Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Usenet -> c.d.o.server -> Re: ASM - AIX 5.3 - 10.2 ( RAW or RAID) ?

Re: ASM - AIX 5.3 - 10.2 ( RAW or RAID) ?

From: DA Morgan <>
Date: Wed, 07 Sep 2005 15:37:14 -0700
Message-ID: <1126132584.778036@yasure> wrote:
> Daniel,
> Thank you for the reply !!
> This is what I mean: (Sorry I don't have much of a clue on disk
> technology)
> create diskgroup DG_1 external redundancy disk '/dev/raw/raw1';
> (My understanding is hardware mirroring is possible while building LVGs
> itself, later we can have either file system of choice or make it RAW.
> So hardware mirroring is nothing to do w/ either it is normal file
> system or RAW device. Correct me if i am wrong)
> Here is the Actual situation:
> VCS is implemented for HA on two Solaris servers. And many other
> database instances are sharing these servers already. I have a new
> system/database that is coming and planning on using ASM for the new
> database. So far we are not using any RAW devices for any of the
> databases. But SA can provide me either RAW devices or normal file
> systems (Veritas) for the new database to use ASM. They use veritas
> Volume Manager to manage the disk ( striping and/or mirroring)
> Now the question would be..
> 1. Will there be any problem in a situation of server failure when we
> use ASM for one of the database?
> ( when a server in a cluster fails, VCS exports and imports file
> systems that are associated to a particular database to other server
> and later oracle instance will be started, but if we implement ASM,
> which actually mounts the disks... Wouldn't it be an issue for VCS ?)
> In another words " how ASM works w/ VCS in the event of server failure
> in HA cluster?
> 2. Below is a snippet from ASM Best Practices...
> "In SAN environments, it assumed that the disks identified are
> appropriately zoned and
> LUN masked. Once the disks are identified, they will need to be
> discovered by ASM.
> This requires that the disk devices (Unix filenames) have their
> ownership changed from
> root to oracle. These candidate disks must already have a valid label
> on it, and should
> not be managed (encapsulated) by any other LVM, such as Veritas."
> In our case Verias manages the disk, So how to use ASM?
> Thanks again,
> Madhu

I am not that current on Solaris but assuming Oracle 10g I would use hardware for RAID, use RAW for storage, and use ASM for the load balancing and to create fail-over group(s).

Dump Veritas: It is a waste of money and CPU with 10g.

Daniel A. Morgan
(replace x with u to respond)
Received on Wed Sep 07 2005 - 17:37:14 CDT

Original text of this message