<?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: BoardReader &#8211; Forum Search Engine</title>
	<atom:link href="http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/</link>
	<description>Everything about MySQL Performance</description>
	<lastBuildDate>Sat, 21 Nov 2009 14:06:19 -0800</lastBuildDate>
	<generator>http://wordpress.org/?v=2.8.4</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Vadim</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-675626</link>
		<dc:creator>Vadim</dc:creator>
		<pubDate>Wed, 11 Nov 2009 17:54:10 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-675626</guid>
		<description>James,

It&#039;s mainly InnoDB.
MyISAM is used to store performance log statistics.</description>
		<content:encoded><![CDATA[<p>James,</p>
<p>It&#8217;s mainly InnoDB.<br />
MyISAM is used to store performance log statistics.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: james</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-675610</link>
		<dc:creator>james</dc:creator>
		<pubDate>Wed, 11 Nov 2009 16:35:47 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-675610</guid>
		<description>Peter, would you please share with us which mysql storage engine is being used to store that much amount of data?

thanks!</description>
		<content:encoded><![CDATA[<p>Peter, would you please share with us which mysql storage engine is being used to store that much amount of data?</p>
<p>thanks!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: peter</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-392711</link>
		<dc:creator>peter</dc:creator>
		<pubDate>Wed, 26 Nov 2008 19:19:01 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-392711</guid>
		<description>The aggregator node is data less and just forwards data so we never seen it failing without all box going down.   This also makes it easy to run multiple copies of it if you need it to.  But technically yes if this one fails you lose 1/6 of the data.</description>
		<content:encoded><![CDATA[<p>The aggregator node is data less and just forwards data so we never seen it failing without all box going down.   This also makes it easy to run multiple copies of it if you need it to.  But technically yes if this one fails you lose 1/6 of the data.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: http://lesdatabases.blogspot.com/</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-392667</link>
		<dc:creator>http://lesdatabases.blogspot.com/</dc:creator>
		<pubDate>Wed, 26 Nov 2008 17:42:10 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-392667</guid>
		<description>Peter,

In High Performance MySQL 2nd Edition (congratulations for all the work done), you (I don&#039;t forget the other contributors) say that there are 6 servers with 4 searchd instances per server, and that one of the four searchd aggregates the result from the other three. In that situation, what happens if this one fail ? Cause, in this case, if I always go to this searchd instance I&#039;ll lose 1/6 of indexes and not just 1/24.</description>
		<content:encoded><![CDATA[<p>Peter,</p>
<p>In High Performance MySQL 2nd Edition (congratulations for all the work done), you (I don&#8217;t forget the other contributors) say that there are 6 servers with 4 searchd instances per server, and that one of the four searchd aggregates the result from the other three. In that situation, what happens if this one fail ? Cause, in this case, if I always go to this searchd instance I&#8217;ll lose 1/6 of indexes and not just 1/24.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: peter</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-318942</link>
		<dc:creator>peter</dc:creator>
		<pubDate>Sat, 28 Jun 2008 05:35:23 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-318942</guid>
		<description>York,

If you always need to have HA keep 2 boxes having same indexes. You can load balance between them and take one to take over full load if other fails.  

We use Dual QuadCore  CPUs now.</description>
		<content:encoded><![CDATA[<p>York,</p>
<p>If you always need to have HA keep 2 boxes having same indexes. You can load balance between them and take one to take over full load if other fails.  </p>
<p>We use Dual QuadCore  CPUs now.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: York</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-318593</link>
		<dc:creator>York</dc:creator>
		<pubDate>Fri, 27 Jun 2008 09:41:12 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-318593</guid>
		<description>Peter,

I know SPHINX very well and I want you explain me how do you deal with the problem of high availability ! 

i.e : If 4 nodes are down, how you can imagine the situation to give to the last user the same results as if they are not down ?

I tried on my own to imagine that there is a bench of old machines contains same indexes just for backup. 
And in sphinx configuration with the SetIndexWeights() function , give those a low weight and those on normal ones a high weight. And then Query all the nodes, backup machines included... what do you think ?

I have an unrelated question : 

do you think does an Quad Xeon processor is better then a 2 * Dual core Xeon ? to implement sphinx with distributed searching !</description>
		<content:encoded><![CDATA[<p>Peter,</p>
<p>I know SPHINX very well and I want you explain me how do you deal with the problem of high availability ! </p>
<p>i.e : If 4 nodes are down, how you can imagine the situation to give to the last user the same results as if they are not down ?</p>
<p>I tried on my own to imagine that there is a bench of old machines contains same indexes just for backup.<br />
And in sphinx configuration with the SetIndexWeights() function , give those a low weight and those on normal ones a high weight. And then Query all the nodes, backup machines included&#8230; what do you think ?</p>
<p>I have an unrelated question : </p>
<p>do you think does an Quad Xeon processor is better then a 2 * Dual core Xeon ? to implement sphinx with distributed searching !</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: peter</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-318420</link>
		<dc:creator>peter</dc:creator>
		<pubDate>Thu, 26 Jun 2008 21:22:32 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-318420</guid>
		<description>York,

It is even more now :)   We use 4 indexes per box each having dedicated hard drive.  We do not use CPU affinity for the search nodes though we could try and see.    There is probably 200Gb forth of index per node so some 50GB per CPU.    The criteria is random distribution. Though indexes themselves are also date partitioned so when you search for last 3 months only you do not have to go over all data.</description>
		<content:encoded><![CDATA[<p>York,</p>
<p>It is even more now <img src='http://www.mysqlperformanceblog.com/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' />    We use 4 indexes per box each having dedicated hard drive.  We do not use CPU affinity for the search nodes though we could try and see.    There is probably 200Gb forth of index per node so some 50GB per CPU.    The criteria is random distribution. Though indexes themselves are also date partitioned so when you search for last 3 months only you do not have to go over all data.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: yokr</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-318304</link>
		<dc:creator>yokr</dc:creator>
		<pubDate>Thu, 26 Jun 2008 10:23:30 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-318304</guid>
		<description>You are now about 1,2 Billion of indexed forum posts ! I know that with Sphinx you can do distributed searching along the nodes! 
My questions are : 

- what is the size of sphinx index you use and what is their number per machine?
- Does one index is fully allocated to one CPU? or a branch of them ?
- What is the criteria to subdivide data into different machines and different index ?

I appreciate your board.

Thanks &amp; BRAVO</description>
		<content:encoded><![CDATA[<p>You are now about 1,2 Billion of indexed forum posts ! I know that with Sphinx you can do distributed searching along the nodes!<br />
My questions are : </p>
<p>- what is the size of sphinx index you use and what is their number per machine?<br />
- Does one index is fully allocated to one CPU? or a branch of them ?<br />
- What is the criteria to subdivide data into different machines and different index ?</p>
<p>I appreciate your board.</p>
<p>Thanks &amp; BRAVO</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: peter</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-168150</link>
		<dc:creator>peter</dc:creator>
		<pubDate>Sat, 15 Sep 2007 09:33:35 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-168150</guid>
		<description>Ryan,

Custom forum software may require some extra work with parsing to be added.
But please feel free to submit it depending on the size of site crawl guys can get to it sooner or later.</description>
		<content:encoded><![CDATA[<p>Ryan,</p>
<p>Custom forum software may require some extra work with parsing to be added.<br />
But please feel free to submit it depending on the size of site crawl guys can get to it sooner or later.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: ryan</title>
		<link>http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/comment-page-1/#comment-167699</link>
		<dc:creator>ryan</dc:creator>
		<pubDate>Fri, 14 Sep 2007 18:08:59 +0000</pubDate>
		<guid isPermaLink="false">http://www.mysqlperformanceblog.com/2006/11/29/boardreader-forum-search-engine/#comment-167699</guid>
		<description>hmm, does boardreader work with custom &quot;home-brew&quot; forum software?</description>
		<content:encoded><![CDATA[<p>hmm, does boardreader work with custom &#8220;home-brew&#8221; forum software?</p>
]]></content:encoded>
	</item>
</channel>
</rss>
