Re: Oracle ASM and High Availability

From: Seth Miller <sethmiller.sm_at_gmail.com>
Date: Wed, 24 Feb 2016 09:34:29 -0600
Message-ID: <CAEueRAUXT_A3zgmSWS=LaGO_jgts1RHCpQobBnpOD_5WgOE4Pw_at_mail.gmail.com>



Sanjay,

Although this is probably not the answer you are looking for but, at the point you are describing the best way is to engage Oracle Support. There are a number of good articles on using kfed/kfod and manual disk repair, and the ASM SMEs at Oracle are really smart and willing to share their information but, once you start messing with the disk header, you are deep in the weeds and certainly unsupported without the direct involvement of Oracle Support.

Seth Miller

On Wed, Feb 24, 2016 at 9:27 AM, Sanjay Mishra <smishra_97_at_yahoo.com> wrote:

> Seth
>
> If header is not repairable then what is the best way to bring environment
> back with less downtime. Scenario is Disk group is 20Tb and Lost disk is
> 512G.
>
> TIA
> Sanjay
>
>
> On Tuesday, February 23, 2016 3:29 PM, Seth Miller <
> sethmiller.sm_at_gmail.com> wrote:
>
>
> Sanjay,
>
> There are ways to fix corrupt ASM disk headers (other than restoring a
> backup of the header) so no, it is not true that all the data would be lost
> if external redundancy disk header is lost.
>
> However, this scenario is exactly why you would not use external
> redundancy of you are unable to protect the disks and the headers at the
> storage layer.
>
> There are scenarios where you would have separate disk groups for each
> database but these would generally not be for protecting data, but
> segregating it for purposes of using storage snapshots.
>
>
> Seth Miller
>
> On Tue, Feb 23, 2016 at 10:48 AM, Sanjay Mishra <
> dmarc-noreply_at_freelists.org> wrote:
>
> Hi
>
> Need some expert opinion on Oracle ASM best practices and high
> availability. It is recommended to have minimum DG using ASM like DATA/FRA
> for multiple database on the Cluster. Now I had seen some scenario, where
> one disk header on External Redundancy DG causes to restore all Database
> and hence Major failure. Does this not make sense to create multiple DG
> like DATA for each Database and FRA can be shared ?
>
> Please share your opinion and just for reference that Disk header in such
> scenario was not recoverable from backup header.
>
> TIA
> Sanjay
>
>
>
>
>

--
http://www.freelists.org/webpage/oracle-l
Received on Wed Feb 24 2016 - 16:34:29 CET

Original text of this message