Re: Snapshot Thin Clone of Physical Standby

From: Seth Miller <sethmiller.sm_at_gmail.com>
Date: Sat, 3 Oct 2015 14:29:42 -0500
Message-ID: <CAEueRAXUD-y8ufPp+WCK8NQmPw5RYiq0z-VBJiQCMBviJtgybw_at_mail.gmail.com>

I've done this successfully without the need for recovery. If you cancel managed recovery before taking the snapshot, the standby will do a checkpoint and give you a consistent database.

Seth Miller
On Oct 3, 2015 11:03 AM, "Stefan Knecht" <knecht.stefan_at_gmail.com> wrote:

> Don
>
> When you're trying to open the database with resetlogs, and you're looking
> in the alert log - does it list a redo logfile there? I don't remember the
> exact ORA- message that goes with it, but if you then try to recover the
> database in SQL*PLUS, and specify to apply that redo log file, the next
> alter database open should succeed.
>
> Could you share the alert log entries that are written just after the
> resetlogs command is issued, until the failure is reported?
>
> Stefan
>
>
> On Fri, Oct 2, 2015 at 11:36 PM, Don Seiler <don_at_seiler.us> wrote:
>
>> Greetings listers,
>>
>> We have an all flash array (Pure) and are working on scripts to make
>> snapshot thin clones from it. However we're stuck when it comes to trying
>> to make a standalone clone from a physical standby database.
>>
>> This is Oracle 12.1.0.2.3 64-bit.
>>
>> The normal process with primary/standalone databases is we take a
>> snapshot of the datafiles and online redo logs, mount the cloned disks and
>> it'll do crash recovery from the cloned online redo logs. This works fine
>> for us. If this is a same-host clone then we run ALTER DATABASE RENAME to
>> rename the files first. The flash array takes a crash-consistent snapshot,
>> so no BACKUP mode is needed, it just does crash recovery from the cloned
>> redo logs. It's very nice.
>>
>> However, when we try this method from the clone of the physical standby,
>> we get an error that we can't rename one of the standby redo logs:
>>
>> ORA-01511: error in renaming log/data files
>> ORA-00261: log 52 of thread 1 is being archived or modified
>>
>> So we go to plan B, which is to try to create a new controlfile. The new
>> CREATE CONTROLFILE script doesn't specify the standby redo logs.
>>
>> When we try to recover or open the database, we get the error that file 1
>> is inconsistent due to failed media recovery session:
>>
>> SQL> alter database open resetlogs;
>> alter database open resetlogs
>> *
>> ERROR at line 1:
>> ORA-01196: file 1 is inconsistent due to a failed media recovery session
>> ORA-01110: data file 1: '/file/name/here.dbf'
>>
>> Every variation of this that we've tried fails similarly.
>>
>> Have any of you fine folks tried something similar and succeeded?
>>
>> Don.
>>
>> --
>> Don Seiler
>> http://www.seiler.us
>>
>
>

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

Received on Sat Oct 03 2015 - 21:29:42 CEST