<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Tuning Oracle to Make a Query Slower</title>
	<atom:link href="http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/feed/" rel="self" type="application/rss+xml" />
	<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/</link>
	<description>Just another Oracle blog</description>
	<lastBuildDate>Fri, 24 May 2013 02:26:20 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.5.1</generator>
	<item>
		<title>By: DanyC</title>
		<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/#comment-161117</link>
		<dc:creator>DanyC</dc:creator>
		<pubDate>Wed, 31 Oct 2012 23:25:23 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=3760#comment-161117</guid>
		<description><![CDATA[One word buddy: Many thanks!!

Dany]]></description>
		<content:encoded><![CDATA[<p>One word buddy: Many thanks!!</p>
<p>Dany</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: osborne</title>
		<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/#comment-160628</link>
		<dc:creator>osborne</dc:creator>
		<pubDate>Tue, 30 Oct 2012 20:27:23 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=3760#comment-160628</guid>
		<description><![CDATA[&lt;pre&gt;
$ cat table_SIZE.SQL
set lines 155
compute sum of totalsize_megs on report
break on report
col owner for a20
col segment_name for a30
col segment_type for a10
col totalsize_megs for 999,999.9
select * from (
select owner, segment_name, segment_type type,
sum(bytes/1024/1024) as totalsize_megs,
tablespace_name
from dba_segments
where owner like nvl(‘&amp;owner’,owner)
and segment_name like nvl(‘&amp;table_name’,segment_name)
and segment_type like nvl(‘&amp;type’,segment_type)
group by owner, segment_name, tablespace_name, segment_type
order by 4 desc
)
where rownum &lt; 30
order by 4;

$ cat realfreemem.sh
#!/bin/ksh

FREEMEM=`cat /proc/meminfo &#124; grep MemFree: &#124; sed -e &#039;s/ */ /g&#039; &#124; cut -f2 -d&#039; &#039;`
CACHED=`cat /proc/meminfo &#124; grep Cached: &#124; grep -v SwapCached: &#124; sed -e &#039;s/ */ /g&#039; &#124; cut -f2 -d&#039; &#039;`
TOTALMEM=`cat /proc/meminfo &#124; grep MemTotal: &#124; sed -e &#039;s/ */ /g&#039; &#124; cut -f2 -d&#039; &#039;`

((FREEMEM=${FREEMEM}/1024))
((CACHED=${CACHED}/1024))
((TOTALMEM=${TOTALMEM}/1024))

((TOTALFREE=${FREEMEM}+${CACHED}))
((REALFREE=$TOTALFREE*100))
((REALFREEPCT=$REALFREE/$TOTALMEM))

if [ &quot;$1&quot; = &quot;-a&quot; ]; then
echo &#039;Free Memory:&#039; ${FREEMEM}&#039;M&#039;
echo &#039;Cached Memory:&#039; ${CACHED}&#039;M&#039;
echo &#039;Total Free Memory:&#039; ${TOTALFREE}&#039;M&#039;
echo &#039;Total Memory:&#039; ${TOTALMEM}&#039;M&#039;
echo &#039;Percent Memory Free (including cache):&#039; ${REALFREEPCT}&#039;%&#039;
else
echo &#039;Percent Memory Free (including cache):&#039; ${REALFREEPCT}&#039;%&#039;
fi

return $REALFREEPC
&lt;pre&gt;]]></description>
		<content:encoded><![CDATA[<pre>
$ cat table_SIZE.SQL
set lines 155
compute sum of totalsize_megs on report
break on report
col owner for a20
col segment_name for a30
col segment_type for a10
col totalsize_megs for 999,999.9
select * from (
select owner, segment_name, segment_type type,
sum(bytes/1024/1024) as totalsize_megs,
tablespace_name
from dba_segments
where owner like nvl(‘&#038;owner’,owner)
and segment_name like nvl(‘&#038;table_name’,segment_name)
and segment_type like nvl(‘&#038;type’,segment_type)
group by owner, segment_name, tablespace_name, segment_type
order by 4 desc
)
where rownum < 30
order by 4;

$ cat realfreemem.sh
#!/bin/ksh

FREEMEM=`cat /proc/meminfo | grep MemFree: | sed -e 's/ */ /g' | cut -f2 -d' '`
CACHED=`cat /proc/meminfo | grep Cached: | grep -v SwapCached: | sed -e 's/ */ /g' | cut -f2 -d' '`
TOTALMEM=`cat /proc/meminfo | grep MemTotal: | sed -e 's/ */ /g' | cut -f2 -d' '`

((FREEMEM=${FREEMEM}/1024))
((CACHED=${CACHED}/1024))
((TOTALMEM=${TOTALMEM}/1024))

((TOTALFREE=${FREEMEM}+${CACHED}))
((REALFREE=$TOTALFREE*100))
((REALFREEPCT=$REALFREE/$TOTALMEM))

if [ "$1" = "-a" ]; then
echo 'Free Memory:' ${FREEMEM}'M'
echo 'Cached Memory:' ${CACHED}'M'
echo 'Total Free Memory:' ${TOTALFREE}'M'
echo 'Total Memory:' ${TOTALMEM}'M'
echo 'Percent Memory Free (including cache):' ${REALFREEPCT}'%'
else
echo 'Percent Memory Free (including cache):' ${REALFREEPCT}'%'
fi

return $REALFREEPC
<pre></pre>
]]></content:encoded>
	</item>
	<item>
		<title>By: Danyc</title>
		<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/#comment-157344</link>
		<dc:creator>Danyc</dc:creator>
		<pubDate>Tue, 23 Oct 2012 09:18:35 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=3760#comment-157344</guid>
		<description><![CDATA[Hi Kerry,

By any chance, can you show the output of the fss/ realfreemem.sh and table_size scripts?

Many thanks,
Dani]]></description>
		<content:encoded><![CDATA[<p>Hi Kerry,</p>
<p>By any chance, can you show the output of the fss/ realfreemem.sh and table_size scripts?</p>
<p>Many thanks,<br />
Dani</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Alex</title>
		<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/#comment-75403</link>
		<dc:creator>Alex</dc:creator>
		<pubDate>Mon, 05 Dec 2011 11:32:10 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=3760#comment-75403</guid>
		<description><![CDATA[an alternative to flushing caches in linux :  echo 3 &gt; /proc/sys/vm/drop_caches]]></description>
		<content:encoded><![CDATA[<p>an alternative to flushing caches in linux :  echo 3 &gt; /proc/sys/vm/drop_caches</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: osborne</title>
		<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/#comment-74305</link>
		<dc:creator>osborne</dc:creator>
		<pubDate>Wed, 23 Nov 2011 20:52:12 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=3760#comment-74305</guid>
		<description><![CDATA[Gwen,

  Absolutely, if we wanted to make this query run fast, then making the buffer cache bigger would do it. But I was trying to make it go slower ;) - that&#039;s why it was interesting (to me anyway). 

Frits,

  Yes, in 32 bit days there were distinct advantages to having big file system cache in some cases (like yours).]]></description>
		<content:encoded><![CDATA[<p>Gwen,</p>
<p>  Absolutely, if we wanted to make this query run fast, then making the buffer cache bigger would do it. But I was trying to make it go slower ;) &#8211; that&#8217;s why it was interesting (to me anyway). </p>
<p>Frits,</p>
<p>  Yes, in 32 bit days there were distinct advantages to having big file system cache in some cases (like yours).</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Frits Hoogland</title>
		<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/#comment-74300</link>
		<dc:creator>Frits Hoogland</dc:creator>
		<pubDate>Wed, 23 Nov 2011 20:06:56 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=3760#comment-74300</guid>
		<description><![CDATA[More or less exactly what I encountered some 5-7 years ago. Large solaris machine (at that time; 32GB memory), database user-data size 1-2GB. In order to make the processing go faster, I routinely advised to mount the data filesystems with &#039;forcedirectio&#039;....resulting in the database performance getting much times slower, because now it needed to be read from disk, instead from the operating system cache.]]></description>
		<content:encoded><![CDATA[<p>More or less exactly what I encountered some 5-7 years ago. Large solaris machine (at that time; 32GB memory), database user-data size 1-2GB. In order to make the processing go faster, I routinely advised to mount the data filesystems with &#8216;forcedirectio&#8217;&#8230;.resulting in the database performance getting much times slower, because now it needed to be read from disk, instead from the operating system cache.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Gwen Shapira</title>
		<link>http://kerryosborne.oracle-guy.com/2011/11/tuning-oracle-to-make-a-query-slower/#comment-74297</link>
		<dc:creator>Gwen Shapira</dc:creator>
		<pubDate>Wed, 23 Nov 2011 19:29:39 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=3760#comment-74297</guid>
		<description><![CDATA[In short, if you can read 10g table into the OS cache and not the SGA buffer cache, you are not sizing your SGA right.]]></description>
		<content:encoded><![CDATA[<p>In short, if you can read 10g table into the OS cache and not the SGA buffer cache, you are not sizing your SGA right.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
