<?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: Provisioning Disaster Recovery with ZFS, iSCSI and VMware</title>
	<atom:link href="http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/feed" rel="self" type="application/rss+xml" />
	<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs</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: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-1427</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Thu, 07 Oct 2010 12:27:00 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-1427</guid>
		<description>John,

If you issued shareiscsi=on this action would use the iscsitgtd service which is not COMSTAR and is now deprecated. The process to make it show up on the replicated target is documented in http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs starting at the heading Export SMF iSCSI configuration. Basically you need to load the target system iscsitgtd service manifest properties from the source and then clone the zfs data to a new floder and then change the service manifest properties to point to it.

Regards,
Mike</description>
		<content:encoded><![CDATA[<p>John,</p>
<p>If you issued shareiscsi=on this action would use the iscsitgtd service which is not COMSTAR and is now deprecated. The process to make it show up on the replicated target is documented in <a href="http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs" rel="nofollow">http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs</a> starting at the heading Export SMF iSCSI configuration. Basically you need to load the target system iscsitgtd service manifest properties from the source and then clone the zfs data to a new floder and then change the service manifest properties to point to it.</p>
<p>Regards,<br />
Mike</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: jhn</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-1424</link>
		<dc:creator>jhn</dc:creator>
		<pubDate>Tue, 05 Oct 2010 18:00:52 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-1424</guid>
		<description>Hi Mike, thank you for you response.

I would guess I&#039;m using COMSTAR from the package inforation, I had just installed all the &quot;scsi&quot; packages from the GUI:

# pkginfo &#124; grep scsi
system      SUNWiscsidmr                    Sun iSCSI Data Mover (Root)
system      SUNWiscsidmu                    Sun iSCSI Data Mover (Usr)
system      SUNWiscsir                      Sun iSCSI Device Driver (root)
system      SUNWiscsitgtr                   Sun iSCSI Target (Root)
system      SUNWiscsitgtu                   Sun iSCSI Target (Usr)
system      SUNWiscsitr                     Sun iSCSI COMSTAR Port Provider (root)
system      SUNWiscsitu                     Sun iSCSI COMSTAR Port Provider
system      SUNWiscsiu                      Sun iSCSI Management Utilities (usr)
system      SUNWmpsvplr                     Sun MP API library for the scsi_vhci driver (Root)

But, all I did to get iSCSI working was:

# zfs set shareiscsi=on tank

The &quot;LUN&#039;s&quot; just inherit it from the pool:

# zfs get shareiscsi
NAME                            PROPERTY    VALUE       SOURCE
tank                            shareiscsi  on          local
tank/target01                   shareiscsi  on          inherited from tank
tank/target01@snapshot01        shareiscsi  on          inherited from tank
rpool                           shareiscsi  off         default
rpool/ROOT                      shareiscsi  off         default
rpool/ROOT/opensolaris          shareiscsi  off         default
rpool/ROOT/opensolaris@install  shareiscsi  off         default
rpool/dump                      shareiscsi  off         default
rpool/export                    shareiscsi  off         default
rpool/export/home               shareiscsi  off         default
rpool/export/home/user          shareiscsi  off         default
rpool/swap                      shareiscsi  off         default

I looked at the comment you sent and tried running the svccfg export command, but didn&#039;t see anything specific to my tagets:

# svccfg export -a stmf



  
    
    
    
      
    
    
      
        
      
    
    
      
        
      
    
    
    
    
      
    
    
      
    
    
      
    
    
    
      
        STMF
      
      
        
        
      
    
  


Do I need to do something more than just setting shareiscsi?

Thank you and kind regards.</description>
		<content:encoded><![CDATA[<p>Hi Mike, thank you for you response.</p>
<p>I would guess I&#8217;m using COMSTAR from the package inforation, I had just installed all the &#8220;scsi&#8221; packages from the GUI:</p>
<p># pkginfo | grep scsi<br />
system      SUNWiscsidmr                    Sun iSCSI Data Mover (Root)<br />
system      SUNWiscsidmu                    Sun iSCSI Data Mover (Usr)<br />
system      SUNWiscsir                      Sun iSCSI Device Driver (root)<br />
system      SUNWiscsitgtr                   Sun iSCSI Target (Root)<br />
system      SUNWiscsitgtu                   Sun iSCSI Target (Usr)<br />
system      SUNWiscsitr                     Sun iSCSI COMSTAR Port Provider (root)<br />
system      SUNWiscsitu                     Sun iSCSI COMSTAR Port Provider<br />
system      SUNWiscsiu                      Sun iSCSI Management Utilities (usr)<br />
system      SUNWmpsvplr                     Sun MP API library for the scsi_vhci driver (Root)</p>
<p>But, all I did to get iSCSI working was:</p>
<p># zfs set shareiscsi=on tank</p>
<p>The &#8220;LUN&#8217;s&#8221; just inherit it from the pool:</p>
<p># zfs get shareiscsi<br />
NAME                            PROPERTY    VALUE       SOURCE<br />
tank                            shareiscsi  on          local<br />
tank/target01                   shareiscsi  on          inherited from tank<br />
tank/target01@snapshot01        shareiscsi  on          inherited from tank<br />
rpool                           shareiscsi  off         default<br />
rpool/ROOT                      shareiscsi  off         default<br />
rpool/ROOT/opensolaris          shareiscsi  off         default<br />
rpool/ROOT/opensolaris@install  shareiscsi  off         default<br />
rpool/dump                      shareiscsi  off         default<br />
rpool/export                    shareiscsi  off         default<br />
rpool/export/home               shareiscsi  off         default<br />
rpool/export/home/user          shareiscsi  off         default<br />
rpool/swap                      shareiscsi  off         default</p>
<p>I looked at the comment you sent and tried running the svccfg export command, but didn&#8217;t see anything specific to my tagets:</p>
<p># svccfg export -a stmf</p>
<p>        STMF</p>
<p>Do I need to do something more than just setting shareiscsi?</p>
<p>Thank you and kind regards.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-1416</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Sat, 02 Oct 2010 04:26:32 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-1416</guid>
		<description>Hi John,

You have not indicated what iSCSI target your running but I suspect it&#039;s COMSTAR. If that&#039;s the case have a look at this comment and see if it&#039;s what your looking for. http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-789

Regards,

Mike</description>
		<content:encoded><![CDATA[<p>Hi John,</p>
<p>You have not indicated what iSCSI target your running but I suspect it&#8217;s COMSTAR. If that&#8217;s the case have a look at this comment and see if it&#8217;s what your looking for. <a href="http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-789" rel="nofollow">http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-789</a></p>
<p>Regards,</p>
<p>Mike</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: jhn</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-1415</link>
		<dc:creator>jhn</dc:creator>
		<pubDate>Sat, 02 Oct 2010 00:48:24 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-1415</guid>
		<description>have been using OpenSolaris 5,11 ZFS to create iSCSI LUN for our ESX servers. I am adding another OpenSolaris 5.11 box for replication purposes. Security is not an issue, so I&#039;m using netcat (nc). 

I have tried this on the source server:
zfs create -V300GB tank/target01

[I attach target01 to esx server as an iSCSI LUN &amp; create VM in LUN ]

zfs snapshot tank/target01@snapshot01
zfs send -R tank/target01@snapshot01 &#124; nc san01 3141

And, this on the destination server:
nc -l -p 3141 -vvv &#124; zfs recv -vd tank

However, the destination LUN does not show up when I rescan the iSCSI host bus adapter.

What am I doing wrong? How should I be doing this? Thank you.</description>
		<content:encoded><![CDATA[<p>have been using OpenSolaris 5,11 ZFS to create iSCSI LUN for our ESX servers. I am adding another OpenSolaris 5.11 box for replication purposes. Security is not an issue, so I&#8217;m using netcat (nc). </p>
<p>I have tried this on the source server:<br />
zfs create -V300GB tank/target01</p>
<p>[I attach target01 to esx server as an iSCSI LUN &amp; create VM in LUN ]</p>
<p>zfs snapshot tank/target01@snapshot01<br />
zfs send -R tank/target01@snapshot01 | nc san01 3141</p>
<p>And, this on the destination server:<br />
nc -l -p 3141 -vvv | zfs recv -vd tank</p>
<p>However, the destination LUN does not show up when I rescan the iSCSI host bus adapter.</p>
<p>What am I doing wrong? How should I be doing this? Thank you.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-802</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Fri, 02 Apr 2010 20:46:18 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-802</guid>
		<description>Maxim,

If you explore /etc/security/auth_attr you will find all the security descriptions and attributes.

e.g.
solaris.smf.read.stmf:::Read STMF Provider Private Data::help=SmfSTMFRead.html

Here is what you would grant zfsadm:

ss1:~#usermod -A solaris.smf.read.stmf zfsadm

Mike</description>
		<content:encoded><![CDATA[<p>Maxim,</p>
<p>If you explore /etc/security/auth_attr you will find all the security descriptions and attributes.</p>
<p>e.g.<br />
solaris.smf.read.stmf:::Read STMF Provider Private Data::help=SmfSTMFRead.html</p>
<p>Here is what you would grant zfsadm:</p>
<p>ss1:~#usermod -A solaris.smf.read.stmf zfsadm</p>
<p>Mike</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: m0ps</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-798</link>
		<dc:creator>m0ps</dc:creator>
		<pubDate>Fri, 02 Apr 2010 17:24:44 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-798</guid>
		<description>I solved this problem. Just added -F option to second zfs recv instance in bash script. But then there was another problem: can not get a snapshot, due to the fact that zvol is busy. It was solved in the following way: stop stmf and iscsi/target services and rebooting OS. Not exactly a nice way, but I hope that I never do it :).
I still could not find solution for problem with exporting stmf configuration with -a option. Problem with user permissions, how can I grant zfsadm user do it?
Thanks,
Maxim</description>
		<content:encoded><![CDATA[<p>I solved this problem. Just added -F option to second zfs recv instance in bash script. But then there was another problem: can not get a snapshot, due to the fact that zvol is busy. It was solved in the following way: stop stmf and iscsi/target services and rebooting OS. Not exactly a nice way, but I hope that I never do it <img src='http://blog.laspina.ca/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> .<br />
I still could not find solution for problem with exporting stmf configuration with -a option. Problem with user permissions, how can I grant zfsadm user do it?<br />
Thanks,<br />
Maxim</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-797</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Fri, 02 Apr 2010 16:51:38 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-797</guid>
		<description>Maxim,

I&#039;m sorry, I missed an important detail, you should only activate (import) the LUNs using a cloned LUN other wise the replication stream will be broken.
You will need to issue a manual send receive to correct the modified state or possibly redo the entire send by deleting it on the target.
Locate the last successful snap on the source (ss1) and issue the following.

ss1:~# pfexec /sbin/zfs send -R sp1/san/volx@time &#124; ssh zfsadm@ss2 pfexec /sbin/zfs receive -F sp1/san/volx

Regard,

Mike</description>
		<content:encoded><![CDATA[<p>Maxim,</p>
<p>I&#8217;m sorry, I missed an important detail, you should only activate (import) the LUNs using a cloned LUN other wise the replication stream will be broken.<br />
You will need to issue a manual send receive to correct the modified state or possibly redo the entire send by deleting it on the target.<br />
Locate the last successful snap on the source (ss1) and issue the following.</p>
<p>ss1:~# pfexec /sbin/zfs send -R sp1/san/volx@time | ssh zfsadm@ss2 pfexec /sbin/zfs receive -F sp1/san/volx</p>
<p>Regard,</p>
<p>Mike</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: m0ps</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-794</link>
		<dc:creator>m0ps</dc:creator>
		<pubDate>Fri, 02 Apr 2010 08:32:59 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-794</guid>
		<description>sory... message is:
cannot receive incremental stream: destination DATA/iSCSI/lun0 has been modified since most recent snapshot</description>
		<content:encoded><![CDATA[<p>sory&#8230; message is:<br />
cannot receive incremental stream: destination DATA/iSCSI/lun0 has been modified since most recent snapshot</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: m0ps</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-793</link>
		<dc:creator>m0ps</dc:creator>
		<pubDate>Fri, 02 Apr 2010 08:00:58 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-793</guid>
		<description>Thanks for your reply Mike!
I try your solution and it work, but i faced with problem:
after enabling of stmf and  iscsi/target:default on ss2 I can&#039;t continue replication of zvol&#039;s from ss1 to ss2. Script output is &quot;cannot receive: failed to read from stream&quot;
Also I can&#039;t do svccfg export -a stmf &gt; stmf.cfg from zfsadm user (but svccfg export stmf &gt; stmf.cfg works fine).</description>
		<content:encoded><![CDATA[<p>Thanks for your reply Mike!<br />
I try your solution and it work, but i faced with problem:<br />
after enabling of stmf and  iscsi/target:default on ss2 I can&#8217;t continue replication of zvol&#8217;s from ss1 to ss2. Script output is &#8220;cannot receive: failed to read from stream&#8221;<br />
Also I can&#8217;t do svccfg export -a stmf &gt; stmf.cfg from zfsadm user (but svccfg export stmf &gt; stmf.cfg works fine).</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mike La Spina</title>
		<link>http://blog.laspina.ca/ubiquitous/provisioning_disaster_recovery_with_zfs/comment-page-1#comment-789</link>
		<dc:creator>Mike La Spina</dc:creator>
		<pubDate>Thu, 01 Apr 2010 19:41:38 +0000</pubDate>
		<guid isPermaLink="false">http://ux1.laspina.ca/?p=57#comment-789</guid>
		<description>Mi Maxim,

You have discovered that COMSTAR is very different.

Here is what you need to do:

Define all your host groups, initiators, targets etc. for both ss1 and ss2 on ss1.

Use the following to export the stmf service config to a file:

&lt;strong&gt;ss1:~# svccfg export -a stmf &gt; stmf.cfg&lt;/strong&gt;

Copy the config file to ss2 and perform the following:

&lt;strong&gt;ss2:~# svcadm disable stmf
ss2:~#  svccfg delete stmf
ss2:~#  svccfg import stmf.cfg
ss2:~#  svcadm refresh stmf
ss2:~#  svcadm enable stmf&lt;/strong&gt;

Once this is complete you will need to import the LU&#039;s (the last time I worked on this task was on snv_126 so this part may not be required any more):

&lt;strong&gt;ss2:~# stmfadm import-lu /dev/zvol/rdsk/sp1/san/vol0&lt;/strong&gt;

And there you go a completely migrated COMSTAR config and LUs.

Enjoy.

Regards,

Mike</description>
		<content:encoded><![CDATA[<p>Mi Maxim,</p>
<p>You have discovered that COMSTAR is very different.</p>
<p>Here is what you need to do:</p>
<p>Define all your host groups, initiators, targets etc. for both ss1 and ss2 on ss1.</p>
<p>Use the following to export the stmf service config to a file:</p>
<p><strong>ss1:~# svccfg export -a stmf > stmf.cfg</strong></p>
<p>Copy the config file to ss2 and perform the following:</p>
<p><strong>ss2:~# svcadm disable stmf<br />
ss2:~#  svccfg delete stmf<br />
ss2:~#  svccfg import stmf.cfg<br />
ss2:~#  svcadm refresh stmf<br />
ss2:~#  svcadm enable stmf</strong></p>
<p>Once this is complete you will need to import the LU&#8217;s (the last time I worked on this task was on snv_126 so this part may not be required any more):</p>
<p><strong>ss2:~# stmfadm import-lu /dev/zvol/rdsk/sp1/san/vol0</strong></p>
<p>And there you go a completely migrated COMSTAR config and LUs.</p>
<p>Enjoy.</p>
<p>Regards,</p>
<p>Mike</p>
]]></content:encoded>
	</item>
</channel>
</rss>

