Re: excessive archive log

From: Don Seiler <don_at_seiler.us>
Date: Fri, 7 Dec 2007 10:57:01 -0600
Message-ID: <716f7a630712070857k724facebg58b01fa158bee7b0@mail.gmail.com>

How are you inserting the data? Just because a tablespace is in NOLOGGING mode doesn't mean INSERTs aren't logged, that only applies to direct-path/append INSERTs. UPDATEs will always be logged regardless.

Don.

On Dec 7, 2007 10:22 AM, <ryan_gaffuri_at_comcast.net> wrote:
> We have an ETL database where virtually all activity are on tablespace in nologging mode. 90% of our activity is inserts and 10% is updates of 1 column that is not indexed. We do have indexes on the tables being inserted to. Though not large numbers(2-4 with column length of 1-3). We do not have really wide columns.
> all of this is in noarchive log tablespaces.
> We are getting relatively speaking alot of redo. We are getting more redo than we are getting data generated.
>
> one thing we may be seeing is that we are using advanced queueing and it is enqueued and dequeued constantly. Could this be causing our archive problems?
>
> how do we investigate this? can I use logminer to research this?

-- 
Don Seiler
http://seilerwerks.wordpress.com
ultimate: http://www.mufc.us
--
http://www.freelists.org/webpage/oracle-l

Received on Fri Dec 07 2007 - 10:57:01 CST