DBAsupport.com Forums - Powered by vBulletin
Page 3 of 3 FirstFirst 123
Results 21 to 26 of 26

Thread: rupd$ in materialized view logs

  1. #21
    Join Date
    Feb 2007
    Posts
    212
    hmmnn....nice explanation dear....sounds similar to log miner...where u can
    see all the process being done on the database and you can audit which
    table was being modified by what user and report some fraudulent tampering
    of data Ok I try read it and give it a test....hhmmn i hope i wont be
    able to sleep

    Thanks again xxxxxxxxxs

  2. #22
    Join Date
    Sep 2001
    Location
    Makati, Philippines
    Posts
    857
    Quote Originally Posted by PAVB

    Wasn't trying to do that. It's clear you have Jen eating from your hand.
    hmmm... I believe that includes you, and boris, and some of us here.


    Anyway, for the Oracle Streams which boris mentioned, from what I know
    the improvement was more in AQ(Advance Queueing), I remember problems using AQs in 9i/10g
    where in queued process sometimes can't continue when let's say the db server gets down abnormally
    and restarted. Then most of the recomendation was to use Streams specially
    when the version is already in 10g. And from what I've read from Tamil, Streams are not yet stable
    if you are going to use it specifically for Replication.
    ---------------

  3. #23
    Join Date
    Feb 2007
    Posts
    212
    Hi all....yes, I got what u meant now .
    While I am testing our replication set-up I noticed that when the network
    connection is broken(long enough) or the server is down. I got "BROKEN" status in my refresh group, and It does not automatically go back to "NORMAL"
    status even when the connection is already GOOD. So, you have to manually
    refresh it using OEM. and if u have a critical data that need to be replicated
    immediately then u have to check the status often.
    Did you encounter the same problem?
    I thinking of a OS scheduled script the will check every minute if the refresh group is "BROKEN" then it will run the progem like this :


    BEGIN
    DBMS_REFRESH.MAKE(
    name => '"MVADMIN"."GROUP1"',
    list => '',
    next_date => SYSDATE,
    interval => '/*10:Secs*/ sysdate + 10/(60*60*24)',
    implicit_destroy => TRUE,
    lax => FALSE,
    job => 0,
    rollback_seg => NULL,
    push_deferred_rpc => TRUE,
    refresh_after_errors => TRUE,
    purge_option => NULL,
    parallelism => NULL,
    heap_size => NULL);
    END;
    /

    Do you have any idea how to make the script?

    Thanks

  4. #24
    Join Date
    Nov 2006
    Location
    Sofia
    Posts
    630
    Jenn,
    I gave not dealed witrh replication for a long time but I remember that the behaviour you see is intended behaviour. Replication operates like that and there is nothing strange here

    The job is not necessarily be OS level job. You can use dbms_job as well.
    Also, It wold be good if U put new questions in new threads instead of adding to existing one. That's much clearer

    Cheers
    Boris

  5. #25
    Join Date
    Feb 2007
    Posts
    212
    Ah ok thanks honey....
    So I was wrong when I explain to my boss that replication is automated....
    even when network is cut-off and restored or server is down and up .

  6. #26
    Join Date
    Nov 2006
    Location
    Sofia
    Posts
    630
    Well, it's like that
    Replication tries to propagate changes. If that fails, rep tries it again say after 1 sec. If it fails again, it tries after 2 sec.If it fails again, it tries after 4 sec and so on, until a treshold is reached (don't remember but after several tries)
    Then the rep gives up, marks the job broken and never tries again. Here u should manually restart the process when everything is fine again

    Cheers

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


Click Here to Expand Forum to Full Width