Results 1 to 3 of 3
  1. #1
    Join Date
    Apr 2004
    Posts
    22

    Arrow Unanswered: help: about turning in partitioned db

    OS:linux
    CPU 4
    RAM 8G
    DELL 6650+CX200

    rows of the test table is 143715052

    2 nodes in one db.

    when i executed the command "select count(*)" , i used 'vmstat' to watch the performance. the CPU wait was 50 and the process blocks were 10.

    ****************
    procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
    r b swpd free buff cache si so bi bo in cs us sy id wa
    1 2 6312 6540648 15600 959276 0 0 28 25 9 32 2 3 94 2
    3 3 6312 6534528 15600 959276 0 0 148416 0 1485 1151 20 4 27 48
    6 1 6312 6530896 15608 959268 0 0 162144 113 1524 1242 25 6 23 47
    2 8 6312 6527816 15608 959268 0 0 174464 0 1541 1231 26 5 18 51
    10 3 6312 6523840 15608 959268 0 0 172032 0 1543 1231 27 5 15 52
    1 7 6312 6520120 15616 959260 0 0 176064 12 1554 1285 27 6 14 54
    0 9 6312 6520048 15616 959260 0 0 154240 0 1499 1140 27 5 14 54
    1 14 6312 6519912 15628 959248 0 0 173920 28 1539 1239 26 4 16 53
    2 12 6312 6519200 15628 959248 0 0 182272 0 1565 1222 27 5 15 53
    3 8 6312 6519144 15628 959248 0 0 159616 0 1512 1128 26 4 18 52
    2 6 6312 6519072 15628 959248 0 0 166240 8 1522 1178 27 4 16 54
    2 8 6312 6519072 15628 959248 0 0 165056 0 1506 1151 25 4 19 52
    1 7 6312 6517864 15640 959236 0 0 132096 56 1420 1031 21 4 15 59

    ****************

    how can i do?

    pool english to express my ideas.

    thanks.
    Last edited by huyuhui; 04-28-04 at 02:38.

  2. #2
    Join Date
    Apr 2004
    Posts
    22

    Who can tell me how to do

    CPU wait were nearly 50 and block progress were nearly 12, who can tell me how to turning the performance .
    thanks a lot !!
    **************************************
    Make progress everyday
    Mail: hoo.jimmy@gmail.com
    **************************************

  3. #3
    Join Date
    Feb 2002
    Location
    Germany
    Posts
    141
    Huyuhui,

    it's very difficult to help you with the data you're giving. You just write you've issued "SELECT COUNT(*) ..."
    I would start checking the system:
    - how big are your bufferpools
    - did you update the statistics for the table(s) involved?
    - do you use different disks for logs and for data?
    - and, and, and ...

    Your design
    - do you have an index defined? (DB2 can use it for count(*))
    - how many containers did you define for your table? are these containers spread over many disks? how did you define prefetch
    - and, and, and...

    My suggestion: just google for "DB2 performance optimization", and you'll find lots of tips. Got some concrete question? post it here.

    HTH
    Rodney Krick

Posting Permissions

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