Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Parallel Query

Re: Parallel Query

From: Shane Murray <smurray_at_bah.com>
Date: 1997/02/04
Message-ID: <32f6b44f.94879047@news>#1/1

I've had bad luck with instance level tuning, so stick to table and hint level. Here's what I would do if I were you:

  1. Pick a large table that you query on regularly.
  2. Issue: alter table X parallel (degree 2);
  3. run an explain plan for a fll table scan query with some group by columns like: select a,b,c from x group by a,b,c
  4. examine plan_table and one of the columns (options?) should have query processes allocated like :Q1234, :Q5678
  5. Do some test and evaluation on changing the degree of parallelism on this table (2 should be good, though, unless you have fast processors). You can set timing on through sqlplus to record the time it takes to retreive the same dataset. The table should not be cached in the buffer cache, though.
  6. Good luck. I have done lots of testing and configuring with the PQO and found we really only benefit with a handful of tables being set parallel.
    >We recently installed 2nd CPU to HPUX 10.01 server. I configured Oracle
    >7.1.6 for Parallel Query. I tried configuring Parallel Query at instance
    >level, table level and query level (using hint). It does not seem like
    >it is working. I do not see any speed improvement on query that suppose
    >to take advantage of parallel query. I can see ora_p000X processes
    >running. I talk to Oracle Support and they said instance level and table
    >level parallel query on 7.1.6 is not reliable and they also want me to
    >install a patch to fix async I/O in HP. Please help.
Received on Tue Feb 04 1997 - 00:00:00 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US