<?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/"
	xmlns:georss="http://www.georss.org/georss" xmlns:geo="http://www.w3.org/2003/01/geo/wgs84_pos#" xmlns:media="http://search.yahoo.com/mrss/"
		>
<channel>
	<title>Comments on: LOB sizing</title>
	<atom:link href="http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/feed/" rel="self" type="application/rss+xml" />
	<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/</link>
	<description>Just another Oracle weblog</description>
	<lastBuildDate>Wed, 22 May 2013 12:40:08 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.com/</generator>
	<item>
		<title>By: Jonathan Lewis</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43561</link>
		<dc:creator><![CDATA[Jonathan Lewis]]></dc:creator>
		<pubDate>Fri, 23 Dec 2011 18:50:52 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43561</guid>
		<description><![CDATA[Amin,

If you want help in understanding how to use the context option (currently called Oracle Text) the appropriate forum is at https://forums.oracle.com/forums/forum.jspa?forumID=71]]></description>
		<content:encoded><![CDATA[<p>Amin,</p>
<p>If you want help in understanding how to use the context option (currently called Oracle Text) the appropriate forum is at <a href="https://forums.oracle.com/forums/forum.jspa?forumID=71" rel="nofollow">https://forums.oracle.com/forums/forum.jspa?forumID=71</a></p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Amin Adatia</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43548</link>
		<dc:creator><![CDATA[Amin Adatia]]></dc:creator>
		<pubDate>Fri, 23 Dec 2011 13:33:23 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43548</guid>
		<description><![CDATA[Jonathan

Thanks. I already have set the memory to the max allowed (2GB) and the default to 1GB. The tablespaces were created with &quot;defaults&quot; which happen to have pctincrease 50%. I am moving to the auto-allocate as I get the chance to do so. We also have partitioned tables (to mainly help with the deletes) and we use tablespaces in a round-robin (tablespace for the day essentially) in case a datafile gets corrupted ad the whole tablespace becomes unavailable until the data files are resolved.

Anyway, how to find out if the memory allocations/parameters are working optimally?]]></description>
		<content:encoded><![CDATA[<p>Jonathan</p>
<p>Thanks. I already have set the memory to the max allowed (2GB) and the default to 1GB. The tablespaces were created with &#8220;defaults&#8221; which happen to have pctincrease 50%. I am moving to the auto-allocate as I get the chance to do so. We also have partitioned tables (to mainly help with the deletes) and we use tablespaces in a round-robin (tablespace for the day essentially) in case a datafile gets corrupted ad the whole tablespace becomes unavailable until the data files are resolved.</p>
<p>Anyway, how to find out if the memory allocations/parameters are working optimally?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jonathan Lewis</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43545</link>
		<dc:creator><![CDATA[Jonathan Lewis]]></dc:creator>
		<pubDate>Fri, 23 Dec 2011 10:06:12 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43545</guid>
		<description><![CDATA[Amin,

Unless you&#039;re on a very old version of Oracle you shouldn&#039;t be thinking about things like initial, next and pctincrease anyway.

As a general guideline for tablespaces (and block sizes) - if you can&#039;t think of a good reason for changing the defaults then use the defaults.

In the case of context it&#039;s probably more important to worry about setting the memory allocation attribute for creating the context index as this dictates the volume of data (e.g. number of files) that the code will read as a unit when creating the index - and the size of the LOB content of the index tables will be dictated by the volume of data read as a unit.]]></description>
		<content:encoded><![CDATA[<p>Amin,</p>
<p>Unless you&#8217;re on a very old version of Oracle you shouldn&#8217;t be thinking about things like initial, next and pctincrease anyway.</p>
<p>As a general guideline for tablespaces (and block sizes) &#8211; if you can&#8217;t think of a good reason for changing the defaults then use the defaults.</p>
<p>In the case of context it&#8217;s probably more important to worry about setting the memory allocation attribute for creating the context index as this dictates the volume of data (e.g. number of files) that the code will read as a unit when creating the index &#8211; and the size of the LOB content of the index tables will be dictated by the volume of data read as a unit.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Amin Adatia</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43419</link>
		<dc:creator><![CDATA[Amin Adatia]]></dc:creator>
		<pubDate>Tue, 20 Dec 2011 18:40:30 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-43419</guid>
		<description><![CDATA[Hi Jonathan

What are the implications of specifying the Storage preference for the $I, $N, etc tables?

v_Attributes := &#039;TABLESPACE  storage(initial 1K next 1K pctincrease 0);
CTX_DDL.Set_Attribute(,I_TABLE_CLAUSE,);

OR is it just better to specify just the tablesapace and take the defaults.

Regards]]></description>
		<content:encoded><![CDATA[<p>Hi Jonathan</p>
<p>What are the implications of specifying the Storage preference for the $I, $N, etc tables?</p>
<p>v_Attributes := &#8216;TABLESPACE  storage(initial 1K next 1K pctincrease 0);<br />
CTX_DDL.Set_Attribute(,I_TABLE_CLAUSE,);</p>
<p>OR is it just better to specify just the tablesapace and take the defaults.</p>
<p>Regards</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jonathan Lewis</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-34706</link>
		<dc:creator><![CDATA[Jonathan Lewis]]></dc:creator>
		<pubDate>Fri, 23 Oct 2009 16:00:32 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-34706</guid>
		<description><![CDATA[Vijay,

Sorry about the delay in replying - sometimes I just lose track of comments because of all the travel I do. I&#039;ll be writing up a short note in the next few days to answer this question.]]></description>
		<content:encoded><![CDATA[<p>Vijay,</p>
<p>Sorry about the delay in replying &#8211; sometimes I just lose track of comments because of all the travel I do. I&#8217;ll be writing up a short note in the next few days to answer this question.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Vijay</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-34553</link>
		<dc:creator><![CDATA[Vijay]]></dc:creator>
		<pubDate>Fri, 02 Oct 2009 14:03:23 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-34553</guid>
		<description><![CDATA[Dear Jonathan,

if you can please put some highlight some options or point out some way using which one can decide which option for LOB storage (CACHE, CACHE READ, or NOCACHE; LOGGING or NOLOGGING) should be used. 

In our system we are having tables with CLOB columns, the system is third party and mostly deletes on the tables with CLOB column is taking a lot of time. I was wondering if changing some storage settings will help us, presently we have ENABLE STORAGE IN ROW CHUNK 8192, NOCACHE LOGGING.

Regards,
Vijay.]]></description>
		<content:encoded><![CDATA[<p>Dear Jonathan,</p>
<p>if you can please put some highlight some options or point out some way using which one can decide which option for LOB storage (CACHE, CACHE READ, or NOCACHE; LOGGING or NOLOGGING) should be used. </p>
<p>In our system we are having tables with CLOB columns, the system is third party and mostly deletes on the tables with CLOB column is taking a lot of time. I was wondering if changing some storage settings will help us, presently we have ENABLE STORAGE IN ROW CHUNK 8192, NOCACHE LOGGING.</p>
<p>Regards,<br />
Vijay.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Igor</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8455</link>
		<dc:creator><![CDATA[Igor]]></dc:creator>
		<pubDate>Tue, 15 May 2007 11:51:46 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8455</guid>
		<description><![CDATA[Hi Jonathan, 
There were a lot of bugs related to LOBS (in 9.2 version)
in combination with ASSM (last one we have seen it seems to be fixed only with latest patchset 10.2.0.3). 
In tests we have done in this context (inserts + delete in batches), it looks like that still with 10.2 version and MSSM - not only is more stable but performed better than BLOB in ASSM.  
(also RETENTION showed better results from performance perspective in all versions but difference was much more important in 9.2 version). 
What is your experience/comment  ? 

How you decided RETENTION vs PCTVERSION ? 

Thank you and regards,
Igor]]></description>
		<content:encoded><![CDATA[<p>Hi Jonathan,<br />
There were a lot of bugs related to LOBS (in 9.2 version)<br />
in combination with ASSM (last one we have seen it seems to be fixed only with latest patchset 10.2.0.3).<br />
In tests we have done in this context (inserts + delete in batches), it looks like that still with 10.2 version and MSSM &#8211; not only is more stable but performed better than BLOB in ASSM.<br />
(also RETENTION showed better results from performance perspective in all versions but difference was much more important in 9.2 version).<br />
What is your experience/comment  ? </p>
<p>How you decided RETENTION vs PCTVERSION ? </p>
<p>Thank you and regards,<br />
Igor</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jonathan Lewis</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8393</link>
		<dc:creator><![CDATA[Jonathan Lewis]]></dc:creator>
		<pubDate>Mon, 14 May 2007 20:27:32 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8393</guid>
		<description><![CDATA[Sid, we weren&#039;t trying to cache the entire LOB segment (which was down to about 64GB) in the recycle pool. We were using the recycle pool so that (a) we could cycle the blocks through a caching mechanism - to avoid direct path reads and writes but (b) avoid thrashing the default pool to death because of the huge volume of LOB that would go through the cache over time. For the activity going on, sizes around 128MB were adequate.

Satish, the minimum size for an &#039;out of line&#039; LOB is one &quot;chunk&quot;, and the minimum size of a chunk is one block. Most of our LOBs were smaller than 8KB, but the tablespace for the LOB segment was 16KB - so most of our LOBs were using less than half a block.  By halving the block size we saved half the space - the half that we were wasting in most blocks.
]]></description>
		<content:encoded><![CDATA[<p>Sid, we weren&#8217;t trying to cache the entire LOB segment (which was down to about 64GB) in the recycle pool. We were using the recycle pool so that (a) we could cycle the blocks through a caching mechanism &#8211; to avoid direct path reads and writes but (b) avoid thrashing the default pool to death because of the huge volume of LOB that would go through the cache over time. For the activity going on, sizes around 128MB were adequate.</p>
<p>Satish, the minimum size for an &#8216;out of line&#8217; LOB is one &#8220;chunk&#8221;, and the minimum size of a chunk is one block. Most of our LOBs were smaller than 8KB, but the tablespace for the LOB segment was 16KB &#8211; so most of our LOBs were using less than half a block.  By halving the block size we saved half the space &#8211; the half that we were wasting in most blocks.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sathish</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8374</link>
		<dc:creator><![CDATA[Sathish]]></dc:creator>
		<pubDate>Mon, 14 May 2007 15:09:40 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8374</guid>
		<description><![CDATA[&quot;Moving the LOB segment to a tablespace with an 8KB block size would automatically save 128GB&quot; - Did not understand this? I am missing a point.]]></description>
		<content:encoded><![CDATA[<p>&#8220;Moving the LOB segment to a tablespace with an 8KB block size would automatically save 128GB&#8221; &#8211; Did not understand this? I am missing a point.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: sid</title>
		<link>http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8372</link>
		<dc:creator><![CDATA[sid]]></dc:creator>
		<pubDate>Mon, 14 May 2007 14:42:34 +0000</pubDate>
		<guid isPermaLink="false">http://jonathanlewis.wordpress.com/2007/05/11/lob-sizing/#comment-8372</guid>
		<description><![CDATA[Jonathan,

How much was &#039;recycle pool&#039; size after caching? I wonder how much buffer cache is required to CACHE 0.25 TB of  LOB.

Thanks]]></description>
		<content:encoded><![CDATA[<p>Jonathan,</p>
<p>How much was &#8216;recycle pool&#8217; size after caching? I wonder how much buffer cache is required to CACHE 0.25 TB of  LOB.</p>
<p>Thanks</p>
]]></content:encoded>
	</item>
</channel>
</rss>
