Mulling over MTBF.

Eric Kuhnke eric at fnordsystems.com
Wed Sep 12 11:29:20 PDT 2001


My thoughts on this:

It's always better to leave a hard drive spinning, rather than power it
up/down...  it puts much more stress on the mechanism to spin up and down
constantly.  This could be compared to the way airliners are rated in
"cycles", ie: takeoffs/landings, where all the mechanical stress occurs.
Unlike hard drives, airplanes can't remain in the air all the time :)

Node waking/sleeping doesn't seem to affect much, but it can be somewhat
unreliable waking up again unless you have quality power supplies.  For
regular ATX desktop/tower supplies I prefer Enermax, Herolchi, and
Enlight.  They're all Taiwanese brands.


Eric Kuhnke
eric at fnordsystems.com
www.fnordsystems.com

-----Original Message-----
From: beowulf-admin at beowulf.org [mailto:beowulf-admin at beowulf.org]On
Behalf Of Carlos O'Donell Jr.
Sent: Wednesday, September 12, 2001 6:04 AM
To: beowulf at beowulf.org
Subject: Mulling over MTBF.


Beowulf,

On a recent trend, I have been discussing with various
colleagues the aspect of power usage and power saving
in clusters.

At first, power saving, through node sleeping or drive
spin down, seemed like a good idea.

Though, I am wary about the following effects:

- Does spindown/spinup on common IDE drives effect MTBF?
 (Any drive for the matter)

- Does node sleeping/wakeup cycles effect MTBF for voltage
  supplies on the motherboard? (Or other componenets, through
  relaxation and transients).

I haven't seen any deep discussion about this. Though
I may want to turn my eye towards a few electrical/computer
engineering papers on the subject.

Cheers,
Carlos O'Donell Jr.
-----------------------------
Computer(Engineering/Science)
University of Western Ontario
http://www.uwo.ca
-----------------------------

_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf





More information about the Beowulf mailing list