DBAsupport.com Forums - Powered by vBulletin
Page 1 of 4 123 ... LastLast
Results 1 to 10 of 37

Thread: checkpoint not complete

  1. #1
    Join Date
    May 2001
    Location
    Delhi
    Posts
    340

    Exclamation

    Hi All,

    I am getting error

    Checkpoint not complete.

    My redo config is

    4 Groups
    2 members in each group
    size of redo files 20 MB

    OS Block Size 2048
    log_checkpoin_interval = 300000
    log_checkpoint_timeout = 0

    i understand that checkpint occur only at switching logs if
    log_checkpoint_interval * size of OS block > actual redo lg file size

    and in my case it is like that only, right ?


    so what should i do and let me know what is the harm of this error.

    HELP ME !!!!
    --------------------------
    The Time has come ....

  2. #2
    Join Date
    Aug 2001
    Location
    Waterloo, On
    Posts
    547
    Your database may be experiencing many INSERTs or LONG/RAW activity. You might have to tune your log buffer size.
    Consider a scenario when the redo log buffer is 1/3 full or the user commits. Now LGWR has to write to an available online redolog. If such a scenario occurs during checkpoint, the user has to wait for log switch to occur before checkpoint is complete and next redo log is available.

    You can review V$SYSSTAT:

    SELECT a.VALUE / DECODE(b.VALUE, 0, 1, b.VALUE)
    FROM V$SYSSTAT a, V$SYSSTAT b
    WHERE a.NAME = 'redo size' AND b.NAME = 'user commits';


    This provides the average number redo records per commit. Now you must determine the average number of commits per second and multiply it by the average number redo records per commit (calculated above). This provides the max log buffers to setup.

    You can then monitor the event 'Log file switch (checkpoint incomplete)'.

    so what should i do and let me know what is the harm of this error.
    Performance degrades. Users keep waiting for a log switch because the session cannot wrap into the next log. Wrapping cannot be performed because the checkpoint for that log has not completed.

    Lets C What Gurus Say
    No way I can claim that at present. But the day is not far off.....

    [Edited by Raminder on 11-09-2001 at 01:41 AM]

    Raminder Singh

    Oracle Certified DBA: Oracle 8i, 9i


    Mail me at raminderahluwalia@rediffmail.com.

  3. #3
    Join Date
    May 2001
    Location
    Delhi
    Posts
    340

    checkpoint error


    raminder - i'll what u hv said and post the reply here. guide me then.


    on checking the alert file again i found that though this is ot occuring very frequently, but still i feel that this need attention.

    --------------------------
    The Time has come ....

  4. #4
    Join Date
    May 2001
    Location
    Delhi
    Posts
    340

    further rinfo

    Hi,

    avg. redo per commit is 38833.7869 (i hope it is correct).
    how to find avg. no of commits per second.

    i think i hv to divide this value by time that a redolog file takes to fill.
    like in my case avg. is 11 min. i.e. 660 sec


    so if i m correct avg. no of commits 38833.7869 / 660
    = 25630299.354 bytes (is it ?).

    i hv log_buffers = 1048576


    now what ?
    --------------------------
    The Time has come ....

  5. #5
    Join Date
    Sep 2001
    Location
    Makati, Philippines
    Posts
    857
    your LOG_CHECKPOINT_INTERVAL(300000) is to small for an interval of a redo log size of 20M(20000000). You've forces to have a checkpoint of every 300000 volume of bytes out of 20000000 written to the redo logs from the log buffer. This phase is fast. I suggest you increase your LOG_CHECKPOINT_INTERVAL to 20000001 or 99999999.

  6. #6
    Join Date
    Jun 2000
    Location
    Madrid, Spain
    Posts
    7,447
    well 300000 os block is 60mb so it´s ok
    paste a few lines from alert log which you see this message might help

  7. #7
    Join Date
    Sep 2001
    Location
    Makati, Philippines
    Posts
    857
    Hi Pando,
    So we consider here the O.S. block size which is 2048? Ok.
    Can you please tell me or refer me a url that discusses this matter with emphasis of the consideration of the O.S. block size. thanks a lot.

  8. #8
    Join Date
    Jun 2001
    Location
    Helsinki. Finland
    Posts
    3,938
    300000 is OK altough I prefer 9999999 :-)

    I wonder why he/she use 2K for a block size. I would say that even 4 is small nowardays.


  9. #9
    Join Date
    Dec 2000
    Location
    Ljubljana, Slovenia
    Posts
    4,439
    Originally posted by julian
    I wonder why he/she use 2K for a block size. I would say that even 4 is small nowardays.
    2K is OS block size, which is quite usual nowadays. I belive you had DB block size in mind with your remark, no?

    [Edited by jmodic on 11-09-2001 at 08:29 AM]
    Jurij Modic
    ASCII a stupid question, get a stupid ANSI
    24 hours in a day .... 24 beer in a case .... coincidence?

  10. #10
    Join Date
    Jun 2001
    Location
    Helsinki. Finland
    Posts
    3,938
    Right, I have overlooked this one, you have the eyes of an eagle :-)


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