<?xml version="1.0" encoding="UTF-8"?>
<!-- generator="bbPress/1.0.3" -->
<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">
	<channel>
		<title>DataStax Support Forums &#187; Topic: Disk speed recommendations</title>
		<link>http://www.datastax.com/support-forums/topic/disk-speed-recommendations</link>
		<description>Software, Support, and Training for Apache Cassandra</description>
		<language>en-US</language>
		<pubDate>Tue, 21 May 2013 12:25:42 +0000</pubDate>
		<generator>http://bbpress.org/?v=1.0.3</generator>
		<textInput>
			<title><![CDATA[Search]]></title>
			<description><![CDATA[Search all topics from these forums.]]></description>
			<name>q</name>
			<link>http://www.datastax.com/support-forums/search.php</link>
		</textInput>
		<atom:link href="http://www.datastax.com/support-forums/rss/topic/disk-speed-recommendations" rel="self" type="application/rss+xml" />

		<item>
			<title>Sven on "Disk speed recommendations"</title>
			<link>http://www.datastax.com/support-forums/topic/disk-speed-recommendations#post-7523</link>
			<pubDate>Fri, 16 Nov 2012 16:26:52 +0000</pubDate>
			<dc:creator>Sven</dc:creator>
			<guid isPermaLink="false">7523@http://www.datastax.com/support-forums/</guid>
			<description>&#60;p&#62;If disk failure was the only factor to consider, then that would be fine. You do want to also cover for other kinds of failures though (network, other hardware). So a smaller replication factor might be good to reduce data shuffeling, but you probably want to keep some replication.
&#60;/p&#62;</description>
		</item>
		<item>
			<title>schalla on "Disk speed recommendations"</title>
			<link>http://www.datastax.com/support-forums/topic/disk-speed-recommendations#post-7505</link>
			<pubDate>Thu, 15 Nov 2012 16:16:19 +0000</pubDate>
			<dc:creator>schalla</dc:creator>
			<guid isPermaLink="false">7505@http://www.datastax.com/support-forums/</guid>
			<description>&#60;p&#62;Thanks sdelmas.&#60;br /&#62;
On the same topic, is it accurate to consider lower replication factor if I am using a RAID10 as I would be getting redundancy/mirror through RAID plus I need not use cassandra to do repairs on disk crashes? What are your thoughts?
&#60;/p&#62;</description>
		</item>
		<item>
			<title>Sven on "Disk speed recommendations"</title>
			<link>http://www.datastax.com/support-forums/topic/disk-speed-recommendations#post-7504</link>
			<pubDate>Thu, 15 Nov 2012 13:34:49 +0000</pubDate>
			<dc:creator>Sven</dc:creator>
			<guid isPermaLink="false">7504@http://www.datastax.com/support-forums/</guid>
			<description>&#60;p&#62;I am not aware of specific performance benchmarks that cover spindle speed.&#60;/p&#62;
&#60;p&#62;While this is an understandable question, it's also a very open ended one. Even just on a hardware layer the spindle speed is just one of many factors that affects performance (what controller type, how much of an on disk cache does the drive have, drive metrics, etc.), Then the next layer is OS, so what OS is this, what file system type, etc.. The next layer up would be, what kind of data is being used, and how is it being accessed.&#60;/p&#62;
&#60;p&#62;The high level recommendation is that faster is obviously better. Raid is certainly helpful if possible. The biggest gain would come from going with SSD. Basically the same rules apply as for any other server kind system.
&#60;/p&#62;</description>
		</item>
		<item>
			<title>schalla on "Disk speed recommendations"</title>
			<link>http://www.datastax.com/support-forums/topic/disk-speed-recommendations#post-7501</link>
			<pubDate>Wed, 14 Nov 2012 22:36:04 +0000</pubDate>
			<dc:creator>schalla</dc:creator>
			<guid isPermaLink="false">7501@http://www.datastax.com/support-forums/</guid>
			<description>&#60;p&#62;Are there any guidelines or recommendations for disk speeds if RAID is not used?&#60;br /&#62;
Is there any benchmark using 7000rpm vs 10000rpm disk speeds?
&#60;/p&#62;</description>
		</item>

	</channel>
</rss>
