Re: ASM and disk partitions offset

From: goran bogdanovic <goran00_at_gmail.com>
Date: Sun, 4 Mar 2012 13:07:37 +0100
Message-ID: <CAGyPXK4ZY7JP3ECf+Psmh4Y+u0x7Ozy+MegFWb43Gfcsw2Fg_A_at_mail.gmail.com>

summarized:
misalignment may cause that physical I/O subsystem must execute more single-block I/O calls than really necessary.

in order to eliminate possible performance penalty, it is better to align Oracle blocks upon hardware device boundaries which can eliminate block crossings completely by introducing an offset whose size is calculated to ensure that Oracle block boundaries begin and end on stripe boundaries.

cheers,
goran

On Sun, Mar 4, 2012 at 12:57 PM, goran bogdanovic <goran00_at_gmail.com> wrote:

> Hi Dimitre,
>
> why (in general) offset i.e. aligning is relevant (or could be relevant)
> has been well explained by your fellow countryman ;-) Christo:
>
> http://www.pythian.com/news/411/aligning-asm-disks-on-linux/
>
> cheers,
> goran
>
>
> On Fri, Feb 17, 2012 at 12:21 AM, Radoulov, Dimitre <cichomitiko_at_gmail.com
> > wrote:
>
>> Thanks again!
>>
>> >> Cynical? Moi?
>>
>> Nooo :):):)
>>
>>
>> Best regards
>> Dimitre
>>
>> On 16/02/2012 23:32, Niall Litchfield wrote:
>> > Paul Drake points out there's a later version of the vmware article I
>> > reference at
>> > http://www.vmware.com/pdf/Perf_Best_Practices_vSphere4.1.pdf (it says
>> > the same thing as far as partition alignment goes though)
>> >
>> > On Thu, Feb 16, 2012 at 6:58 PM, Radoulov, Dimitre
>> > <cichomitiko_at_gmail.com <mailto:cichomitiko_at_gmail.com>> wrote:
>> >
>> >
>> > So thanks again Niall,
>> > now I'm reading the articles you mentioned.
>> >
>> >
>> > --
>> > Niall Litchfield
>> > Oracle DBA
>> > http://www.orawin.info
>>
>>
>>
>> --
>> http://www.freelists.org/webpage/oracle-l
>>
>>
>>
>

--
http://www.freelists.org/webpage/oracle-l

Received on Sun Mar 04 2012 - 06:07:37 CST