Re: ImpDp performance

From: Andrew Kerber <andrew.kerber_at_gmail.com>
Date: Wed, 13 May 2015 09:49:27 -0500
Message-Id: <E248D88C-1EC2-4F9D-9B65-73BB800302AB_at_gmail.com>

That won't really gain you any time. Regardless you will want to gather stats on the new system.

Sent from my iPhone

> On May 13, 2015, at 9:02 AM, sumit Tyagi <dba.tyagisumit_at_gmail.com> wrote:
> 
> Thanks will look into it ... And what about exclude=statistics
> 

>> On 13 May 2015 18:09, "Andrew Kerber" <andrew.kerber_at_gmail.com> wrote:
>> Those all will help. The advantage from nologging is minimal if your redo logs are adequately sized.
>>
>> Sent from my iPad
>>
>> > On May 13, 2015, at 3:20 AM, sumit Tyagi <dba.tyagisumit_at_gmail.com> wrote:
>> >
>> >
>> > Hi oracle-L Team ,
>> >
>> > One of our client has suggested the below to improve the impdp performance . Please suggest on same , Can this help in performance improvement of impdp . I have checked there are several restrictions
>> > on network_link usage .
>> >
>> > 1. Use network_link with impdp instead of export/copy/import
>> > 2. Replace tables but exclude indexes and constraints
>> > 3 Use indexfile to generate the index/constraint creation script and Write a script that will:
>> >
>> > o Create the indexes in parallel (not the parallel=X option but launch multiple create indexes simultaneously since we are on standard edition)
>> > o Increase the sort_area_size for the session before creating the indexes
>> > o Change the constraint create to use novalidate
>> > o Create the indexes with nologging.
>> >
>> >
>> > --
>> >
>> > --
>> > BR
>> > Sumit

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

Received on Wed May 13 2015 - 16:49:27 CEST