<?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: Using Virtualisation for DR</title>
	<atom:link href="http://thestoragearchitect.com/2007/09/07/using-virtualisation-for-dr/feed/" rel="self" type="application/rss+xml" />
	<link>http://thestoragearchitect.com/2007/09/07/using-virtualisation-for-dr/</link>
	<description>Storage, Virtualisation &#38; Cloud</description>
	<lastBuildDate>Fri, 10 Feb 2012 17:47:20 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.2.1</generator>
	<item>
		<title>By: Nigel</title>
		<link>http://thestoragearchitect.com/2007/09/07/using-virtualisation-for-dr/#comment-154</link>
		<dc:creator>Nigel</dc:creator>
		<pubDate>Fri, 07 Sep 2007 17:16:00 +0000</pubDate>
		<guid isPermaLink="false">http://thestoragearchitect.wordpress.com/2007/09/07/using-virtualisation-for-dr/#comment-154</guid>
		<description>Chris,&lt;br/&gt;&lt;br/&gt;Sorry that I dont actually represent any of the vendors you are asking.  But I do have an opinion and some experience with UVM....... and I am fairly confident that UVM does NOT destage with application write sequencing in mind.&lt;br/&gt;&lt;br/&gt;I &quot;understand&quot; that the same destage algorithms are used for both internal and external disk.&lt;br/&gt;&lt;br/&gt;After all, UVM is not sold as a DR or distance replication technology.  So I would be surprised to see HP or HDS rubber stamp your proposed solution.  I wouldn&#039;t buy it from you just because of your comment &quot;..although technically, the third copy doesn&#039;t actually exist..&quot;.  Even my wife, with her limited storage knowledge, would raise an eyebrow at that  :-D&lt;br/&gt;&lt;br/&gt;Nigel</description>
		<content:encoded><![CDATA[<p>Chris,</p>
<p>Sorry that I dont actually represent any of the vendors you are asking.  But I do have an opinion and some experience with UVM&#8230;&#8230;. and I am fairly confident that UVM does NOT destage with application write sequencing in mind.</p>
<p>I &#8220;understand&#8221; that the same destage algorithms are used for both internal and external disk.</p>
<p>After all, UVM is not sold as a DR or distance replication technology.  So I would be surprised to see HP or HDS rubber stamp your proposed solution.  I wouldn&#8217;t buy it from you just because of your comment &#8220;..although technically, the third copy doesn&#8217;t actually exist..&#8221;.  Even my wife, with her limited storage knowledge, would raise an eyebrow at that  <img src='http://thestoragearchitect.com/wp-includes/images/smilies/icon_biggrin.gif' alt=':-D' class='wp-smiley' /> </p>
<p>Nigel</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mark</title>
		<link>http://thestoragearchitect.com/2007/09/07/using-virtualisation-for-dr/#comment-153</link>
		<dc:creator>Mark</dc:creator>
		<pubDate>Fri, 07 Sep 2007 16:47:00 +0000</pubDate>
		<guid isPermaLink="false">http://thestoragearchitect.wordpress.com/2007/09/07/using-virtualisation-for-dr/#comment-153</guid>
		<description>Invista is stateless. &lt;br/&gt;&lt;br/&gt;No I/O is cached by the Data Path Controllers, there is no store &amp; forward operation.</description>
		<content:encoded><![CDATA[<p>Invista is stateless. </p>
<p>No I/O is cached by the Data Path Controllers, there is no store &#038; forward operation.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: BarryWhyte</title>
		<link>http://thestoragearchitect.com/2007/09/07/using-virtualisation-for-dr/#comment-152</link>
		<dc:creator>BarryWhyte</dc:creator>
		<pubDate>Fri, 07 Sep 2007 16:40:00 +0000</pubDate>
		<guid isPermaLink="false">http://thestoragearchitect.wordpress.com/2007/09/07/using-virtualisation-for-dr/#comment-152</guid>
		<description>Note however that what you propose in fig1 is not possible with SVC. The backend storage can only be presented to one cluster. We do not permit the sharing of disks with SVC and the host (the same disks that is) You can however configure &#039;split controller&#039; where some LUNs are presented to SVC and some direct to a host - just not the same LUNs. The risk of data-miscompares is too great to support such as configuration. I suspect this would be the case for other virtualizers.</description>
		<content:encoded><![CDATA[<p>Note however that what you propose in fig1 is not possible with SVC. The backend storage can only be presented to one cluster. We do not permit the sharing of disks with SVC and the host (the same disks that is) You can however configure &#8216;split controller&#8217; where some LUNs are presented to SVC and some direct to a host &#8211; just not the same LUNs. The risk of data-miscompares is too great to support such as configuration. I suspect this would be the case for other virtualizers.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: BarryWhyte</title>
		<link>http://thestoragearchitect.com/2007/09/07/using-virtualisation-for-dr/#comment-151</link>
		<dc:creator>BarryWhyte</dc:creator>
		<pubDate>Fri, 07 Sep 2007 16:18:00 +0000</pubDate>
		<guid isPermaLink="false">http://thestoragearchitect.wordpress.com/2007/09/07/using-virtualisation-for-dr/#comment-151</guid>
		<description>Needless to say, yes SVC&#039;s GlobalMirror and MetroMirror functions ensure consistency of the remote image by ensuring the I/O is submitted to the remote site before its is cached locally. The exact behavior depends on whether its synchronous or asynchronous.&lt;br/&gt;&lt;br/&gt;This is quite a complex area to explain and I&#039;ll post a more in-depth answer over on my blog as I think I&#039;ll need some pictures to explain it properly.</description>
		<content:encoded><![CDATA[<p>Needless to say, yes SVC&#8217;s GlobalMirror and MetroMirror functions ensure consistency of the remote image by ensuring the I/O is submitted to the remote site before its is cached locally. The exact behavior depends on whether its synchronous or asynchronous.</p>
<p>This is quite a complex area to explain and I&#8217;ll post a more in-depth answer over on my blog as I think I&#8217;ll need some pictures to explain it properly.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Ron</title>
		<link>http://thestoragearchitect.com/2007/09/07/using-virtualisation-for-dr/#comment-150</link>
		<dc:creator>Ron</dc:creator>
		<pubDate>Fri, 07 Sep 2007 13:15:00 +0000</pubDate>
		<guid isPermaLink="false">http://thestoragearchitect.wordpress.com/2007/09/07/using-virtualisation-for-dr/#comment-150</guid>
		<description>While your scenario is technically possible, I don&#039;t know that I would trust it.  Too many &quot;big yellow cable pullers&quot; out there that could really mess up your day.&lt;br/&gt;&lt;br/&gt;With virualization you don&#039;t have consistency groups, so I&#039;m not sure how they would guarantee a &quot;crash consistent&quot; copy of your data.&lt;br/&gt;&lt;br/&gt;I know we&#039;ve asked before about how cache is destaged and what happens to the data if there are &quot;issues&quot;.  I have yet to see a technical answer.</description>
		<content:encoded><![CDATA[<p>While your scenario is technically possible, I don&#8217;t know that I would trust it.  Too many &#8220;big yellow cable pullers&#8221; out there that could really mess up your day.</p>
<p>With virualization you don&#8217;t have consistency groups, so I&#8217;m not sure how they would guarantee a &#8220;crash consistent&#8221; copy of your data.</p>
<p>I know we&#8217;ve asked before about how cache is destaged and what happens to the data if there are &#8220;issues&#8221;.  I have yet to see a technical answer.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

