HP3000-L Archives

December 2004, Week 1

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Craig Lalley <[log in to unmask]>
Reply To:
Craig Lalley <[log in to unmask]>
Date:
Fri, 3 Dec 2004 09:56:01 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (39 lines)
--- Northern Credit Bureaus Inc <[log in to unmask]> wrote:

> What does it mean when the disk bar reaches 100%? Does it mean it saturates the scsi card
> throughput? Does it mean a certain threshold has been reached in the # of io's/s that an average
> jbod disk should accept?
>

I'll take a stab at this one, and this is only my opinion.

What does it mean when the disk bar reaches 100%?

It depends on what type of disc you have and it means virtually nothing.  The algorithms that
Glance uses (and other performance tools) are based on a queuing theory of first in first out.

With today's high speed raid arrays, caching technolgies etc... that algorithm is no longer
effective.  Remember glance can only see logical I/O, i.e. from the system point of view, it is
collecting data as it goes to and from the SCSI bus.  It know's nothing about hardware mirroring,
cacheing or even raid.

You better off getting the disc performance statistics from the "other" disc side.   I.E. XP512 or
even the MOD 20 has some data that can be looked at.

It is an interesting question though, I started writing a paper on it a couple of years ago, I
never finished it.

-Craig





__________________________________
Do you Yahoo!?
All your favorites on one personal page – Try My Yahoo!
http://my.yahoo.com

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

ATOM RSS1 RSS2