DBAsupport.com Forums - Powered by vBulletin
Results 1 to 6 of 6

Thread: checkpoint

  1. #1
    Join Date
    Jun 2002
    Posts
    62
    Does Checkpoint occur at log switch?
    Regards,
    Jack

  2. #2
    Join Date
    Sep 2001
    Location
    Düsseldorf, Germany.
    Posts
    588
    According to oracle checkpotints occures primiraily when (While database working normally)

    1. When a logswitch occures
    2. based on the parameter value LOG_CHECKPOINT_INTERVAL
    3. based on the parameter value LOG_CHECKPOINT_TIMEOUT


    HTH

    Sameer

  3. #3
    Join Date
    Jun 2000
    Location
    Madrid, Spain
    Posts
    7,447
    4. FAST_START_IO_TARGET
    5. LOG_FILE_SIZE

  4. #4
    Join Date
    Jun 2002
    Posts
    62
    Pando,

    Could you say a bit more abt LOG_FILE_SIZE? I read in Sybex book that checkpoint occurs when 90% of smallest logfile is full.
    Regards,
    Jack

  5. #5
    Join Date
    Jun 2000
    Location
    Madrid, Spain
    Posts
    7,447
    well incremental checkpoints is based on 2, 3, 4 and 5 we described, actually it is based on the minimum value of those 4 events

    if you look v$instance_recovery you will a couple of columns which says ACTUAL_REDO_BLKS and TARGET_REDO_BLKS and ACTUAL_REDO_BLKS is always less than TARGET_REDO_BLKS, when it is bigger checkpoints occurs, the TARGET_REDO_BLKS is the minimum of 4 events, if let´s say I have


    LOG_CHECKPOINT_INTERVAL very BIG say 1000000
    LOG_CHECKPOINT_TIMEOUT big as well such as 1800
    and
    FAST_START_IO_TARGET = 0

    then you will see different values in v$instance_recovery but the TARGET will be LOG_FILE_SIZE_REDO_BLKS since most probably (I am assumming here) it will have the smallest value compare with other 3 (well actually FAST_START_IO_TARGET will show NULL since it´s disabled by setting it to 0) and in this case incremental checkpoint will base its redo block address calculation on log file size. If you take this TARGET_REDO_BLKS value and do some calculation you will actually see it is 90% of the smallest redo log file size (here the Sybex book is taking about)

    If you mix three parameters (specially the time interval, I cant track it!) and the file size then it´s harder to track it by looking v$instance_recovery but at the end of the day it is Oracle that is doing the calculations




  6. #6
    Join Date
    Jun 2002
    Posts
    62
    Pando, thanks a lot for your promt response.
    Regards,
    Jack

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