FAQ
Dear List,
I am trying to find out if setting undo_retention to non-zero value will generate more redo log or not. I dummy up some update and insert transactions and ran it in a 9.2 db with undo_retention=0 and undo_retention=900 (the default), I monitored the redo size by querying 'redo size' stat in v$sysstat. And I found out that the amount of redo generated under both undo_retention is about the same. I would have thought that with undo_retention=non-zero value will generate more redo because of the undo entries and these undo will be in the redo stream as well. However, from the test result it doesn't seem to be the case. Is it my understanding of undo is wrong or the test I did is not right?
Josh

Find local movie times and trailers on Yahoo! Movies.

Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org
put 'unsubscribe' in the subject line.

--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Search Discussions

  • Lex de Haan at May 13, 2004 at 12:50 pm
    undo retention does not generate more or less of anything;
    it only influences how fast Oracle will reuse inactive undo segment
    extents...
    under 9i this is always on a "best effort" basis;
    under 10g this can (optionally) be made a guarantee.
    Kind regards,
    Lex.

    visit my website at http://www.naturaljoin.nl

    -----Original Message-----
    From: oracle-l-bounce_at_freelists.org
    On Behalf Of Jos
    Sent: Thursday, May 13, 2004 13:07
    To: oracle-l_at_freelists.org
    Subject: redo log generation and undo_retention

    Dear List,
    I am trying to find out if setting undo_retention to non-zero value will
    generate more redo log or not. I dummy up some update and insert
    transactions and ran it in a 9.2 db with undo_retention=0 and
    undo_retention=900 (the default), I monitored the redo size by querying
    'redo size' stat in v$sysstat. And I found out that the amount of redo
    generated under both undo_retention is about the same. I would have thought
    that with undo_retention=non-zero value will generate more redo because of
    the undo entries and these undo will be in the redo stream as well. However,
    from the test result it doesn't seem to be the case. Is it my understanding
    of undo is wrong or the test I did is not right?
    Josh

    Binary/unsupported file stripped by Ecartis --
    Type: text/x-vcard
    File: Lex de Haan.vcf

    Please see the official ORACLE-L FAQ: http://www.orafaq.com

    To unsubscribe send email to: oracle-l-request_at_freelists.org
    put 'unsubscribe' in the subject line.
    --
    Archives are at http://www.freelists.org/archives/oracle-l/
    FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
  • Daniel W. Fink at May 13, 2004 at 3:35 pm
    Josh,

    Undo entries are always written into the redo stream. This is the only
    way to reconstruct undo/rollback segments in order to perform the
    rollback tasks of recovery. Undo is always 'inserted', never 'updated'
    nor 'deleted' (this is a good way to think about it, though it is not
    100% accurate). When an existing block is used by a new transaction, the
    block is basically renewed. (again, not 100% accurate, but close enough
    for 99% of the issues). The difference in undo_retention settings will
    be how long Oracle attempts to keep undo entries from being overwritten.
    As far as Oracle is concerned, undo entries into a new block or reused
    block are pretty much the same.

    Daniel Fink

    Jos wrote:
    Dear List,
    I am trying to find out if setting undo_retention to non-zero value will generate more redo log or not. I dummy up some update and insert transactions and ran it in a 9.2 db with undo_retention=0 and undo_retention=900 (the default), I monitored the redo size by querying 'redo size' stat in v$sysstat. And I found out that the amount of redo generated under both undo_retention is about the same. I would have thought that with undo_retention=non-zero value will generate more redo because of the undo entries and these undo will be in the redo stream as well. However, from the test result it doesn't seem to be the case. Is it my understanding of undo is wrong or the test I did is not right?
    Josh
    Please see the official ORACLE-L FAQ: http://www.orafaq.com

    To unsubscribe send email to: oracle-l-request_at_freelists.org
    put 'unsubscribe' in the subject line.

    --
    Archives are at http://www.freelists.org/archives/oracle-l/
    FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
    -----------------------------------------------------------------

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
grouporacle-l @
categoriesoracle
postedMay 13, '04 at 10:59a
activeMay 13, '04 at 3:35p
posts3
users3
websiteoracle.com

People

Translate

site design / logo © 2022 Grokbase