Re: very large RAC database copy to another RAC
- From: Harel Safra <harel.safra@xxxxxxxxx>
- To: dd.yakkali@xxxxxxxxx
- Date: Thu, 20 Jan 2011 11:31:13 +0200
"The problem with RMAN backup ( whether cold or hot) is, the backup copies
are created on a filesystem and we do not have that big space to hold the
backupfiles"
You can duplicate the database with rman from the live database on the other
side, without writing the backups to disk.
Have you thought about storage replication?
Harel Safra
Sent from my phone
On Jan 20, 2011 11:15 AM, "dd yakkali" <dd.yakkali@xxxxxxxxx> wrote:
> Subodh,
>
> thanks for your reply. The problem with RMAN backup ( whether cold or
> hot) is, the backup copies are created on a filesystem and we do not have
> that big space to hold the backupfiles, thats why the datafiles have to be
> copied directly to the target RAC's ASM.
>
> We are in the process of testing some of the speeds to check the
feasibility
> of DBMS_FILE_TRANSFER
>
> Thanks
> Deen
>
> On Thu, Jan 20, 2011 at 12:44 AM, Subodh Deshpande <
> deshpande.subodh@xxxxxxxxx> wrote:
>
>> RMAN requires archive log I suppose..
>>
>>
>>
http://asktom.oracle.com/pls/apex/f?p=100:11:0::::P11_QUESTION_ID:40370408605260
>>
>> think of putting it in archive log
>>
>> the new infrastructure is good so in your case, in one hour how much
gb/tb
>> of data can be backedup and restored, how many NICs and multiple
processess
>> for copying of datafiles..you can have..
>>
>> have you taken any tiral/pilot runs of datatransfer in this new
>> environment...if 48 hours down time window is non-negotiable then use
only
>> proven and known methods as per your comfort level..
>>
>> these are few inputs from my side...subodh
>>
>>
>> On 20 January 2011 13:43, dd yakkali <dd.yakkali@xxxxxxxxx> wrote:
>>
>>> Dear DB admin gurus,
>>>
>>> I need some thoughts on this. We have a very large NOArchivelog database
>>> 11.1.0.7 RAC 20 node RHEL4, 300+T spread on about 50 tablespaces, we
want
>>> copy this to another RAC cluster RHEL5 30 nodes in a different data
center
>>> connected with pretty good network infrastructure between them. They
both
>>> use ASM of course.
>>>
>>> We are probably allowed to be down on the source for about 48 Hours max.
>>>
>>> So our aim is to parallelize(increase throughput by utilizing all the
>>> nodes of both RAC systems ) the copy from one ASM to another ASM
directly.
>>> I want to know what techniques are available to do this. Few techniques
>>> that came up in the discussion are RMAN duplicate database,
>>> dbms_file_transfer via dblink to copy datafiles using transportable
>>> tablespaces technique etc.
>>>
>>> I want to know if any of you out there did dbms_file_transfer between
>>> databases and if you did, can you please share your experiences.
>>>
>>> Any thoughts/pointers are greatly appreciated.
>>>
>>> thanks
>>> Deen
>>>
>>
>>
>>
>> --
>> ==============================
>> DO NOT FORGET TO SMILE TODAY
>> ==============================
>>
Other related posts:
