Re: gdm troubles. How to 4 X servers?

[Aivils S. <Aiv...@un...>]

>This is really weird! When I activate the display manager startup
>script, /etc/rc.d/init.d/dm, the first server comes up all right, but
>the second and third Xservers crash with a message saying that fb
>can't access /dev/nvidia1 and /dev/nvidia2 respectivelly. I have
>commented out the restrictions on <fb> and <dri> from the
>/etc/security/console.perm file.

This is Nvidia internal trouble.

>Also, I do a probe only ath the
>beginning of the dm file (more on that in a moment). You can see the
>messages in the include log files. However, if I run the prefdm script
>(called from within dm) on its' own from the command line (including
>the probeonly), all three X servers come up all right.

rule - You should keep good stating/probing order.
"rmmod nvidia" and try again.

>Can't figure it
>out at all. Do you have any suggestions. BTW, There are situations
>when doing a probeonly on one board is not sufficient. I tried using
>another one and then the first one failed. What seems to work
>consistently is to do a probeonly on servers 1 and 2 (not zero).

To this day probeonly trick works for all. You might try Nvidia suggested
Option "UseInt10Module" "1" in all Device sections (Nvidia readme).
Of course You can wait for official 2.6.0 support.

>Next question: I wish to attempt using four X servers. I will define
>DumbCon=3 on the boot command line. Question is which VT numbers do
>you suggest? Currently using VT1, VT17 and VT18 (2.6.0-test9 kernel)

I suggest read out dmesg.
$ cat /var/log/dmesg | grep Console

Aivils