Forum Stats

  • 3,732,979 Users
  • 2,246,665 Discussions
  • 7,856,452 Comments

Discussions

Suddenly pass huge redo data

MOJIBUL HOQUE
MOJIBUL HOQUE Member Posts: 159 Blue Ribbon

I have a 2 RAC database with data guard . Suddenly huge redo data passed around 60 MB per seconds to standby database through data guard. Bandwidth between primary database and standby database is 100MB. Database was slow  during passing huge redo data. I want to identify actual reason why suddenly huge redo passe. Is it possible ? How ?

Best Regards

Mojibul Hoque

Tagged:
L. Fernigriniuser13328581

Answers

  • John Thorton
    John Thorton Member Posts: 14,493 Silver Crown
    edited November 2019
    MOJIBUL HOQUE wrote:I have a 2 RAC database with data guard . Suddenly huge redo data passed around 60 MB per seconds to standby database through data guard. Bandwidth between primary database and standby database is 100MB. Database was slow during passing huge redo data. I want to identify actual reason why suddenly huge redo passe. Is it possible ? How ?Best RegardsMojibul Hoque

    DBMS_LOGMNR can report content of REDO log file

  • MOJIBUL HOQUE
    MOJIBUL HOQUE Member Posts: 159 Blue Ribbon
    edited November 2019

    But supplemental log is not enable.  If i have enable it, Will it effect on database performance ?

    And Could I analyze data when supplemental log was not enabled ?

    Best Regards

    MOJIBUL HOQUE

  • Kamran Agayev A.
    Kamran Agayev A. Member Posts: 5,520 Bronze Crown
    edited November 2019

    It depends on the volume of the transactions in your database, Check the following blog post regarding how supplemental logging can affect db performance

    https://www.performatune.com/en/supplemental-logging-when-to-use-it-and-when-not/

    Enable it, do some log mining and then disable it afterwards

    BR,

    Kamran Aghayev A.

  • Jonathan Lewis
    Jonathan Lewis Member Posts: 9,599 Gold Crown
    edited November 2019

    Before chasing too much detail, I would ask myself what could cause an unusual volume of redo.

    Is this VERY unusual, how long did it go on for ?

    A simple explanation might be that someone did a "create table as select", or "alter index rebuild" or similar. Is that worth checking ?  A check on dba_objects.created or dba_objects.last_ddl_time MIGHT offer a clue if it showed an object with a timestamp around the time of the redo generation (and the object is a rough match for the volume of redo)

    Regards

    Jonathan Lewis

    L. Fernigriniuser13328581
Sign In or Register to comment.