[WBEL-users] Updates to WB Respin1 and X-windows problems

[Brian Thies]

I run WB Linux on a wide variety of workstations and servers.
It's a hobby, so I am not suffering any financial loss when something goes 
wrong, just losing some time.

On 4/2/05, I decided to update 15 machines. Most of them had not been 
updated for about 6 months, but wiith the release of Respin2, I decided to 
bring my systems to current standards.  I copied the updates directory from 
ftp.gtlib.cc.gatech.edu to an ftp server on my local subnet. I add the entry 
for the ftp server to the yum.conf file in the updates section of each 
machine to be updated using  SSH.
I then connected to the machines and ran "yum update"  logged in as 
superuser to each of the machines. It took most of a day to finish updating 
everything. I typed "reboot" and then "exit" to leave the Secure Shell of 
each machine. (No errors were noted in the yum messages.)

Eleven of the machines had no problems. Three sat with X-windows constantly 
crashing and restarting. Number 4 just sat with a blank screen. None of the 
local virtual consoles could be accessed on any of the machines. However the 
SSH daemon was still working and could be accessed.
Samba was running and the shares were still accessible.

Machine 1 looked the same with a single CPU kernel as the SMP 686 kernel - 
constant x-windows crash and restart


Machine 1 Hostname Noir
Manufacturer Intergraph Model TD-425
Dual CPU PII
512M ECC Registered RAM
60G IDE
CDROM -ATAPI
30 G Ultra-W SCSI  Adaptec dual port controller
Trident Gui 9440 2 Meg Video Card.
Onboard Crystal 4336B PNP Sound
Onboard 3C590 NIC

Machine 2 Hostname Armitage
Manufacturer Gateway Model P133
P-I 133 mhz
256M RAM
HP Megaraid 1M Raid-5 Controller
ATI Rage II video.
Intel 10/100 NIC
Vibra16 Sound Card
Standard IDE controller for CD-ROM
Connected to Proliant Array of 7) 4 gig drives.

Machine 3  Hostname Gantz
SIS chipset MSI motherboard
Promise Ultra-33 Disk Controller
Promise Ultra-66 Disk Controller
52X LG CD-ROM
8.4 Gig Goldstar Drive
127 M Quantum Drive
80 Gig Seagate Drive
540 M WD Drive
128 M Ram
Intel 10/100 NIC
ATI Rage II Video

Machine 4  Hostname - Gatekeeper
Video-4meg  Virge/DX 375 S3 chipset.
Gateway PII-300
128M Ram
40G Hard Drive
Intel 10/100
Crystal ISA sound card.

For some reason the XFree86 VESA drver stopped working with it.

I keep the data and the root directories on separate partitions.
It was a relatively simple matter to reload the root directories with WB 
Respin 2, instructing Disk Druid to retain the data in my other partitions.

Machines 1-3 came back up with no problems. I added the users and configured 
samba, and they were back on the network. I ran yum and the machines were at 
a current software level.


Machine 4 absolutely refused to run X-windows.
I loaded WB Respin2 - no help
s3virge, s3, vesa drivers were attempted after setting the /etc/inittab to 
3.
I used startx from the command line to test dozens of configurations.
Most attempts hung the system so badly, that a manual power-off was required 
to release it.
Auto-detect did not work.
The vga driver brought up x-windows. (Definately not something to leave 
permenantly!)
VGA with 4 bit color at 640x480 was the only combination that worked.
Monitor is a NEC Multisynch FE990 - I manually selected the FE950 when the 
DCC probe did not work. (It used the same settings)

Note:
On Redhat 8.0, the installer had a test button to test the configuration 
before rebooting to x-windows. No such feature in EL3.

I finally pulled the S3Virge/DX card and installed an ATI Rage II 4 meg 
video card.
The card was auto-detected and XFree86Config was correctly updated.
Setting /etc/inittab to 5 and x windows started automatically.
I ran yum update and everything is now fine.

Anyone have any insight as to what happened to cause the vesa driver to stop 
working on machine 4?

Any comments on the constant x-windows crashing on the first three machines?

Brian