Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Mailing Lists -> Oracle-L -> OTRE: DBWR slaves in NT???

OTRE: DBWR slaves in NT???

From: Mohan, Ross <>
Date: Thu, 01 Mar 2001 13:42:52 -0800
Message-ID: <>

an old trick of mine. Embed something unsavory, wicked, or delirious
into a
"status report".
<FONT face=Arial color=#0000ff

If you
get called in on the carpet, at least you know your superiors are really
<FONT face=Arial color=#0000ff

size=2>reviewing your work....if not.....<EG>

  <FONT face="Times New Roman"
  size=2>-----Original Message-----From: Kevin Kostyszyn   []Sent: Thursday, March 01, 2001 3:56   PMTo: Multiple recipients of list ORACLE-LSubject: RE:   DBWR slaves in NT???
  <FONT face=Arial color=#0000ff
  size=2>Ummm, Ross, was that a Freudian slip?  db_block what?    hehehh   

    <FONT face=Tahoma
    size=2>-----Original Message-----From:     []On Behalf Of Mohan, RossSent:     Thursday, March 01, 2001 3:41 PMTo: Multiple recipients of list     ORACLE-LSubject: RE: DBWR slaves in NT???     I agree. I focussed my answer almost exclusively     on Raghu's last line. "My import is killing     me."
    In my hazily-recalled modicum of experience, this     probably has a simple solution.
    Or, hell, maybe I am wrong. It *could* require     doing something slick and esoteric, like calling in     Rivest from RSA to recalculate the prime number seed     for the db_block_hash_fuckits parameter and then     rewriting the kxscvngr subroutine in assembler     to more aggressively clump and scavenge the dirty     blocks in mode "1016" from the warm end of the     buffer cache.
    But, I doubt it.

    Remember, the DBWRs write dirty buffers (data in memory that     has been modified from it's counterpart in a     datafile) from the buffer cache (memory) to the     datafiles (disk).  In O7, the DBWR I/O clump size, or how many     dirty blocks could be written at once (also     sometimes referred to as the DBWR's "internal write     batch size"), was calculated as:
    size=2>DB_FILE_SIMULTANEOUS_WRITES*DB_FILES/2.     In O8, the DBWR I/O clump size is static, and is different     from platform to platform.  Thus, for Oracle8,     the advise to change DB_FILES is bogus.     HTH!


    System/Database Administrator <FONT     size=2>                 Quad/Tech International, Sussex, WI USA     Disclaimer:  After taking the Oracle8 Perf Tuning     class, I think I know more than I do.  Don't     blame me for your (in)actions based on my opinion!     -----Original Message----- Sent:
    Thursday, March 01, 2001 12:27 To: Multiple     recipients of list ORACLE-L
    not a dbwriter problem. on a
    "single-disk wonder", your best bet is to DUMP all     indexes on loadable tables, then load your data,     then rebuild the indexes. of
    course, that might take just as long. any possiblity     of getting more disk? like, several? a raid     controller? mit gluck!
    -----Original Message----- Sent:
    Thursday, March 01, 2001 11:26 AM To: Multiple     recipients of list ORACLE-L
    Hi Friends I have toad tool for my
    NT database DBWR Average Scan Depth 1024      **Number of DB_Files too high?? I got doubt due to     from v$waitstat I got data block waits some time 8000 and     some times higher number. How can I increase the DBWR slaves     on NT?? although I set in development box, How can I     see as background process really working for     me?? I have my datafile and indexes same on E disk     drive, My hit ratios are okay,
    My import is killing me. Any help
    appreciated. Thanks <FONT
    size=2>Raghu. <FONT
    Get Your Private, Free E-mail from MSN Hotmail at <A     target=_blank
    href=""> <FONT     size=2>-- Please see the official ORACLE-L FAQ: <A     target=_blank href="">

    This message has been scanned for viruses with Trend Micro's     Interscan VirusWall. <FONT
    size=2>-- Please see the official ORACLE-L FAQ: <A     target=_blank href="">

Received on Thu Mar 01 2001 - 15:42:52 CST

Original text of this message