Re: X with ssh, (was Novell to acquire Suse)

[Dan Coutu]

Michael ODonnell wrote:

Well, I have had no problem with this when interacting with other
flavors of Linux or even FreeBSD.  So I'm certain it is not my
local system that's the problem.
I'm using ssh to connect remotely like this:

ssh -XCA -l mylogin remote.system.name
   



OK - so you're instructing the SSH clients (via their
command lines) to do X forwarding, compression and
to allow X authentication traffic to pass.  That's a
good start.  Your ~/.ssh/config files should also
have lines in them like this:
ForwardX11 true
 

Okay, added it. No difference.

What about on the desktop boxes?  Have you verified
that /etc/ssh/sshd_config and /etc/default on the
desktop boxes don't have any entries in them that
could be causing trouble?  You want entries like
these in your /etc/ssh/sshd_config:
X11Forwarding    yes
X11DisplayOffset 10
 

I forgot to mention that I had also done this step that Cole had pointed 
out.
This was on the server that I'm trying to run the applications on.  So I
just tried also insuring that the workstation has the same settings. They
were okay.

...and you want to be sure that those capabilities
aren't being overridden in the /etc/default/ssh files,
which are read by the SSH daemons on startup.
 

There are no /etc/default/ssh files on either end.

It is only Debian (3.0) server systems that do not define a DISPLAY
environment variable on the remote end (ssh normally does this
automatically when you use the switches I've specified above) and
so X applications will not display on my local system's X server.
   



Since a number of us have Debian boxes that allow
remote X connections just fine, I can say with a high
degree of confidence that it's much more likely that
the problem here is simply that your local config
needs tweaking.
 

Oh I'm sure that's the problem. I just don't know what the missing piece 
is. :-(

I tried Cole's suggestion of installing xbase-clients but that
didn't do it.  I had been thinking that perhaps I needed to
start with a remote environment that already had a valid DISPLAY
defined because it was a workstation.  Can anyone verify or deny
this theory?
   



BTW, the SSH clients and servers emit very useful
debug info when instructed to be verbose - have you
tried that?
 

I tried that. It nicely told me that it was requesting X forwarding with 
authentication
but still there is no DISPLAY variable defined and thus no X forwarding is
possible.

Dan

_______________________________________________
gnhlug-discuss mailing list
[EMAIL PROTECTED]
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss