[rescue] E250 temperatures

Ron Wickersham rjw at alembic.com
Sat Apr 12 16:04:18 CDT 2008

On Fri, 11 Apr 2008, Joshua Boyd wrote:

> I just noticed that the thermal error light was lit on my e250 running
> solaris 10.  A quick login, and prtdiag shows:
> System Temperatures (Celsius):
> ------------------------------
>      CPU0    64    WARNING
>      CPU1    47
>       MB0    41
>       MB1    31
>       PDB    35
>      SCSI    29

just checking my system:

System Temperatures (Celsius):
       CPU0    41
       CPU1    44
        MB0    34
        MB1    29
        PDB    28
       SCSI    25

> Yet, the fans still aren't full speed:
> Fan Bank :
> ----------
> Bank      Speed     Status
>         (0-255)
> ----      -----     ------
> SYS       219        OK

Fan Bank :

Bank      Speed     Status
----      -----     ------
  SYS       171        OK

so i show a slower fan and lower temperatures.   i've been worried that
my room temperature is too high at around 75-degrees F, but have never
gotten a temp warning light in the five years this machine has been running.

this machine is not mounted in a rack, just sitting on the floor, in the
stand-up position and has another E250 on the top of it.

i am running both power supplies and 3 disk drives.

========================= CPUs =========================

                     Run   Ecache   CPU    CPU
Brd  CPU   Module   MHz     MB    Impl.   Mask
---  ---  -------  -----  ------  ------  ----
SYS     0     0      400     2.0   US-II    10.0
SYS     1     1      400     2.0   US-II    10.0

========================= Memory =========================

        Interlv.  Socket   Size
Bank    Group     Name    (MB)  Status
----    -----    ------   ----  ------
   0      none     U0701   128      OK
   0      none     U0801   128      OK
   0      none     U0901   128      OK
   0      none     U1001   128      OK

the two CPUs are the same...are your CPUs different?   this machine
is running Solaris 10 with 12 zones and in some of the zones there
are 20 virtual hosts of apache, and some heavy usage in perl scripts
for bulletin board system called Discus, so the perfmeter frequently
shows both CPUs maxed out and the load average up to ten so the
machine isn't just sitting idle.

when you have the machine apart, i'd check that all three fans are
putting out the same volume of air, feeling the flow with your hand
should be adequate for that.  i could imagine that if one fan were
running slower then the lower airflow could make a difference on the
parts cooled by it.

> What is the real danger region for the E250, and will the light turn off
> on its own if the temp goes down?
> Also, does anyone know any tricks to make it balance the thermal load
> between the two CPUs better?

what kind of computational load are you running?

> Any other tricks?  I would venture to say that the area where the E250
> sits is still generally under 70 degrees.

this is cooler than the external temperature that my machine is running

> The e250 is near the bottom
> of a short rack that doesn't actually have all that much running in it.

that should be the coolest place in the rack

> Below the e250 is a smartups 1400.

is the e250 laying down in the rack-mount orientation or standing up in the
"normal" orientation shown in the product pictures?   and assuming that
you have the side panels on the e250 which make the airflow go the way
intended by the design.

> The only things running above it are
> a SS20, a GigE switch, and a FastE switch (the FC array, T1, and Compaq
> DL360 are off).  On top of the rack is a Ultra 5, and a few feet away is
> an Ultra 80, which doesn't seem to have the same sort of monitoring on
> it.  I may have to get a remote device to monitor the area...

in feeling the e250 there is no warmth from anyside of the side panels,
the airflow keeps the case absolutely at room temperature.   so i can't
imagine that any gear either above or below would affect its temperature 
in any measureable way.


More information about the rescue mailing list