Xserve G4 stability (random processes crashing)

[Kevin Day]

On Apr 5, 2010, at 1:28 PM, Nathan Whitehorn wrote:

> Kevin Day wrote:
>> On Apr 3, 2010, at 3:48 PM, Nathan Whitehorn wrote:
>>  
>>> Since you say UP kernels have the same problems, other G4 machines seem not to have issues, and SMP G5 Xserves are completely stable, that points at some G4 Xserve-specific piece of hardware. I'd guess the ATA controller. Could you try chroot to an NFS volume mounted from a known-stable machine, or a USB or Firewire disk, and trying the same things?
>>> -Nathan
>>>    
>> 
>> Okay, i've done some more playing... The problem still happens even if TMPDIR, /usr/src and /usr/obj are NFS mounted to another system.
>> 
>> I'm fiddling more, but I think that rules out ATA then. 
>> The problem seems to take a long while to first appear, but once it does appear it happens pretty fast repeatedly after that. Is it possible the fan controls aren't working right?
>>  
> That's possible. The fan control settings are done completely by hardware, though. Can you try with the whole system on NFS (i.e. a chroot or netbooting)?
> -Nathan



Even pure NFS (running inside a jail with all of the jail chroot over NFS) was still crashing.

But, I think I may have figured out the issue...

This box only has 1GB of DIMMs installed, but FreeBSD is somehow seeing 1.25GB of RAM and is apparently trying to use it. If I put 2GB of RAM in there, it correctly detects 2GB and (so far) buildworld is running fine after three reboots.

Mac OS X is only seeing 1GB, and seems to reliably detect that. I'm going to do some more digging to figure out where the wrong memory size is coming from.