Re: One control file/redo member with ASM ok?

From: Ronald Rood <devnull_at_ronr.nl>
Date: Mon, 17 Jan 2005 05:49:14 +0100
Message-ID: <0001HW.BE11025A00316047F03055B0@news.individual.net>

On Mon, 17 Jan 2005 03:13:15 +0100, Connor McDonald wrote (in article <41EB1F3B.A1F_at_yahoo.com>):

> DA Morgan wrote:
>> 
>> Thomas Kyte wrote:
>>

>>> In article <beolu01umcqbaft3pmdmu5isst1c4as0bg_at_4ax.com>, Sybrand Bakker
>>> says...
>>>

>>>> On 16 Jan 2005 10:25:25 -0800, domenicg_at_hotmail.com (Domenic) wrote:
>>>> 
>>>> 
>>>>> Is it acceptable to use one control file and one redo log member only
>>>>> (per group) on a database using ASM, where ASM is doing the mirroring?
>>>>> What's Oracle's recommendation/opinion on this?  Wouldn't two control
>>>>> files and two redo members per group be overkill when ASM is doing the
>>>>> mirroring also?
>>>>> 
>>>>> Domenic.
>>>> 
>>>> Oracle requires 2 control files. It won't start without 2 control
>>>> files.
>>>> Mirroring only protects you against hardware failure.
>>>> It doesn't protect you against someone inadvertently deleting a
>>>> controlfile.
>>>> Also the controlfile is usually only a few Mb's big. It would be very
>>>> strange to try to save a few Mb's and instead increase the risk
>>>> someone blows up your database.
>>>> 
>>>> 
>>>> --
>>>> Sybrand Bakker, Senior Oracle DBA

>>>
>>>
>>> Really?
>>>
>>> SQL> startup pfile=initora9ir2.ora
>>> ORACLE instance started.
>>>
>>> Total System Global Area 269554828 bytes
>>> Fixed Size 451724 bytes
>>> Variable Size 201326592 bytes
>>> Database Buffers 67108864 bytes
>>> Redo Buffers 667648 bytes
>>> Database mounted.
>>> Database opened.
>>> SQL> show parameter control
>>>
>>> NAME TYPE VALUE
>>> ------------------------------------ -----------
>>> ------------------------------
>>> control_file_record_keep_time integer 7
>>> control_files string
>>> /home/ora9ir2/oradata/ora9ir2/
>>> control01.ctl
>>> SQL>
>>>
>>>
>>> (they have the control files mirrored via ASM mirroring)

>> 
>> I have mixed feelings on this one.
>> 
>> While technically you are correct ... advising someone that they can do
>> this invites a very dangerous behaviour. Companies may suffer greviously
>> because someone reads this as Tom Kyte saying it is ok. And even I would
>> gleefully do what you recommend or what I assume you have sanctioned.
>> --
>> Daniel A. Morgan
>> University of Washington
>> damorgan_at_x.washington.edu
>> (replace 'x' with 'u' to respond)
>> 
>> ----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet 
>> News==----
>> http://www.newsfeeds.com The #1 Newsgroup Service in the World! >100,000 
>> Newsgroups
>> ---= East/West-Coast Server Farms - Total Privacy via Encryption =---
> 
> "can" != "should"
> 
>> -)
>

On a filesystem I would certainly use multiple controlfiles devided over several disks. For the online redolog files I also would use at least 2 groups. This question is about ASM. ASM already handles the redundancy and with ASM you have no manual control over the database files. With ASM you could drop an ASM disk and that's what you are protected against. Personally I use OMF, with a datafile location and a recovery area using different disks. In that case you get a copy of a controlfile in both destinations and the same is for the online redolog groups. Doing so will be a good practice and I advice my customers to do so.

With kind regards / met vriendelijke groeten, Ronald

http://ronr.nl/unix-dba
http://homepage.mac.com/ik_zelf/oracle Received on Sun Jan 16 2005 - 22:49:14 CST