Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> RE: ASM mirroring vs SAM mirroring

RE: ASM mirroring vs SAM mirroring

From: Tony van Esch <tony.van.esch_at_xs4all.nl>
Date: Fri, 27 Jul 2007 20:51:54 +0200
Message-Id: <200707271852.l6RIpt0R072820@smtp-vbr12.xs4all.nl>

Hi,

The storageboxes are in the same 'SAN', so ASM doesn't have to worry about that. From the ASM perpective it just gets presented disks which are provisioned in the appropriate Failure group. Also 1Km is not really a big problem for networks (I think you could 'stretch' it to 100Km, before latency would be an issue).

Regards,
Tony van Esch

-----Oorspronkelijk bericht-----
Van: Hemant K Chitale [mailto:hkchital_at_singnet.com.sg] Verzonden: vrijdag 27 juli 2007 16:03
Aan: oracle-l_at_freelists.org
CC: p.mclarty_at_cqu.edu.au; Tony van Esch
Onderwerp: RE: ASM mirroring vs SAM mirroring

Not having done ASM but how does ASM do mirroring across a geographical ("stretched")
cluster when the two storages are seperated by a significant distance ? I know that CA mirrors at the storage level but how does ASM read the disks (devices) from the two storage units together ? What network does it use to replicate between the two failure groups ? How does it handle latency ?

At 08:36 PM Friday, Peter McLarty wrote:
>Hi Tony
>
>Your information is much appreciated.
>
> From all the HP Papers it was never clear about such stuff, I guess
> you had to have the presentation with them.
>
>
>________________________________
>
>From: Tony van Esch [mailto:tvesch_at_xs4all.nl]
>Sent: Fri 27/07/2007 8:12 PM
>To: Peter McLarty
>Cc: oracle-l_at_freelists.org
>Subject: Re: ASM mirroring vs SAM mirroring
>
>
>
>Hi,
>
>we have a similar config running (DUAL HP EVA, mirroring and RAC) and
>asked the suppliers (HP & Oracle) what would be a certified solution.
>In the end ASM was the only viable solution.
>
>
>1> mirroring on SAN level with EVA is called 'Continuous Access'. You
>1> only
>get presented the primary LUN's, but not the copy. The copy is NOT
>presented to the racnodes. So if the storagebox/site with the primary
>LUN's fails, you lose your disks and your database is gone and you have
>downtime. Not really flexible. the mirror woulf have to be presented to
>the racnodes to get things up & running.
>
>2> Mirroring with ASM (host-based mirroring). Is this case the primary
>2> and
>the copy are both presented to the racnodes and placed inside the
>correct failuregroups (FG1=site1/storagebox1 and
>FG2=site2/storagebox2). If one storagebox/site fails, only one
>failuregoup is lost, but the database will still be available.
>

Hemant K Chitale
http://hemantoracledba.blogspot.com

--
http://www.freelists.org/webpage/oracle-l
Received on Fri Jul 27 2007 - 13:51:54 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US