Need Help Recovering Drive with "Xserver cannot be started....." error. Solaris 8

Appreciate any help with this. my knowledge with Solaris is limited. I've dug into past posts on here, with the same error, but doesn't seem to solve the issue I have.
Running Solaris 8 on a Sun Sparc machine. This drive is used as a test hard drive for machines coming in for repair. System will boot the operating system, but no CDE. Error is "Xserver cannot be started on display :0".
This drive was fully operational with CDE until a few months back when we were using it to test a repair when the repair went into panic mode. The unit shut down and from that point on, the drive lost the CDE.

  1. I have run fsck, and in the past it has worked when the CDE would not come up. Not working this time around.
  2. The disk has plenty of free capacity. Original thought was during panic, machine may have dumped or written data to disk. Not the case, plenty of room.
    Main problem is I do not know Solaris or UNIX based systems. My knowledge is in DOS, so boot structure and file processing in a Solaris system is pretty foreign to me. Not really sure where to start or what to look for. Bottom line it appears that the last few steps prior to the CDE coming up are not correct, and a file has been corrupted. If anyone can walk me through the boot sequence and what brings up the CDE, files involved, etc...would really appreciate it.

Welcome!

The Xserver (normally Xsun on Sparc systems) is crashed. It does all graphics, including the CDE login screen, so you are in a terminal-only console. You can log in to this console and get a shell.

Are there any detailed errors in the /var/dt/Xerrors?
cat /var/dt/Xerrors

Are mouse and keyboard connected? (I remember it crashes otherwise.)
Does your graphics card work? A general front end is the fbconfig command:
fbconfig -prconf

Thanks for the reply!
Yes Sir. Terminal only and can log in as root.
Keyboard, Mouse, and graphics card are all good. The machines we set up have a couple of different hardware configurations. The hard drives we use are basically identical and a second drive on this machine brings up the CDE. Cloning the good drive will not work in this case, the problem drive contains software to perform a function that is not available on the working drive.
Errors found on Xerrors:
" error (pid 451): Server for display :0 terminated unexpectedly 1
/dev/fbs/ffb1: No such device or address
/dev/fbs/ffb1: No such device or address
Graphics Adapter /dev/fbs/ffb1 is of unknown type
InitOutput: Error loading module for /dev/fbs/ffb1

I checked the fbs directory. There is a filename for both ffb0 and ffb1.
I do not know their contents at this point.
Again, I really appreciate the help.

Under /dev/ you find devices (special files). The devices have two individual numbers that address the device driver in the kernel.
Some, but not all, devices behave like a normal file.
You can examine any file with the file command:
file /dev/fbs/ffb1

After running fbconfig:

---Hardware Configuration for /dev/fb---
Type: Double-buffered ffb2+ with z buffer
Board: rev 2 (vertical)
Prom Information: @(#)ffb2p.fth 2.9 98/07/14
FBC: version 0x3241906d
DAC: Brooktree 9070 , version 1 (PAC2)
3DRAM: Mitsubishi 130a, version 1
EDID Data: Available - EDID version 1 rev. 3
Monitor Sense ID: 6 (unknown 43X32 RGB Color Monitor)
Current resolution setting: 1280X1024X67

Ah okay then the kernel driver for the current graphics hardware works.
Can you examine the /dev/fb?
ls -l /dev/fb
file /dev/fb

contents of both ffb0 and ffb1 simply state "special character (194/0) or (194/1).

for /dev/fb, it also comes back with "special character (194/0)."

So there are ffb0 and ffb1 device files (or links to device files).
And the Xserver stumbles over the ffb1, while the fbconfig works with the ffb0.

What yields
fbconfig -list
?
I got it from

Maybe the device file name corresponds to the PCI slot.
The article recommends a "reconfiguration reboot" that can be done with command
reboot -- -r
It discovers the current hardware, creates the device files, and removes non-working device files.

-list comes back with a two column output: "Device Filename" and "Specific Config Program.
For device ffb0 the config program is SUNWffb_config
For device ffb1 it is returning "not configurable".

I have attempted reboot -r prior from both this terminal level and the OSD environment (ok> prompt). I'll give that another try. I'm about to get pulled into the afternoon meetings, so I'll have to break off for now. I'll check back in tomorrow or over the weekend. I'll also do a check on these files on the operational drive we have to see if there are any glaring differences so far. I really appreciate your taking the time to help me out on this.

Update....Issue is resolved.
I'm not sure of the inner workings of the files, but here is what was found.
The machines we work on have a configuration for both single and dual video. At the time this drive was last working correctly, it was setup in the /etc/dt/config/Xservers file as single video when testing the repair. When the repair went panic, and possibly after running fsck to attempt to repair the drive, the Xserver file may have been replaced with a copy configured for dual video. I installed a second video card this morning....sure enough the CDE came up. Went into the text editor and checked the file, and it was configured for dual. Since I'm almost completely clueless to Solaris, that's my best guess as to what happened. So a quick thanks for your time last week. I really appreciate it.

Ah, /etc/dt/config/Xservers
Thanks for the update @KB20164 !