AlphaServer DS15 BCache Errors?

[bennyb]

We have a couple of DS15's that are reporting the following:

SROM V1.0-0 CPU # 00 @ 1000Mhz
SROM program starting
Configuration Error on CPU #0

then further down the POST it lists;

***Error - SROM detected system configuration error

after Sho Config it lists:

Processors
CPU 0 Alpha EV68CB Pass 4.0 1000 Mhz 0MB cache.


Can anyone tell me what would cause a system to not show the CPU's Cache? The systems both come up but are running v e r y s l o w. Im sure due to lack of access to the cache. Any help would be much appreciated.

[Hoff]

A similar and common error across multiple boxes? That's somewhat unusual.

Have these systems been powered down for a long time, exposed to unusual environments (thermal, magnetic or radiation) or otherwise?

Here, I'd reseat everything, reset and reflash the SRM, strip the box down to minimal pieces (to see if some common widget is causing this), check the clock chip (dallas, coin cell, etc) battery (I don't immediately recall which one of these is used in the AlphaServer DS15), and then contact your preferred hardware service organization.

And I'd (for very obvious reasons) suspect bad bcache. This is usually either a motherboard replacement, or sometimes replacing the processor chip.

[bennyb]

They worked perfectly before we moved them to another location. We have reseated all seatable items. We will try the reflash now and see what happens. Thanks!

[bennyb]

These two systems were actually powered off for 4 months. They did test prior to the relocation though. Does having them down for extended period of time cause any known issues?

[Hoff]

If re-flashing and selecting OS_TYPE OpenVMS and re-seating and otherwise resetting the box et al doesn't clear this, then I would look at swapping the batteries.

Some boxes use a coin cell (often a 2032), some use a Dallas chip, and older boxes used a NiCd pack. They all eventually fail, and having the mains disconnected does draw power from the battery.

http://64.223.189.234/node/441

What this particular box uses off-hand, I don't recall. (If you sort that detail out, do let me know and I'll post it at the URL above.)

I would not immediately expect a dead NVRAM battery to affect the processor cache in this fashion, but it will cause some perturbations at power-up. And -- if swapping the battery does work -- it's far cheaper than swapping the motherboard, given that there is apparently no hardware support contract in force here. (The theory here being that trying the somewhat more obscure and unlikely but cheaper fixes first is better... Not that this is the best or most likely fix.)

[bennyb]

Hoff, thank you for all of your help thus far.

We have checked and it is a Dallas ds12887 real time 0434b1 342903 philippines. We have another battery on an ES45 board that is nearly the same, do you think this one would work?

es45 has Dallas12887 0106a2 301466
ds15 has Dallas12887 0434b1 342902

Do the last two numbers have any significance?

[Hoff]

All of what I know about the Dallas, I've posted at that URL.
Specs for the parts are linked from the same page, IIRC.

[Guzzi]

Hi everyone,
I have here a DS15 with the same kind of error, beep code 2-1-2 and 0MB cache found.
This only happens if feature_0 is been used
(jumper at J22 15/16),otherwise the system is totally dead.
Is there a way to onfigure the CPU in a way that the B-cache can be enabled ? probably the MLB eeprom.
(Dallas DS12887 has been changed allready..)

The srm port at J29 gives me:
SROM V1.0-0 CPU # 00 @ 1000 MHz
SROM program starting
Configuration error on CPU #0
Memory sizing in progress
Memory configuration in progress
Memory data test in progress
Memory address test in progress
Memory pattern test in progress
Memory initialization
Loading console

[bennyb]

what is this Feature_0 you talk about?

This only happens if feature_0 is been used
(jumper at J22 15/16),otherwise the system is totally dead.
Is there a way to onfigure the CPU in a way that the B-cache can be enabled ? probably the MLB eeprom.

what would someone need to do this for?

[Guzzi]

FTR0 is a MLB EEPROM bypass to power-on the CPU,i need to do this to get into the srm
prompt.I did a firmware update this morning
but this did not help,what else can i do to fix this board ?