This is the mail archive of the cygwin-xfree@cygwin.com mailing list for the Cygwin XFree86 project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: XWin works on Win2K but not on some WinXP clients


Kirk,

Woellert, Kirk D. wrote:

I'm NOT giving up. This is now my mission in life (when I'm not doing what
I'm supposed to be doing). The guy with Win2K can login into the linux box
again (must not have rebooted the box or something for my last post). WinXP
clients no dice.

Okay, I just wanted to make sure you weren't getting discouraged. Sometime after getting this far I get blind sided with an "I'm giving up" message.


I physically changed the the runlevel on the linux box by
invoking:

# init 3
gdm -debug 10 -no daemon

This is good. This is precisely what you need to be doing for the debugging phase.


2nd line invokes the display manager, which by default puts the system back
in run level 5?

The primary difference between run level 3 and run level 5 is that a display manager is running. Trying to move up to run level 5 after starting gdm on your own would cause init to attempt to start gdm again, which would not be good.


Hence, that's why I didnt need to explicity change to run
level 5 as I erroneously did in previous posts?

That's pretty much correct.


Also explains why the system
would display the message "display 0 already active, launching display 1"?

Exactly.


greps taken following that command:

[root@gaia log]# ps -ef |grep xdm
root      4039     1 12 08:10 ?        00:00:01 xdm
root      4043  3646  0 08:11 pts/0    00:00:00 grep xdm

Why is xdm running? I thought that in a clean reboot you found out that only gdm was running? There should only be one of xdm, gdm, kdm running at any time.


We really have to nail down precisely which one of the three is running at startup and make sure that that is the one we debug.

[root@gaia log]# ps -ef |grep kdm
root      4045  3646  0 08:11 pts/0    00:00:00 grep kdm

Good. kdm should not be running.


[root@gaia log]# ps -ef |grep gdm
root      3282  3252  0 07:44 tty1     00:00:00 /usr/bin/gdm-binary -debug
10 -nodaemon
root      3320  3282  0 07:44 tty1     00:00:00 /usr/bin/gdm-binary -debug
10 -nodaemon

I don't think this is good. I don't think there should be two instances of gdm running. Make sure you reboot the machine before doing each test... you don't want to worry about an old process laying around, which would prevent you from seeing the debug messages.


root      3321  3320  0 07:44 ?        00:00:02 /usr/X11R6/bin/X :0 -auth
/var/gdm/:0.Xauth vt7

That's fine.


I deleted all the log files from the /var/log/gdm folder. They were a lot of
them, and they didnt seem to be changing from one day to the next. Following
deletion, of these files I asked the Win2K guy to login, and then I tried to
login, and no new 0:log files were generated in this sub-directory.
Apparently, individual login attempts are not logged here.

Okay, well, you do have a problem above where you are running two instances of gdm and an instance of gdm. All three of those instances are trying to listen on the same network port for XDMCP logins... it is a crapshoot as to which of them is actually handling connection attempts. If it happens to be xdm, then there will not be new log files in /var/log/gdm (since xdm doesn't log there).


 I need more specific guidance from the forum. I keep pasting the log entrys
for the gdm into my mail, and they look the same - debug mode or not. Are
these the log entries you are looking for? How do I discern which entry
corresponds to the Win2K attempt vice the WinXP attempt?

I think we need to chat on the phone. This should have taken only a few minutes to fix. I will contact you off list to make arrangements.


Harold


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]