Re: Intel xSeries 336 boot failure

[Bryan Allen]

On Oct 16, 2005, at 10:52 PM, Stephen Nelson wrote:


I am having a problem booting an Intel xSeries 336 from cdrom. I have
tried both the 3.7 cd image and the latest snapshot image and both  
have

the same symptoms.


The same machine starts without a problem with Linux and Windows boot
cds. It has a p4 cpu, diskless, and dual broadcom BCM5721 gigabit
ethernet adapters.


Do the x366s come in P4 flavors? Looking at IBM's site suggests they  
only come in Xeon.


Our x336s boot the AMD64 image quite happily. On our boxes I had to  
drop RAM below 4GB, but after that everything was shiny.

--
Bryan Allen
[EMAIL PROTECTED]
http://bda.mirrorshades.net
Cyberpunk is dead. Long live cyberpunk.

Intel xSeries 336 boot failure

[Stephen Nelson]

I am having a problem booting an Intel xSeries 336 from cdrom. I have 
tried both the 3.7 cd image and the latest snapshot image and both have 
the same symptoms.

When the machine boots, it loads the kernel into memory correctly, and 
begins probing the pci bus. I haven't got the exact dmesg as the machine 
never reaches a point at which I can capture it, but the console shows 
something like this:

pcibios0 at bios0: ...
pcibios0: PCI IRQ Routing Table ... (11 Entries)
pcibios0: PCI Exclusive IRQs: ...
pcibios0: PCI Interrupt Router at ... (Intel 82801EB/ER LPC ...)

At this point, the machine hangs for a few seconds, then the screen 
clears and the machine reboots. On the next boot the system fails with 
an error relating to a NMI at 0x0 which I presume is related but may not 
be. The system log does not contain any errors.

I can get into the boot_config utility by using the -c switch, but the 
machine crashes when I proceed in the same manner.

The same machine starts without a problem with Linux and Windows boot 
cds. It has a p4 cpu, diskless, and dual broadcom BCM5721 gigabit 
ethernet adapters.

Is there anything that I can try to get the system to boot? BIOS 
settings perhaps? I've tried disabling the ethernet adapters and 
hyperthreading already with no effect.

Thanks,

Stephen