Re: ASM on SAN

From: Kellyn Pedersen <>
Date: Mon, 22 Feb 2010 08:21:54 -0800 (PST)
Message-ID: <>

I'll second that question-  I'm always "I/O bottlenecked" with the amount of data we move daily and any more knowledge I can gleam to help me out in that area is greatly appreciated! :)


  • On Mon, 2/22/10, Martin Bach <> wrote:

From: Martin Bach <> Subject: Re: ASM on SAN
Cc:, "oracle-l" <> Date: Monday, February 22, 2010, 9:17 AM

Hi Greg,

great reply, I second your approach. May I ask a question about the below paragraph please:

On 02/21/2010 10:23 PM, Greg Rahn wrote:
> The other thing is that most of the storage engineers (people that
> design this stuff) or performance engineers (people that perf test it)
> will tell you there is a sweet spot for the LUN size and
> characteristics.  This varies from array to array, but I would seek
> out this information.  It usually has to do with things like which
> drives are on what controllers and primary vs secondary IO paths, etc.

I remember a MOS document "Note 810484.1 "Recommendation for 4MB ASM Allocation Unit (AU) size for Oracle 11g" which seems to have disappeared when I tried to look it up is recommending au_size > 1MB. I found reference to it under the exadata best practices but I wondered when you'd set the AU_SIZE to a non-default value.

Our system here isn't too large, 2.5 TB, disk groups ctrllog VRAID 1 for online redo logs and control files, DATA for data (VRAID 5) and FRA with VRAID 5 as well. All ASM LUNs are 250G and use external redundancy. Unfortunately there is a severe lack of spindles and I would have preferred VRAID 1 for DATA but my comments about IO performance and checksum'd writes weren't listened to.



Martin Bach
OCM 10g

Received on Mon Feb 22 2010 - 10:21:54 CST

Original text of this message