Re: Can I tryst in "physical ... direct" figures?

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Fri, 24 Sep 2004 07:01:07 +0100
Message-ID: <017a01c4a1fb$e5b4c870$6702a8c0@Primary>

As Terry points out, there could be a lot of sorting going on. Your figure would not be unreasonable (for example) in a data warehouse that was doing lots of 'Create table as select'/'report table'.

To increase confidence - you could cross-check these figures against the tablespace I/O figures. Direct writes are usually multiblock writes. Direct reads are usually multiblock reads - the tablespace (and file) stats will give you a further indication of what proportion of your reads and writes were direct.

Regards

Jonathan Lewis

http://www.jlcomp.demon.co.uk/faq/ind_faq.html The Co-operative Oracle Users' FAQ

http://www.jlcomp.demon.co.uk/seminar.html Optimising Oracle Seminar - schedule updated Sept 19th

Original Message ----- From: <J.Velikanovs_at_alise.lv> To: <oracle-l_at_freelists.org> Sent: Thursday, September 23, 2004 11:09 PM Subject: Can I tryst in "physical ... direct" figures?

9.2.0.5 EE on Lintel RH AS 2.1
OLTP ~200 connections.
Today a have observed figures below in statspack report

Statistic                                      Total     per Second    per

Trans
--------------------------------- ------------------ --------------

physical reads                            19,327,036          429.5 205.8
physical reads direct                     10,482,393          232.9 111.6
physical writes                           16,465,203          365.9 175.3
physical writes direct                    16,343,723          363.2 174.0

It seams very strange to me that almost all writes and more then 1/2 of reads is direct.
What do you think, can I trust on those figures? Can it be truth that all writes are direct "bypassing the buffer cache, as written in Docs?

PS Report interval 750.02 (mins)
PS I know - SP aggregates live. Don't tell me do not look on it.

Jurijs
+371 9268222 (+2 GMT)

Thank you for teaching me.
http://otn.oracle.com/ocm/jvelikanovs.html

--

http://www.freelists.org/webpage/oracle-l

--

http://www.freelists.org/webpage/oracle-l Received on Fri Sep 24 2004 - 00:56:42 CDT