<?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: NFS, the slow performance Network File System</title>
	<atom:link href="http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html/feed" rel="self" type="application/rss+xml" />
	<link>http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html</link>
	<description>the blog that matters...</description>
	<lastBuildDate>Tue, 27 Dec 2011 18:10:28 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.3</generator>
	<item>
		<title>By: Robert</title>
		<link>http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html/comment-page-1#comment-2144</link>
		<dc:creator>Robert</dc:creator>
		<pubDate>Mon, 15 Feb 2010 14:58:35 +0000</pubDate>
		<guid isPermaLink="false">http://blog.rarecore.eu/?p=22#comment-2144</guid>
		<description>NFS isn&#039;t an option for scaleout at all. In general, there is no need to write the session handler for MySQL for yourself, as there is a much faster solution: http://websupport.sk/~stanojr/projects/session_mysql/

There is also a session handler for ndb, but memcache should also do the job for the moment.

The problem of sharing cache files between application servers can also be solved without NFS using an asynchron cache purging system. But you should not cache the complete site, if you offer the possebility of commenting or any other dynamic features. If you have something dynamically on the page think about another method, such as blockwise caching using memcache, apc, or in the easiest case by storing serialized objects on the hard drive.

Robert</description>
		<content:encoded><![CDATA[<p>NFS isn&#8217;t an option for scaleout at all. In general, there is no need to write the session handler for MySQL for yourself, as there is a much faster solution: <a href="http://websupport.sk/~stanojr/projects/session_mysql/" rel="nofollow">http://websupport.sk/~stanojr/projects/session_mysql/</a></p>
<p>There is also a session handler for ndb, but memcache should also do the job for the moment.</p>
<p>The problem of sharing cache files between application servers can also be solved without NFS using an asynchron cache purging system. But you should not cache the complete site, if you offer the possebility of commenting or any other dynamic features. If you have something dynamically on the page think about another method, such as blockwise caching using memcache, apc, or in the easiest case by storing serialized objects on the hard drive.</p>
<p>Robert</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: web 2.0 development company</title>
		<link>http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html/comment-page-1#comment-2135</link>
		<dc:creator>web 2.0 development company</dc:creator>
		<pubDate>Mon, 09 Nov 2009 09:00:39 +0000</pubDate>
		<guid isPermaLink="false">http://blog.rarecore.eu/?p=22#comment-2135</guid>
		<description>thanks for this &amp; here is many relevant content written for developers to read 
simply like it most . .</description>
		<content:encoded><![CDATA[<p>thanks for this &#038; here is many relevant content written for developers to read<br />
simply like it most . .</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Ruth</title>
		<link>http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html/comment-page-1#comment-340</link>
		<dc:creator>Ruth</dc:creator>
		<pubDate>Fri, 27 Mar 2009 05:01:00 +0000</pubDate>
		<guid isPermaLink="false">http://blog.rarecore.eu/?p=22#comment-340</guid>
		<description>I recently came across your blog and have been reading along. I thought I would leave my first comment. I don&#039;t know what to say except that I have enjoyed reading. Nice blog. I will keep visiting this blog very often.

Ruth

http://laptopmessengerbag.info</description>
		<content:encoded><![CDATA[<p>I recently came across your blog and have been reading along. I thought I would leave my first comment. I don&#8217;t know what to say except that I have enjoyed reading. Nice blog. I will keep visiting this blog very often.</p>
<p>Ruth</p>
<p><a href="http://laptopmessengerbag.info" rel="nofollow">http://laptopmessengerbag.info</a></p>
]]></content:encoded>
	</item>
	<item>
		<title>By: joe</title>
		<link>http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html/comment-page-1#comment-18</link>
		<dc:creator>joe</dc:creator>
		<pubDate>Mon, 27 Oct 2008 19:15:09 +0000</pubDate>
		<guid isPermaLink="false">http://blog.rarecore.eu/?p=22#comment-18</guid>
		<description>No.
Saving the session and smarty cache in the DB is not as fast as saving it in the main memory. By looking at your page I see that NFS/caching is probably not your real problem. It is Apache. Try the lighttpd or litespeed webservers, they are significantly faster than Apache. By using one of those + memcached, you should be able to serve about 1000 users/second if you use a second server that handles the database.
By the way: I don&#039;t understand your argument, that saving something in the DB can be faster than saving it in the memory - that is against all laws of computer science since the DB is just another layer of abstraction which does require more code than saving something to the memory.</description>
		<content:encoded><![CDATA[<p>No.<br />
Saving the session and smarty cache in the DB is not as fast as saving it in the main memory. By looking at your page I see that NFS/caching is probably not your real problem. It is Apache. Try the lighttpd or litespeed webservers, they are significantly faster than Apache. By using one of those + memcached, you should be able to serve about 1000 users/second if you use a second server that handles the database.<br />
By the way: I don&#8217;t understand your argument, that saving something in the DB can be faster than saving it in the memory &#8211; that is against all laws of computer science since the DB is just another layer of abstraction which does require more code than saving something to the memory.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Exi</title>
		<link>http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html/comment-page-1#comment-10</link>
		<dc:creator>Exi</dc:creator>
		<pubDate>Fri, 10 Oct 2008 05:56:25 +0000</pubDate>
		<guid isPermaLink="false">http://blog.rarecore.eu/?p=22#comment-10</guid>
		<description>Hi,

the best on my solution is that you could choose an extra database server for the cache (smarty &amp; session) and use 2 or more different servers for the cache.

Save the result in the database (smarty) is much fast as saving the result of a query in the ram. the webserver don&#039;t  have to create the part of the site again, it&#039;s already done. He only have to print the string.

Greatz</description>
		<content:encoded><![CDATA[<p>Hi,</p>
<p>the best on my solution is that you could choose an extra database server for the cache (smarty &#038; session) and use 2 or more different servers for the cache.</p>
<p>Save the result in the database (smarty) is much fast as saving the result of a query in the ram. the webserver don&#8217;t  have to create the part of the site again, it&#8217;s already done. He only have to print the string.</p>
<p>Greatz</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Spacefish</title>
		<link>http://blog.rarecore.eu/nfs-the-slow-performance-network-file-system.html/comment-page-1#comment-9</link>
		<dc:creator>Spacefish</dc:creator>
		<pubDate>Wed, 08 Oct 2008 19:29:40 +0000</pubDate>
		<guid isPermaLink="false">http://blog.rarecore.eu/?p=22#comment-9</guid>
		<description>Hey there, why don´t you mirror the files on both webserver and sync them with rsync? or do you know memcached? it´s we use it to accelerate smarty cache. It is decentralized (you could build a memcached cluster) and is EXTREME!! fast!
http://www.danga.com/memcached/ &lt;- it´s open source ;-) you could avoid a lot of DB Querys with it also by caching results from the DB for the news maybe. If someone posts a comment or post new news, you flush the cachekey for that specific page. It can accelerate Webserver a lot. You could also cache parts of HTML in a website, so you don´t have to rerender it on every request. You could cache the commentboxes or the menu or whatever ^^ results in lesser disk reads and lesser CPU usage. We save aprox. 70% mysql querys with it.</description>
		<content:encoded><![CDATA[<p>Hey there, why don´t you mirror the files on both webserver and sync them with rsync? or do you know memcached? it´s we use it to accelerate smarty cache. It is decentralized (you could build a memcached cluster) and is EXTREME!! fast!<br />
<a href="http://www.danga.com/memcached/" rel="nofollow">http://www.danga.com/memcached/</a> &lt;- it´s open source ;-) you could avoid a lot of DB Querys with it also by caching results from the DB for the news maybe. If someone posts a comment or post new news, you flush the cachekey for that specific page. It can accelerate Webserver a lot. You could also cache parts of HTML in a website, so you don´t have to rerender it on every request. You could cache the commentboxes or the menu or whatever ^^ results in lesser disk reads and lesser CPU usage. We save aprox. 70% mysql querys with it.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

