[Beowulf] Re: Purdue Supercomputer

[Alan Louis Scheinine]

I had written:
 > The air-conditioning installment was not finished, the rack doors
 > were open but the lower part had many cables that basically
 > covered the rear part of the nodes.  That area became so hot
 > that the rubber on the power supply cords became soft.
Perry E. Metzger <perry at piermont.com> replied:
> Were the power supplies physically turned off or were these machines
> "soft off"? Modern machines have both switches on the power supplies
> that turn off the whole power supply, and soft on/off buttons. My bet
> is that the power supplies on these were not actually in the "off"
> position.
> 
> Under circumstances like this, I would have left off all the hard
> power switches until the time was ready to bring the cluster up.

I needed to go to the computer room to refresh my memory.
I was referring to Supermicro cases with Supermicro boards,
purchased last year.  I see that these do not have a power
supply switch in back.  I also checked a recently installed rack
of IBM 1U nodes with AMD processors and a "baseboard" control
processor.  These also do not have a switch near the power supply.
In contrast, a Tyan case that we bought several years ago has a
power supply switch.  Dr. Metzger may be mistaken when he makes the
generalization that "Modern machines have both switches on the
power supplies that turn off the whole power supply, and soft
on/off buttons."

I could, of course, just pull the plug to eliminate the source of
heat.  What I wanted to point-out is the following.  The nodes
were in a room that was cool, though not cold.  The doors of the
racks were open.  I am assuming that the heat is being generated
by the IPMI, but I am not 100 percent certain.  Leaving aside the
part blocked by cables, more generally the stagnant air when there
is no fan blowing, given that the cases are packed tightly one above
the other in a rack, causes a heat build-up when in standby.

What comes to mind, figuratively speaking, it is like an old tube radio.
Having one on the table is not a problem, natural convection removes
the heat.  But putting a group of such radios together would be a problem.

Has anyone else seen a similar problem?  (Perhaps not. :-) )

Best regards,
Alan Scheinine

-- 

  Centro di Ricerca, Sviluppo e Studi Superiori in Sardegna
  Center for Advanced Studies, Research, and Development in Sardinia

  Address from June 2008:
  Alan Scheinine
  5010 Mancuso Lane, Apt. 621
  Baton Rouge, LA 70809

  Email: ascheinine at acm.org

  Phone: 070 9250 238  [+39 070 9250 238]
  Fax:   070 9250 216 or 220  [+39 070 9250 216 or +39 070 9250 220]
  Operator at reception: 070 9250 1  [+39 070 9250 1]
  Mobile phone Italy: 347 7990472   [+39 347 7990472]
  Mobile phone USA:   225 288 4176  [+1 225 288 4176]