[rescue] Sun E450 - 'CPU Fan 2 Failure' but not failed

[Greg Jurman]

On Sun, Nov 10, 2013 at 1:15 AM,  <microcode at zoho.com> wrote:
> On Sun, Nov 10, 2013 at 12:51:08AM -0500, Greg Jurman wrote:
>> Hi,
>>
>> I just acquired a E450 that is running fine except for the
>> environmental monitor constantly throwing that CPU Fan 2 has failed.
>> Solaris is refusing to boot with this error even though CPUs are
>> properly cooled.
>
> I don't have any from the E-series so I don't know for sure, but I have
> other Sun boxes and they will boot with various hardware faults.
>
> The CPU fan may be below the safe range even though you can't tell by
> looking or taking measurements. Or a wire may be disconnected somewhere so
> the obp can't get a valid reading.

So to see if there was a fault I spliced the fan connections around
and swaped the wiring of Fan 1 with Fan 2. If it was the fan, the
fault should of moved and it didn't. The fans don't send tach like a
normal CPU fan, but rotor locked/stall (high if stalled, low if
spinning). Grounding that signal line directly also didn't work.
O-scope didn't show any problems with voltages or show any signal
anomalies.

>
> Is this your first/only Sun box? I have not heard of Solaris refusing to
> boot on Sun hardware because of obp faults. Could you say a little more
> about how you are trying to boot and what message you get?

This is currently my only Sun box. I netbooted Solaris 10 to do a
clean install. After it started to come up it threw a message like
this:

Warning: CPU Fan Number 2 Failed
Fan Failure, system shutdown.

It then power-cycled itself back to obp.

> From the obp prompt you should be able to type boot and it should boot. If
> not it should give you some message(s).
>
> If you type showenvironment at the obp prompt it should show you all the
> critical voltages and fan speeds and flag problematic parts. This is
> supported on the Sun Fire boxes, I don't know what the E-series had but I
> guess it was similar.

It is .env at the obp prompt for me. It does not show speeds as the
fans to not do tach. Everything looks to be in proper operating temp
ranges. The only error it is reporting is 'CPU/Electronics Bay Fan 2:
FAIL'. env-monitor is set to disabled.

>
>> Does anyone know if there a way to bypass this in-order to get Solaris
>> to suck it up and run or a known fix to correct this problem properly
>> (other than replace the main-board)?
>
> I suspect it will just be a fan replacement at the most. You don't want to
> get into buying a new mainboard!

Agreed, I do not want to be buying a new mainboard seeing as this one
works fine EXCEPT for this one problem. I have a fan of very similar
specs on its way. I doubt it'll help as the fault seems to reside
specifically on Fan 2's channel.

> _______________________________________________
> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue