<?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: Exadata Offload &#8211; The Secret Sauce</title>
	<atom:link href="http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/feed/" rel="self" type="application/rss+xml" />
	<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/</link>
	<description>Just another Oracle blog</description>
	<lastBuildDate>Wed, 19 Jun 2013 09:07:28 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.5.1</generator>
	<item>
		<title>By: osborne</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-138035</link>
		<dc:creator>osborne</dc:creator>
		<pubDate>Mon, 20 Aug 2012 15:19:44 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-138035</guid>
		<description><![CDATA[Cool. Thanks for the follow up.]]></description>
		<content:encoded><![CDATA[<p>Cool. Thanks for the follow up.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: HarshadMark</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-138032</link>
		<dc:creator>HarshadMark</dc:creator>
		<pubDate>Mon, 20 Aug 2012 15:14:11 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-138032</guid>
		<description><![CDATA[Thanks a lot for your reply.
The issue is resolved with the cellsrv software upgrade. The versions are :
Old : 11.2.2.1.1     
New : 11.2.2.4.2 

I&#039;d even tried with alter session set “_kcfis_storageidx_disabled”=true; but the null issue still remained on old version.
Now with the new cell sw upgrade the results are as expected. 

Thanks
Harshad]]></description>
		<content:encoded><![CDATA[<p>Thanks a lot for your reply.<br />
The issue is resolved with the cellsrv software upgrade. The versions are :<br />
Old : 11.2.2.1.1<br />
New : 11.2.2.4.2 </p>
<p>I&#8217;d even tried with alter session set “_kcfis_storageidx_disabled”=true; but the null issue still remained on old version.<br />
Now with the new cell sw upgrade the results are as expected. </p>
<p>Thanks<br />
Harshad</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: osborne</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-136759</link>
		<dc:creator>osborne</dc:creator>
		<pubDate>Thu, 16 Aug 2012 03:38:25 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-136759</guid>
		<description><![CDATA[11.2.0.1 was the first release on the V2&#039;s and it had a few wrong results bugs. Many were related to storage indexes. You didn&#039;t mention what cellsrv version you were running but there were also some issues caused by patching cellsrv and leaving the db unpatched.

You said you had tried turning off storage indexes. Just verifying you used this syntax.

alter session set &quot;_kcfis_storageidx_disabled&quot;=true;

You can also check the stat &quot;cell physical IO bytes saved by storage index&quot; in v$mystat or v$sesstat to make sure storge indexes have not been used. I published a little script called &lt;a href=&quot;http://kerryosborne.oracle-guy.com/scripts/mystats.sql&quot;  rel=&quot;nofollow&quot;&gt;mystats.sql&lt;/a&gt; that does that you can use for that. You can find more info about it here: &lt;a href=&quot;http://kerryosborne.oracle-guy.com/2010/08/oracle-exadata-storage-indexes/&quot; rel=&quot;nofollow&quot;&gt;Oracle Exadata – Storage Indexes&lt;/a&gt;

At any rate, you should definitely consider upgrading, as 11.2.0.1 is several years old and there have been many bug fixes and enhancements since then.]]></description>
		<content:encoded><![CDATA[<p>11.2.0.1 was the first release on the V2&#8242;s and it had a few wrong results bugs. Many were related to storage indexes. You didn&#8217;t mention what cellsrv version you were running but there were also some issues caused by patching cellsrv and leaving the db unpatched.</p>
<p>You said you had tried turning off storage indexes. Just verifying you used this syntax.</p>
<p>alter session set &#8220;_kcfis_storageidx_disabled&#8221;=true;</p>
<p>You can also check the stat &#8220;cell physical IO bytes saved by storage index&#8221; in v$mystat or v$sesstat to make sure storge indexes have not been used. I published a little script called <a href="http://kerryosborne.oracle-guy.com/scripts/mystats.sql"  rel="nofollow">mystats.sql</a> that does that you can use for that. You can find more info about it here: <a href="http://kerryosborne.oracle-guy.com/2010/08/oracle-exadata-storage-indexes/" rel="nofollow">Oracle Exadata – Storage Indexes</a></p>
<p>At any rate, you should definitely consider upgrading, as 11.2.0.1 is several years old and there have been many bug fixes and enhancements since then.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: HarshadMark</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-136634</link>
		<dc:creator>HarshadMark</dc:creator>
		<pubDate>Wed, 15 Aug 2012 18:23:16 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-136634</guid>
		<description><![CDATA[Sorry.. 
When I change the setting to 
alter session set cell_offload_processing=FASLE;

