Re: Huge Table Insert - take long time?

Thanks Mark! But, I don't think our exadata DBA team will go for that in production. Can you say very decentralized?

Thanks though!

On Thu, Jul 26, 2018 at 12:11 PM Mark J. Bobak <mark_at_bobak.net> wrote:

> Also, just a thought:
>
> For the duration of this load, disable FORCE LOGGING, then re-enable it
> when load completes, and catch up the standby with an RMAN Incremental
> backup?
>
> -Mark
>
> On Thu, Jul 26, 2018 at 10:30 AM, Mladen Gogala <gogala.mladen_at_gmail.com>
> wrote:
>
>> Hi Woody,
>>
>> Yes it does avoid UNDO, because it extends the table above the high-water
>> mark and inserts the data into empty blocks. Empty block do not contain any
>> information that would need to be saved in the UNDO segments. That is the
>> principal benefit of doing /*+ APPEND */. You may want to try using
>> DBMS_PARALLEL_EXECUTE instead of the /*+ PARALLEL */ hint. My experiences
>> with DBMS_PARALLEL_EXECUTE are quite good.
>>
>> Regards
>>
>>
>>
>> On 07/23/2018 08:32 PM, Woody McKay wrote:
>>
>>>
>>> Didn't realize that about APPEND avoiding undo even if FORCE LOGGING.
>>> Thanks
>>>
>>>
>> --
>> Mladen Gogala
>> Database Consultant
>> Tel: (347) 321-1217
>>
>> --
>> http://www.freelists.org/webpage/oracle-l
>>
>>
>>
>

-- 
Sincerely,

WoodyMcKay

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