<?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: Performance tuning: HugePages and Linux</title>
	<atom:link href="http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/feed/" rel="self" type="application/rss+xml" />
	<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/</link>
	<description>Discussions about Oracle performance tuning, RAC, Oracle internal &#38; E-business suite.</description>
	<lastBuildDate>Sun, 19 May 2013 12:13:20 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.com/</generator>
	<item>
		<title>By: Riyaj Shamsudeen</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-1474</link>
		<dc:creator><![CDATA[Riyaj Shamsudeen]]></dc:creator>
		<pubDate>Sun, 03 Jun 2012 01:16:08 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-1474</guid>
		<description><![CDATA[Hello Jigar
  Thank you. Corrected the typo.
Cheers
Riyaj]]></description>
		<content:encoded><![CDATA[<p>Hello Jigar<br />
  Thank you. Corrected the typo.<br />
Cheers<br />
Riyaj</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jigar Doshi</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-1473</link>
		<dc:creator><![CDATA[Jigar Doshi]]></dc:creator>
		<pubDate>Sun, 03 Jun 2012 00:45:44 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-1473</guid>
		<description><![CDATA[Minor correction. I think it should be /etc/security/limits.conf instead of /etc/securities/limits.conf. Please correct]]></description>
		<content:encoded><![CDATA[<p>Minor correction. I think it should be /etc/security/limits.conf instead of /etc/securities/limits.conf. Please correct</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Hugepages in Linux &#171; Oracle</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-681</link>
		<dc:creator><![CDATA[Hugepages in Linux &#171; Oracle]]></dc:creator>
		<pubDate>Tue, 08 Mar 2011 18:02:34 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-681</guid>
		<description><![CDATA[[...] from Riyaz Shamshuddin quite identical to the issue we had faced. I would just link it here &#8211; Performance tuning: HugePages and Linux    GA_googleAddAttr(&quot;AdOpt&quot;, &quot;1&quot;); GA_googleAddAttr(&quot;Origin&quot;, &quot;other&quot;); [...]]]></description>
		<content:encoded><![CDATA[<p>[...] from Riyaz Shamshuddin quite identical to the issue we had faced. I would just link it here &#8211; Performance tuning: HugePages and Linux    GA_googleAddAttr(&quot;AdOpt&quot;, &quot;1&quot;); GA_googleAddAttr(&quot;Origin&quot;, &quot;other&quot;); [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Lynn Conrad</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-336</link>
		<dc:creator><![CDATA[Lynn Conrad]]></dc:creator>
		<pubDate>Mon, 29 Jun 2009 20:36:16 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-336</guid>
		<description><![CDATA[What was swappiness set too, on aix 5.3 my standard is lru_file_repage=0, memory_affinity=0, v_pinshm=0, maxpin%=94 chuser oracle capabilites=CAP_BYPASS_RAC_VMM, CAP_PROPAGATE, minperm%=5 maxperm%=90 db init parameter lock_sga=true]]></description>
		<content:encoded><![CDATA[<p>What was swappiness set too, on aix 5.3 my standard is lru_file_repage=0, memory_affinity=0, v_pinshm=0, maxpin%=94 chuser oracle capabilites=CAP_BYPASS_RAC_VMM, CAP_PROPAGATE, minperm%=5 maxperm%=90 db init parameter lock_sga=true</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: lscheng</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-244</link>
		<dc:creator><![CDATA[lscheng]]></dc:creator>
		<pubDate>Mon, 16 Mar 2009 07:56:53 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-244</guid>
		<description><![CDATA[I had a customer using 11gR1 in Linux x86-64 platform and automatic memory management feature went through some memory starvation problems (kswapd went crazy freeing memory) a couple of months ago and in order to implement Hugepages AMM had to be turned off. I am not sure if this would be fixed in the future because looks like AMM implementation cannot make use of hugepages.]]></description>
		<content:encoded><![CDATA[<p>I had a customer using 11gR1 in Linux x86-64 platform and automatic memory management feature went through some memory starvation problems (kswapd went crazy freeing memory) a couple of months ago and in order to implement Hugepages AMM had to be turned off. I am not sure if this would be fixed in the future because looks like AMM implementation cannot make use of hugepages.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Krishna Manoharan</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-131</link>
		<dc:creator><![CDATA[Krishna Manoharan]]></dc:creator>
		<pubDate>Sat, 27 Dec 2008 21:32:55 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-131</guid>
		<description><![CDATA[Hi Riyaj,

cat /proc/meminfo &#124;grep HugePages
HugePages_Total:  6144
HugePages_Free:   1655 &lt;-- Free pages are less than total pages.
Hugepagesize:     2048 kB

Keeping HugePages free ( in this example - 1655 - ~3.2GB) would result in physical memory being locked up and not available outside of the Oracle Shared memory realm. If not planning on dynamically resizing the SGA beyond what is set initially, it probably would not make sense to leave so many free huge pages.

Since huge pages are locked in memory, they would not get swapped/paged out. And if huge pages are sized more than required, it can result in more aggresive swapping/paging since the total available physical memory for PGA and other OS needs would be reduced. 

Thanks
Krishna Manoharan]]></description>
		<content:encoded><![CDATA[<p>Hi Riyaj,</p>
<p>cat /proc/meminfo |grep HugePages<br />
HugePages_Total:  6144<br />
HugePages_Free:   1655 &lt;&#8211; Free pages are less than total pages.<br />
Hugepagesize:     2048 kB</p>
<p>Keeping HugePages free ( in this example &#8211; 1655 &#8211; ~3.2GB) would result in physical memory being locked up and not available outside of the Oracle Shared memory realm. If not planning on dynamically resizing the SGA beyond what is set initially, it probably would not make sense to leave so many free huge pages.</p>
<p>Since huge pages are locked in memory, they would not get swapped/paged out. And if huge pages are sized more than required, it can result in more aggresive swapping/paging since the total available physical memory for PGA and other OS needs would be reduced. </p>
<p>Thanks<br />
Krishna Manoharan</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: orainternals</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-82</link>
		<dc:creator><![CDATA[orainternals]]></dc:creator>
		<pubDate>Mon, 17 Nov 2008 18:58:31 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-82</guid>
		<description><![CDATA[Hi Vivek
  Paging/swapping is applicable to IBM AIX as well. But, hugepages may not be directly applicable to AIX. 
  Nevertheless, try to understand why there is memory starvation and resolve it.   
  
Cheers
Riyaj]]></description>
		<content:encoded><![CDATA[<p>Hi Vivek<br />
  Paging/swapping is applicable to IBM AIX as well. But, hugepages may not be directly applicable to AIX.<br />
  Nevertheless, try to understand why there is memory starvation and resolve it.   </p>
<p>Cheers<br />
Riyaj</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Vivek</title>
		<link>http://orainternals.wordpress.com/2008/11/13/performance-tuning-hugepages-and-linux/#comment-81</link>
		<dc:creator><![CDATA[Vivek]]></dc:creator>
		<pubDate>Sat, 15 Nov 2008 18:49:39 +0000</pubDate>
		<guid isPermaLink="false">http://orainternals.wordpress.com/?p=146#comment-81</guid>
		<description><![CDATA[Hi Riyaz,

A nice one. Is it applicable for IBM AIX as well ? I have a 2 node RAC on IBM AIX and one of the node evicts during heavy pagins and page outs.

Regards
Vivek]]></description>
		<content:encoded><![CDATA[<p>Hi Riyaz,</p>
<p>A nice one. Is it applicable for IBM AIX as well ? I have a 2 node RAC on IBM AIX and one of the node evicts during heavy pagins and page outs.</p>
<p>Regards<br />
Vivek</p>
]]></content:encoded>
	</item>
</channel>
</rss>
