Comments on: But it’s in the manual! http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/ Just another Oracle weblog Wed, 19 Jun 2013 11:02:03 +0000 hourly 1 http://wordpress.com/ By: Book Review: Oracle Database 11g Performance Tuning Recipes « Charles Hooper's Oracle Notes http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-41676 Sat, 10 Sep 2011 21:13:49 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-41676 [...] Tuning Manual (with a small, but notable correction), but that SQL statement potentially produces misleading information. (pages [...]

]]> By: Jonathan Lewis http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-30610 Fri, 21 Mar 2008 06:41:48 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-30610 Alan,
From 9i onwards, you might want to look at v$segstat rather than v$bh, it’s a more direct representation of the information you want.

You might want to look at what I’ve done in this post to avoid using tables for holding intermediate results – you may be able to use a /*+ no_merge */ union all view with a delay built in to do the job of your two scratch tables.

]]> By: Alan Kendall http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-30605 Thu, 20 Mar 2008 23:23:06 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-30605 By taking a snapshot of v$bh, you get what is being read into memory at this moment. So v$bh and dba_objects can be used to identify what is in memory, and what is doing disk reads. This can be very handy to see what is running at this moment. In the following example the system is almost idle in I/O, reading in only 68 blocks.


SQL > @v$bh2

BUFFERS_READ_INTO_MEMORY OBJECT_IN_MEMORY
------------------------ ----------------------------------------
                      2 IDX_LINKAGES_RELATION_NAME
                      66 ALAN9
------------------------
                      68

COLUMN OBJECT_in_memory FORMAT A40
COLUMN NUMBER_OF_BLOCKS FORMAT 999,999,999,999
column buffers_read_into_memory format 9999999999

break on report
compute sum of buffers_read_into_memory on report

set termout off
drop table alan9;
drop table alan10;

create table alan9 nologging as
select OBJD,FILE#,BLOCK# from v$bh;

create table alan10 nologging as
select OBJD,FILE#,BLOCK# from v$bh
minus
select OBJD,FILE#,BLOCK# from alan9;

set termout on

SELECT COUNT(*) buffers_read_into_memory,
o.OBJECT_NAME Object_in_Memory
    FROM DBA_OBJECTS o, alan10 bh
    WHERE o.DATA_OBJECT_ID = bh.OBJD
      AND o.OWNER        != 'SYS'
    GROUP BY o.OBJECT_NAME
    having count(*)>0
    ORDER BY COUNT(*);

set termout off
drop table alan9;
drop table alan10;
set termout on

]]> By: Saibabu Devabhaktuni http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-22578 Tue, 16 Oct 2007 08:17:23 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-22578 Jonathan,

_db_block_max_cr_dba is literally killing our database performance, too much CPU is being burned to identify the right buffer , waiting on cache buffer chains latch. We are on Oracle 8i. Even though the default value of _db_block_max_cr_dba is 6, it is not going up after 8 in 10g and in Oracle 8i it is limiting at around 30. How do we get rid of the cloned buffers, setting SGA to a lower value would help but it causes physical I/O to go up. Appreciate your thoughts on this.

Thanks,
Sai.

]]> By: Cache Buffers Chains and Latch Spelunking : Ardent Performance Computing http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-20666 Fri, 14 Sep 2007 14:59:04 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-20666 [...] But it’s in the manual! [...]

]]> By: Jonathan Lewis http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-73 Sat, 04 Nov 2006 16:24:44 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-73 Following up a comment to the Clarity, Clarity, Clarity post – I might also point out that:
(a) I would normally swap the order of the objects in the where clause to indicate the expected order of execution, and
(b) I don’t normally write keywords in capitals and right-align them.

]]> By: Christo Kutrovsky http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-64 Fri, 03 Nov 2006 21:11:11 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-64 Jonathan,

Your post reminded me on something I’ve developed a while ago. So I made a post about it.

Basically same idea, but enable it for RAC + some cache fusion efficiency numbers that you can extract:

http://www.pythian.com/blogs/282/oracle-rac-cache-fusion-efficiency-a-buffer-cache-analysis-for-rac

]]> By: Jonathan Lewis http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-56 Fri, 03 Nov 2006 08:18:22 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-56 Mladen,
I think the book by Dan Tow is worth reading. I don’t agree 100% with some of his side-comments, but the analytical method he describes is very useful.

]]> By: Mladen Gogala http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-49 Fri, 03 Nov 2006 03:46:25 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-49 Jonathan, thanks for adding a new page to my everyday reading routine.I am looking for articles about the general principles of writing SQL. Do you have any site or article that might help me with that?

]]> By: Doug C http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-47 Fri, 03 Nov 2006 00:52:35 +0000 http://jonathanlewis.wordpress.com/2006/11/02/but-its-in-the-manual/#comment-47 I just want to say.. I am THRILLED you have this new blog. I will read it every day.

]]>