[prev in list] [next in list] [prev in thread] [next in thread] 

List:       veritas-vx
Subject:    Re: [Veritas-vx] Calculating Stripe width
From:       "Robert Lucente - MindSpring" <rlucente () pipeline ! com>
Date:       2001-05-27 22:21:42
[Download RAW message or body]

There is a white paper from Sun which states that there is little
gain after 9 stripes. If you are interested, I could email
you the pdf and save you some leg work.

Please note that I am new to Veritas, new to this mailing list, ... .
So request your patience.

Wide Thin Disk Striping

By Bob Larson - Strategic Applications Engineering

Sun BluePrintsT OnLine - October 2000







----- Original Message ----- 
From: "Ryn" <matty91@mindspring.com>
To: <veritas-vx@mailman.eng.auburn.edu>
Sent: Tuesday, May 22, 2001 6:26 PM
Subject: [Veritas-vx] Calculating Stripe width


> Hello folks,
> 
> I am creating a large VxVM  0+1 volume and
> wondered what the
> optimal stripe width would be? We have 24 disks in
> 2 D1000's
> and plan to stripe across 12 disks and mirror to
> the second set of
> 12 disks.
> 
> Is the calculation:
> 
> Stripe Width = PAGE_SIZE * Ncolumn?
> 96k = 8k * 12
> 
> If this calculation is correct, then I would
> create the RAID 0+1 volume
> with a Strip width of 96k I believe.  This should
> write 8k to each disk
> for each full stripe write if I am not mistaken.
> Also, does anyone know
> what the maximum chunk size is for HP-UX and
> Solaris? I am curious
> what the largest chunk of data these OS's can
> write in a single write
> operation.
> 
> Thanks for any info,
> 
> - Ryn
> 
> _______________________________________________
> Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx


[prev in list] [next in list] [prev in thread] [next in thread]