After many many crashes and a new board: (Was: Re: Fatal LOR: PV ENTRY (UMA zone) @/home2/src/sys/vm/uma_core.c:2033)

From: "Willem Jan Withagen" <wjw_at_withagen.nl>

> Well the Tyan board is finally in. Look like a quality PCB, althoug some of
the
> connectors are somewhat far away for the large case I'm using..
> First need to do some paid work, but then I'll start rebuilding the system.
>
> I guess we'll know more by tomorrow. :)

It is depressing in a sort of way.....
But I've been beating the pulp out of the new board and thus far no crashes.
Load has gone a high as 35, with over 200 processes in the queues. All from
different make processes....
The baoard was my favourite since the Tyan boards where told the run with a lot
of heat, which they really do. But better correct than cold.

I'll put a note in the amd64-hardware FAQ as to this aspect.
The board is still in my pile of "spare" hardware. Bad thing is that swapping
will require messing around with some heat-conductor paste. Which I do not have
on stock.
But otherwise, if Alan thinks things should have improved on the board I could
organise a test failry simple. But like he says:

| Oh, dear.  I also have one dual Opteron using a VIA chipset.  I think
| it's the same board.  It's sitting in the corner turned off.  It was
| completely unreliable running as a dual.  With SMP disabled I did not,
| however, have problems.  In a nutshell, given current priorities, it
| wasn't worth my time to debug.

MSI has its european office here in Eindhoven, and they even are a customer of
my former company. Unfortunatly are the only people I know management Bob's, so
hinting about anything is not going to be of much use. And then perhaps even
they could not fix the board, since it could be the chipset or combination of
parts with the chipset.

--WjW