Prev: "The '/' slice extends beyond HBA cylinder 1023"
Next: Cron problem under solaris 10 zone
From: Oscar del Rio on 20 Dec 2005 10:38
We are getting frequent crashes on a Sun Blade 1500, Solaris 8,
and the following is the last message before every crash

pcf8584: [ID 318190 kern.warning] WARNING: i2c_0 lost arbitration;
Controller = 0x28a321 addr = 0x5c

What does it mean? faulty hardware?
From: Michael Laajanen on 20 Dec 2005 15:38
HI,

Oscar del Rio wrote:
> We are getting frequent crashes on a Sun Blade 1500, Solaris 8,
> and the following is the last message before every crash
>
> pcf8584: [ID 318190 kern.warning] WARNING: i2c_0 lost arbitration;
> Controller = 0x28a321 addr = 0x5c
>
> What does it mean? faulty hardware?
Most likely if the machine worked before and nothing changed.

However there rings a bell in my mind regarding a I2C interface problem,
but I think it was on a Blade 100. Is it a Red or a Silver?

Is Openboot updated,I think there is a Openboot rev from mid 2005 on the
Red.

If it is a Red it could be the same problem as the Blade 100 since it is
a few years old to, that problem was a temp sensor to the CPU and fixed
if I recall right with a patch of some kind, Openboot or Solaris.

How frequent it the problem and do you have the possibility to try S9 or
S10 as a test to see of it is a Solaris that causes it maybe run a test
during a weekend?


/michael
From: Oscar del Rio on 20 Dec 2005 15:55
Michael Laajanen wrote:

> If it is a Red it could be the same problem as the Blade 100 since it is
> a few years old to, that problem was a temp sensor to the CPU and fixed
> if I recall right with a patch of some kind, Openboot or Solaris.

Thanks Michael, it is a red one. I've applied the latest Solaris recommended
patch cluster and will try updating the OB. (No chance to try Sol 9 or 10)

I will also replace the PSU with one that Sun sent us recently as part of a
recall. No idea if it would be related to this error though.

It crashed about 10 times in the past month, twice in the last two days.
From: Michael Laajanen on 20 Dec 2005 18:39
HI,

Oscar del Rio wrote:
> Michael Laajanen wrote:
>
>> If it is a Red it could be the same problem as the Blade 100 since it
>> is a few years old to, that problem was a temp sensor to the CPU and
>> fixed if I recall right with a patch of some kind, Openboot or Solaris.
>
>
> Thanks Michael, it is a red one. I've applied the latest Solaris
> recommended
> patch cluster and will try updating the OB. (No chance to try Sol 9 or 10)
>
> I will also replace the PSU with one that Sun sent us recently as part of a
> recall. No idea if it would be related to this error though.
>
> It crashed about 10 times in the past month, twice in the last two days.
HI, if it the same problem I suspect Sun had a HW timing issue that
popped up after some time when component started to age abit and in
certian temperatures.

And remember always only change ONE thing at a time, if you change two
things at the same time, later trying to back them both out will not get
you back ;)


/michael
 | 
Pages: 1
Prev: "The '/' slice extends beyond HBA cylinder 1023"
Next: Cron problem under solaris 10 zone