Re: Convert 2-node RAC to single instance
- From: Tony Sequeira <tony@xxxxxxxxxxxxxxx>
- To: Oracle List <oracle-l@xxxxxxxxxxxxx>
- Date: Thu, 05 Feb 2009 06:29:28 +0000
Hi Steve,
This was my first reaction, but I was told that node 1 has files
essential to the application, and if it's down, the application won't run.
Begs the question why they went RAC in the first place, but that's the
way it is.
Regards.
steve montgomerie wrote:
> can the standby handle the job? why not swithover to the standby, then
> once things are running smooth on the standby, drop the former
> primary, rebuild on the original primary, add that as a standby and
> then switchover to our shinny new config?
>
> On Wed, Feb 4, 2009 at 3:10 PM, Tony Sequeira <tony@xxxxxxxxxxxxxxx
> <mailto:tony@xxxxxxxxxxxxxxx>> wrote:
>
> Hi,
>
> This is complicated by the Oracle Installation being on the SAN,
> which is due to be retired. They are going to disconnect it
> totally, keep it on standby in case they have issues with the new
> instance.
>
> Guess, I could clone the ORACLE_HOME, but it seems to me to be
> easier to reinstall on local disk, and build a new instance.
> Maybe I'm wrong? But it wouldn't take long to copy the files and
> clone the database once that's done.
>
> I'm for keeping it simple.
>
> Regards.
>
>
> Dan Norris wrote:
>> Sorry for the late follow up on this.
>>
>> I'd probably follow this process to ensure there's an easy fall back:
>> 1. Shutdown 2nd instance
>> 2. set cluster_database=false
>> 3. Run application for long enough period to declare the database
>> performance acceptable on a single node (maybe a week, month, or
>> quarter depending on your business rules to declare "success")
>> 4. Drop extra thread of redo and extra undo tablespace
>> 5. relink binaries to disable RAC
>> 6. Possibly remove database and listener from Clusterware
>> configuration (using srvctl)
>> 7. Possibly consider using your second node for a failover
>> cluster for this new single-instance database you have configured
>> (see
>>
>> http://www.oracle.com/technology/products/database/clusterware/pdf/SI_DB_Failover_11g.pdf)
>>
>> Good catch on the relink--I forgot about that step.
>>
>> Dan
>>
>> bao jiejie wrote:
>>> Dan,
>>>
>>> "There must be a fall back to the original RAC DB if
>>> performance suffers."
>>>
>>> If we changed backup to single node, is that easy to handle this
>>> requirement?
>>> Just change the cluster_database parameter and add undo and redo
>>> logs to the second node?
>>>
>>> And ... Do we need relink oracle binary to off the rac option?
>>> like make -f ins_rdbms.mk <http://ins_rdbms.mk> rac on/off ??
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Sun, Feb 1, 2009 at 6:02 PM, LS Cheng <exriscer@xxxxxxxxx
>>> <mailto:exriscer@xxxxxxxxx>> wrote:
>>>
>>> No, that is a 10g feature so bad luck....
>>>
>>>
>>> --
>>> LSC
>>>
>>>
>>>
>>> On Sat, Jan 31, 2009 at 10:37 PM, Tony Sequeira
>>> <tony@xxxxxxxxxxxxxxx <mailto:tony@xxxxxxxxxxxxxxx>> wrote:
>>>
>>> Guillermo Alan Bort wrote:
>>> > Consider using RMAN's switch to copy capability to
>>> minimize downtime
>>> > Alan Bort
>>> > Oracle Certified Associate
>>> >
>>> Is this available at the database level in 9iR2?
>>>
>>> --
>>> S. Anthony Sequeira
>>> ++
>>> I dote on his very absence.
>>> -- William Shakespeare, "The Merchant of
>>> Venice"
>>> ++
>>>
>>> --
>>> //www.freelists.org/webpage/oracle-l
>>>
>>>
>>>
>>>
>>>
>>>
>>> --
>>>
>>> ("'-''-/").___..--''"'-._
>>> '7_ 7 ) '_. ( ).'-.__.')
>>> (_Y_.) ._ ) '._ '. ''-..-'
>>> _.'--'_..-_/ /--'_.' ,'
>>> (il),-'' (li),' ((!.-'
>>>
>>> Best regards，
>>> Yours sincerely House
>>> baojiejie@xxxxxxxxx <mailto:baojiejie@xxxxxxxxx>
>>>
>>>
>
>
Other related posts:
