<?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: Multi Protocol Storage Provisioning with COMSTAR</title>
	<atom:link href="http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/feed" rel="self" type="application/rss+xml" />
	<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with</link>
	<description>Blogging for technical minds.</description>
	<lastBuildDate>Fri, 27 Jan 2012 17:53:51 -0600</lastBuildDate>
	<generator>http://wordpress.org/?v=2.9.2</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: it answers</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-3441</link>
		<dc:creator>it answers</dc:creator>
		<pubDate>Sun, 25 Sep 2011 03:21:46 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-3441</guid>
		<description>&lt;strong&gt;IT Questions and answers...&lt;/strong&gt;

[...]Ubiquitous Talk &#187; Multi Protocol Storage Provisioning with COMSTAR[...]...</description>
		<content:encoded><![CDATA[<p><strong>IT Questions and answers&#8230;</strong></p>
<p>[...]Ubiquitous Talk &raquo; Multi Protocol Storage Provisioning with COMSTAR[...]&#8230;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-1062</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Mon, 31 May 2010 04:26:55 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-1062</guid>
		<description>Hi Josh,

Yes, ZFS currently is not a clustered file system. To provision HA as you indicated requires a second head and the use of the zpool import functionality. There is a delay while the import occurs. COMSTAR can easily serve the same transitioned LUs or NFS shares on the second head with minor protocol retry events. It&#039;s not all that different from many of the formal HA solutions out there. It can be an active/active configuration by using multiple zpools but not to the same LU. I am not aware of an open source product that runs on OpenSolaris which can provision clustered LVMs.

Regards,

Mike</description>
		<content:encoded><![CDATA[<p>Hi Josh,</p>
<p>Yes, ZFS currently is not a clustered file system. To provision HA as you indicated requires a second head and the use of the zpool import functionality. There is a delay while the import occurs. COMSTAR can easily serve the same transitioned LUs or NFS shares on the second head with minor protocol retry events. It&#8217;s not all that different from many of the formal HA solutions out there. It can be an active/active configuration by using multiple zpools but not to the same LU. I am not aware of an open source product that runs on OpenSolaris which can provision clustered LVMs.</p>
<p>Regards,</p>
<p>Mike</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Josh</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-1059</link>
		<dc:creator>Josh</dc:creator>
		<pubDate>Mon, 31 May 2010 03:23:57 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-1059</guid>
		<description>On the one hand, this is a really good concept, and I have many customers who are excited about deploying Comstar. On the other hand, they need HA storage controllers. These days, &quot;table stakes&quot; for controllers is active/active presentation of LUNs. The minimum credible product would be two Comstar nodes each projecting the same LUN set at the same time. The problem is that ZFS isn&#039;t a clustered LVM. I looked into using Veritas CVM, but the cost is so high that it&#039;s cheaper to go with an enterprise RAID array. Do you know of an open source cluster LVM option for OpenSolaris? Otherwise, enterprise customers will need to either stick with the traditional storage approach, or perhaps use the Linux equivalent of Comstar. (SCST.) That would allow the use of the RHEL cluster LVM.</description>
		<content:encoded><![CDATA[<p>On the one hand, this is a really good concept, and I have many customers who are excited about deploying Comstar. On the other hand, they need HA storage controllers. These days, &#8220;table stakes&#8221; for controllers is active/active presentation of LUNs. The minimum credible product would be two Comstar nodes each projecting the same LUN set at the same time. The problem is that ZFS isn&#8217;t a clustered LVM. I looked into using Veritas CVM, but the cost is so high that it&#8217;s cheaper to go with an enterprise RAID array. Do you know of an open source cluster LVM option for OpenSolaris? Otherwise, enterprise customers will need to either stick with the traditional storage approach, or perhaps use the Linux equivalent of Comstar. (SCST.) That would allow the use of the RHEL cluster LVM.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-777</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Wed, 31 Mar 2010 03:15:36 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-777</guid>
		<description>Thanks for the comment.

The SCST project is very interesting and both the SCST and COMSTAR architectures are exceptionally good. I have explored both and I believe that the COMSTAR architecture has some distinct advantages with it&#039;s OpenSolaris hosting and ZFS storage provisioning.

Regards,

Mike</description>
		<content:encoded><![CDATA[<p>Thanks for the comment.</p>
<p>The SCST project is very interesting and both the SCST and COMSTAR architectures are exceptionally good. I have explored both and I believe that the COMSTAR architecture has some distinct advantages with it&#8217;s OpenSolaris hosting and ZFS storage provisioning.</p>
<p>Regards,</p>
<p>Mike</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Shankhadeep Shome</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-775</link>
		<dc:creator>Shankhadeep Shome</dc:creator>
		<pubDate>Tue, 30 Mar 2010 06:24:06 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-775</guid>
		<description>You might also look at the very mature but politically charged SCST project for Linux. It is extremely powerful, much like COMSTAR.</description>
		<content:encoded><![CDATA[<p>You might also look at the very mature but politically charged SCST project for Linux. It is extremely powerful, much like COMSTAR.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-632</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Thu, 04 Feb 2010 18:21:12 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-632</guid>
		<description>Thank-you.

You most definitely should use COMSTAR to provision your iSCSI targets. It&#039;s a kernel based implementation and runs almost twice the speed of iscsitgt.

To layout your disks I think RaidZ2 is a good choice for archive, I would use 4 groups of 7 and 2 groups of 8 this would leave 2 boot drives and 2 spares. You do not want very large groups as it increases the possibility of a total array failure event. Remember the rule of thumb &quot; its not if your disks fails&quot;,&quot;its when your disks fails&quot;.

Regards,

Mike</description>
		<content:encoded><![CDATA[<p>Thank-you.</p>
<p>You most definitely should use COMSTAR to provision your iSCSI targets. It&#8217;s a kernel based implementation and runs almost twice the speed of iscsitgt.</p>
<p>To layout your disks I think RaidZ2 is a good choice for archive, I would use 4 groups of 7 and 2 groups of 8 this would leave 2 boot drives and 2 spares. You do not want very large groups as it increases the possibility of a total array failure event. Remember the rule of thumb &#8221; its not if your disks fails&#8221;,&#8221;its when your disks fails&#8221;.</p>
<p>Regards,</p>
<p>Mike</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Eff Norwood</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-629</link>
		<dc:creator>Eff Norwood</dc:creator>
		<pubDate>Thu, 04 Feb 2010 13:53:58 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-629</guid>
		<description>Mike - thanks for another impressive article. I love your blog and thank you for it. I am currently getting ready to release for testing an OpenSolaris SNV 131 system for backup and VMWare, all provisioned as ZFS backed iSCSI. Would you advise using the traditional iSCSI methodology or the COMSTAR method outlined here?</description>
		<content:encoded><![CDATA[<p>Mike &#8211; thanks for another impressive article. I love your blog and thank you for it. I am currently getting ready to release for testing an OpenSolaris SNV 131 system for backup and VMWare, all provisioned as ZFS backed iSCSI. Would you advise using the traditional iSCSI methodology or the COMSTAR method outlined here?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-54</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Thu, 25 Jun 2009 12:55:56 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-54</guid>
		<description>If you are using it as a backup store then you will really like the deduplication feature that targeted for Q4/2009.</description>
		<content:encoded><![CDATA[<p>If you are using it as a backup store then you will really like the deduplication feature that targeted for Q4/2009.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Per</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-53</link>
		<dc:creator>Per</dc:creator>
		<pubDate>Thu, 25 Jun 2009 06:22:26 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-53</guid>
		<description>Really impressive, I&#039;m using OpenSolaris as a backup store today to a VMware cluster so this was really enlightning.</description>
		<content:encoded><![CDATA[<p>Really impressive, I&#8217;m using OpenSolaris as a backup store today to a VMware cluster so this was really enlightning.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Curmudgeon</title>
		<link>http://blog.laspina.ca/ubiquitous/multi_protocol_storage_provisioning_with/comment-page-1#comment-10</link>
		<dc:creator>Curmudgeon</dc:creator>
		<pubDate>Tue, 16 Jun 2009 12:25:49 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=75#comment-10</guid>
		<description>Thank you Mike for a great lesson in OpenSolaris, greatly appreciated.</description>
		<content:encoded><![CDATA[<p>Thank you Mike for a great lesson in OpenSolaris, greatly appreciated.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

