell where it's broken at.  Worse, it's
got to be something that qtnx is doing differently that's causing the
brokenness.

It also LOOKS like qtnx sets up the ssh connections in a way that can
also allow ssh agent forwarding to work so you don't need a password and
can use hardened encryption keys end-to-end for authentication if you
have your authorized key in both the nx users authorized_keys file and
in your authorized_keys file which is another nice (and bloody obvious)
feature that is totally missing from the NoMachine close source client.
Gag.

My next effort is to try and shim the two ssh sessions and see what the
difference is under the hood in the command passing and primary channel.

> Kind regards

Regards,
Mike
-- 
Michael H. Warfield (AI4NB) | (770) 985-6132 |  mhw at WittsEnd.com
   /\/\|=mhw=|\/\/          | (678) 463-0932 |  http://www.wittsend.com/mhw/
   NIC whois: MHW9          | An optimist believes we live in the best of all
 PGP Key: 0x674627FF        | possible worlds.  A pessimist is sure of it!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 482 bytes
Desc: This is a digitally signed message part
Url : http://lists.fedoraproject.org/pipermail/users/attachments/20110102/7ef4e81c/attachment.bin 