[Beowulf] GPFS on Linux (x86)

Joe Landman landman at scalableinformatics.com
Thu Sep 14 16:31:27 PDT 2006


Craig Tierney wrote:

[...]

> These are the things that I like to see discussed.  I know that Los
> Alamos has a large investment in Panasas, because now it 'just works'.
> But it didn't always 'just work'.  One issue with Panasas is that it can
> be quite expensive  (which is relative) and you have to use their
> hardware.  Getting it to work was probably much easier because it uses
> their hardware.

This is part of their value proposition.  It is quite similar to other
vendors of proprietary disk products.

> However, what if you know you need better access time or more reliable
> better disks (SCSI is still more reliable than SATA in most cases) and

Hmmm.  I have asked this question many times, and all I ever get is
"look at the MTBFs", which actually don't really tell you very much of
value.

CERN did some historical analysis here
http://lpsc.in2p3.fr/cdsagenda/askArchive.php?base=agenda&categ=a0589&id=a0589s1t0/transparencies
 and have a look at page 6, where they don't really see what you state.

As the disks are made on the same process lines, the reliability ought
not to be part of the manufacturing process.  If there is a hyperplane
dividing SCSI and SATA/IDE, it should be evidenced in real reliability
data (e.g. *not* MTBFs which are statistical estimates based upon
specific assumptions of use pattern, but actual measurements ... ala CERN).

> want to invest in 10k or 15k SCSI disks?  What if you care more about a
> scalable metadata engine than just raw bandwidth (or vise-versa)?  What
> if you want to run the Panasas traffic over a higher performing network
> or avoid TCP/IP because of its overhead?  Panasas isn't as flexible as
> some of the software based solutions.  However, maintaining a system
> like this is probably much easier than the software based solutions.

If someone would be so kind as to help me find *real* data that
demonstrates higher SATA/IDE failure rates as compared to SCSI, I would
most appreciate it.

Joe

> 
> Craig

-- 
Joseph Landman, Ph.D
Founder and CEO
Scalable Informatics LLC,
email: landman at scalableinformatics.com
web  : http://www.scalableinformatics.com
phone: +1 734 786 8423
fax  : +1 734 786 8452 or +1 866 888 3112
cell : +1 734 612 4615



More information about the Beowulf mailing list