Syslinux - Aug 2009 - My last try to fix memdisk boot problem

Hi community,

since more than one week I try now to boot my client via PXE.

More informations about my equipment:

Client:

Board ASUS K8V-X SE Amibios with BIOS Revision 1501 (newest)
Intel UNDI, PXE-2.0 (build 082)


Server:

Tyan Tomcat K7M Board with Eisfair 1.6.1 (www.eisfair.org)

With exactly the same hardware unattended boots where working for more  
than one year perfectly without any trouble.




INETD.CONF regarding PXE:

START_TFTPD    = yes
TFTPD_ADD_ARGS 




DHCPD.CONF (only PXE part)

DHCPD_PXE_N        =   1
DHCPD_PXE_1_NAME   =   test
DHCPD_PXE_1_ACTIVE =   yes
DHCPD_PXE_1_KERNEL =   unattended/memdisk
DHCPD_PXE_1_INITRD DHCPD_PXE_1_ROOT   DHCPD_PXE_1_APPEND =  
initrd=unattended/fetnd.imz

I also tried the options "raw" and "bigraw" at the end of
the append
line, without success.

fetnd.imz is the correct .imz file that works with my client. I tried  
this out with a floppy disk and everything was working.


Generated (by my dhcp server) default file:

#
# Do not edit this file, edit /etc/config.d/dhcpd
# Creation date: Wed Aug  5 22:44:43 CEST 2009

DEFAULT test
PROMPT 1
TIMEOUT 150
DISPLAY pxelinux.cfg/boot.msg
F1 pxelinux.cfg/boot.msg

LABEL test
KERNEL unattended/memdisk
APPEND initrd=unattended/fetnd.img


S Y M P T O M S:

When my client boots it gets an IP from my DHCP Server. Then it starts  
pxelinux.0.

Then the following happens:

Loading memdisk.............
Loading fetnd.imz...........
Ready

This happens when i Use MEMDISK 3.36.

When I use newer MEMDISK version (I tried until 3.83-pre6) the same  
messages appear, but then the PC reboots itself and does this in an  
endless loop.

So two possibilities:

* freezing after "Ready"
* rebooting after "Ready"

I absolutely dont know what is going on here and I tried everything. I  
believe that all of my settings are correct, and think it must have to  
do with memdisk.

AGAIN: It worked already with the same hardware....no changes

If you have any futher questions regarding hard- or software details  
please tell me and I will post it here.

Hopefully we can fix this problem....it really sucks :)

regards
Bernd

reg at krautmann.net wrote:
--snip--
> AGAIN: It worked already with the same hardware....no changes
> 
--snip--
> 
> regards
> Bernd
Bernd,

Not knowing when the last time it worked was, have you ran memtest or 
some other memory test? Have you checked to see if the system is 
overheating (loose/bad fan/heat sink/clip).

I am more of a hardware guy, so I see hardware issue. Both should be 
relatively painless to confirm, assuming the box is local.

Memory goes bad, fans fail, capacitors blow, in other words, 'no 
changes.. that you are aware of'.

Hardware failure seems logical, if one assumes nothing has changed. You 
didn't mention if you have changed any software prior to the breakage.

Was any software changed prior to things not working / before you 
started making changes to configs and newer versions of memdisk? Also 
you didn't mention that you used a newer version of pxelinux. Memdisk 
MUST be matched with the same version of pxelinux, otherwise the results 
are unpredictable.

Hope this helps,

-Andrew