Oracle returns no rows ( expected output ).
The current setting  is cell_offload_processing=TRUE, and when query runs under this setting it returns nulls for col1&lt;0.]]></description>
		<content:encoded><![CDATA[<p>Sorry..<br />
When I change the setting to<br />
alter session set cell_offload_processing=FASLE;</p>
<p>Oracle returns no rows ( expected output ).<br />
The current setting  is cell_offload_processing=TRUE, and when query runs under this setting it returns nulls for col1&lt;0.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: HarshadMark</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-136618</link>
		<dc:creator>HarshadMark</dc:creator>
		<pubDate>Wed, 15 Aug 2012 16:33:43 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-136618</guid>
		<description><![CDATA[I&#039;m running the query on Exdata v2 (Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production).
The table is partitioned and the targeted partition contains about 5 million rows.

Partitiion : Range by date 
subpartition : list 
col1 : number data type and it is NOT part of partition/subpartition scheme.
Query : select col1 from t1 partiton (p_20110101) where col1 &lt; 0 ;

I see Oracle returning rows whre col1 is null.

When I run the same query by changing the -
alter session set cell_offload_processing=TRUE;

I see no output.

I could not understand why Exadata treats nulls different than it treats on regular Oracle database.
I thought Storage index might be the cuase of this behaviour..I tried changing &quot;_kcfis_storageidx_disabled&quot;
but I still get the same o/p ie Oracle retruns nulls for the condtition col1 &lt; 0.

Have you seen this before ?
Please let me know if you need more info.

Thanks
Harshad]]></description>
		<content:encoded><![CDATA[<p>I&#8217;m running the query on Exdata v2 (Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 &#8211; 64bit Production).<br />
The table is partitioned and the targeted partition contains about 5 million rows.</p>
<p>Partitiion : Range by date<br />
subpartition : list<br />
col1 : number data type and it is NOT part of partition/subpartition scheme.<br />
Query : select col1 from t1 partiton (p_20110101) where col1 &lt; 0 ;</p>
<p>I see Oracle returning rows whre col1 is null.</p>
<p>When I run the same query by changing the -<br />
alter session set cell_offload_processing=TRUE;</p>
<p>I see no output.</p>
<p>I could not understand why Exadata treats nulls different than it treats on regular Oracle database.<br />
I thought Storage index might be the cuase of this behaviour..I tried changing &quot;_kcfis_storageidx_disabled&quot;<br />
but I still get the same o/p ie Oracle retruns nulls for the condtition col1 &lt; 0.</p>
<p>Have you seen this before ?<br />
Please let me know if you need more info.</p>
<p>Thanks<br />
Harshad</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: osborne</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-136583</link>
		<dc:creator>osborne</dc:creator>
		<pubDate>Wed, 15 Aug 2012 13:39:17 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-136583</guid>
		<description><![CDATA[Joshua - When the IO_SAVED_% column is negative it means that more data was transferred than exists in the object(s) that are being scanned. There are numerous reasons this can happen. Sorts that spill over into temp for example often do a lot of i/o and if the sum is great than the IO_CELL_OFFLOAD_ELIGIBLE_BYTES then the calculation goes negative. Hope that helps. 

