North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: Quick question.

  • From: Alexei Roudnev
  • Date: Sun Aug 01 01:15:46 2004

SuperMicro 6010H and 6010L (excellent 1U servers, I am very sad that they do
not produce them anymore and that
they are not easily available on auctions becausee are too perfect for many
tasks) have 2 P-III slots, can work with 1 emty and 1 filled it, or with 1
good CPU and 1 broken, without problems. I can not say the same about their
modern 1U servers, which are overperformed and over-complicated.

========================================
> Michael Loftis wrote:
> Eh really?  Whenever I've lost a second CPU (primary or secondary)
> the machine was a brick until the secondary CPU was gutted and for
> Piii slotted systems a terminator board was installed in the
> secondary slot.

That was with PIIs, PPros and PIIIs; and not always. When I said "Xeon"
I assumed a P4 Xeon; nobody's sold PIII system for a while. I have seen
number of PII and PIII dual-capable motherboards that work with one CPU
without a processor terminator though, including the very successful
slot-1 440BX and 440GX based workhorses (that power the majority of PII
and PIII Xeons). Since the 7501 Intel chipset, BIOSes are able to detect
processor failures all the time, apparently.
(O/T) that's why I avoid RDRAM like the plague too: does not work
without a terminator in unused slots.

AFAIC, the main annoyance of dual-capable mobos with single processors
has historically been that you needed to install M$ with two processors
if you wanted to upgrade to dual-proc later; this has changed with
hyperthreading, as now even single-processor systems are seen by Windows
as a duallie.


> What motherboard(s) you using that are holding up to failures
> like this?

The last two failures I can remember were an Intel whitebox with
SE7501BR2 motherboard and a Dell mobo also based on the 7501 (can't
remember the P/N) (some Dell boards are close derivatives of the Intel
design).

For hardware I have touched myself recently, I can assure that none of
the following require a processor terminator:

http://indigo.intel.com/mbsg/details.aspx?compID=1915
http://indigo.intel.com/mbsg/details.aspx?compID=1887
Proliant DL380.
I have tested numerous Tyan and Supermicro 7501 and 7505 boards as well
and none required a processor terminator.
I have never seen an AMD board that required terminators either.

I plan to test the following two DDRII-400 7525 boards when released:
http://www.tyan.com/products/html/thunderi7525.html
http://www.supermicro.com/products/motherboard/Xeon800/E7525/X6DA8-G2.cf
m
and they don't require terminators either.

Terminators are a thing of the past; as a matter of fact, in California
and especially in Sacramento they're called governators now.


> My experience has shown PSU and motherboard failures are
> faaaaar more common than CPUs.

It used to be true, but not anymore. A 2.4 GHz Xeon currently _retails_
for $200 something (care to guess what the OEM price for 1,000 is?), it
can't be as reliable as when we used to pay $2,800 for it. Also, there
was a time where fan failure would account for half of the PS failures,
not true anymore either.

Michel.