Kerry]]></description>
		<content:encoded><![CDATA[<p>Joshua &#8211; When the IO_SAVED_% column is negative it means that more data was transferred than exists in the object(s) that are being scanned. There are numerous reasons this can happen. Sorts that spill over into temp for example often do a lot of i/o and if the sum is great than the IO_CELL_OFFLOAD_ELIGIBLE_BYTES then the calculation goes negative. Hope that helps. </p>
<p>Kerry</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: joshua</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-136159</link>
		<dc:creator>joshua</dc:creator>
		<pubDate>Tue, 14 Aug 2012 02:57:31 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-136159</guid>
		<description><![CDATA[Hi Kerry , 
Do you have any idea why IO_SAVED_% can be negative even when I am not using compression

SQL_ID         CHILD  PLAN_HASH  EXECS  AVG_ETIME AVG_PX OFFLOAD IO_SAVED_% SQL_
------------- ------ ---------- ------ ---------- ------ ------- ---------- ----
2uka1yym51jfg      1 3599640982      1       4.06     13 Yes         -55.58 inse]]></description>
		<content:encoded><![CDATA[<p>Hi Kerry ,<br />
Do you have any idea why IO_SAVED_% can be negative even when I am not using compression</p>
<p>SQL_ID         CHILD  PLAN_HASH  EXECS  AVG_ETIME AVG_PX OFFLOAD IO_SAVED_% SQL_<br />
&#8212;&#8212;&#8212;&#8212;- &#8212;&#8212; &#8212;&#8212;&#8212;- &#8212;&#8212; &#8212;&#8212;&#8212;- &#8212;&#8212; &#8212;&#8212;- &#8212;&#8212;&#8212;- &#8212;-<br />
2uka1yym51jfg      1 3599640982      1       4.06     13 Yes         -55.58 inse</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Fun with _SERIAL_DIRECT_READ &#171; ORAganism</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-46103</link>
		<dc:creator>Fun with _SERIAL_DIRECT_READ &#171; ORAganism</dc:creator>
		<pubDate>Mon, 27 Jun 2011 20:33:31 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-46103</guid>
		<description><![CDATA[[...] getting a reasonable amount of attention since 11g was released. Doug Burns, Dion Cho (and here), Kerry Osborne and Miladin Modrakovic have all written on the subject. Tanel Põder also covered it in his [...]]]></description>
		<content:encoded><![CDATA[<p>[...] getting a reasonable amount of attention since 11g was released. Doug Burns, Dion Cho (and here), Kerry Osborne and Miladin Modrakovic have all written on the subject. Tanel Põder also covered it in his [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Exadata &#171; Oracle Scratchpad</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-22083</link>
		<dc:creator>Exadata &#171; Oracle Scratchpad</dc:creator>
		<pubDate>Tue, 14 Sep 2010 19:00:32 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-22083</guid>
		<description><![CDATA[[...] Offloading [...]]]></description>
		<content:encoded><![CDATA[<p>[...] Offloading [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: osborne</title>
		<link>http://kerryosborne.oracle-guy.com/2010/06/exadata-offload-the-secret-sauce/#comment-14518</link>
		<dc:creator>osborne</dc:creator>
		<pubDate>Thu, 08 Jul 2010 15:57:09 +0000</pubDate>
		<guid isPermaLink="false">http://kerryosborne.oracle-guy.com/?p=2531#comment-14518</guid>
		<description><![CDATA[Yes SPARC. We&#039;ve loaded up about a 1T database and run a batch process that takes 4-5 hours on the Sun in about 45min on a quarter rack. That&#039;s without doing anything other than loading the data and running it (across both nodes of course). But in general, I was just talking about how it feels. Everything on the Sun feels like it runs at a glacial speed vs. what we&#039;ve gotten used to on the Exadata.]]></description>
		<content:encoded><![CDATA[<p>Yes SPARC. We&#8217;ve loaded up about a 1T database and run a batch process that takes 4-5 hours on the Sun in about 45min on a quarter rack. That&#8217;s without doing anything other than loading the data and running it (across both nodes of course). But in general, I was just talking about how it feels. Everything on the Sun feels like it runs at a glacial speed vs. what we&#8217;ve gotten used to on the Exadata.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
