<?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: Apache SSL oddity</title>
	<atom:link href="http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/feed/" rel="self" type="application/rss+xml" />
	<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/</link>
	<description>Sven's occasional log</description>
	<lastBuildDate>Sat, 07 Jan 2012 19:52:43 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.0.5</generator>
	<item>
		<title>By: hynek</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2912</link>
		<dc:creator>hynek</dc:creator>
		<pubDate>Wed, 06 Oct 2010 10:11:09 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2912</guid>
		<description>Thank you sven. I was struggling with that for alost a day, until I found you article. 
I&#039;ve got server in DMZ, which is pure HTTPS and has the different domain names to Internet and to internal network, meaning two different certs also.

Appreciate it very much, now working for me correctly.
Thanks, Hynek</description>
		<content:encoded><![CDATA[<p>Thank you sven. I was struggling with that for alost a day, until I found you article.<br />
I&#8217;ve got server in DMZ, which is pure HTTPS and has the different domain names to Internet and to internal network, meaning two different certs also.</p>
<p>Appreciate it very much, now working for me correctly.<br />
Thanks, Hynek</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: sven</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2591</link>
		<dc:creator>sven</dc:creator>
		<pubDate>Sat, 28 Mar 2009 12:24:39 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2591</guid>
		<description>Was I really this misunderstandable? I was neither talking about name-based virtual hosts (mine are purely IP/port based) nor about SSL server name indication!
I will not approve any more comments which talk about those unless they show how they are related to my problem (IP/port based virtual hosts with the same ServerName).</description>
		<content:encoded><![CDATA[<p>Was I really this misunderstandable? I was neither talking about name-based virtual hosts (mine are purely IP/port based) nor about SSL server name indication!<br />
I will not approve any more comments which talk about those unless they show how they are related to my problem (IP/port based virtual hosts with the same ServerName).</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: pietro</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2590</link>
		<dc:creator>pietro</dc:creator>
		<pubDate>Sat, 28 Mar 2009 10:03:31 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2590</guid>
		<description>I wrote about sni a while ago :
https://mancoosi.org/~abate/namebased-virtual-hosting-with-ssl

but it still doesn&#039;t work properly with many browsers...</description>
		<content:encoded><![CDATA[<p>I wrote about sni a while ago :<br />
<a href="https://mancoosi.org/~abate/namebased-virtual-hosting-with-ssl">https://mancoosi.org/~abate/namebased-virtual-hosting-with-ssl</a></p>
<p>but it still doesn&#8217;t work properly with many browsers&#8230;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: goundoulf@gmail.com</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2588</link>
		<dc:creator>goundoulf@gmail.com</dc:creator>
		<pubDate>Thu, 26 Mar 2009 22:36:07 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2588</guid>
		<description>Well yes it is supported by Apache and it is called Server Name Indication (&lt;a href=&quot;http://en.wikipedia.org/wiki/Server_Name_Indication&quot; rel=&quot;nofollow&quot;&gt;http://en.wikipedia.org/wiki/Server_Name_Indication&lt;/a&gt;), but the problem is that the client has to be at least IE 7 (and Vista, not XP), or Firefox &gt; 2.0, or Opera &gt; 8.0 or Google Chrome or Safari 3.2.1...

And a lot of visitors still use IE 7 and XP, or IE &lt; 7...</description>
		<content:encoded><![CDATA[<p>Well yes it is supported by Apache and it is called Server Name Indication (<a href="http://en.wikipedia.org/wiki/Server_Name_Indication">http://en.wikipedia.org/wiki/Server_Name_Indication</a>), but the problem is that the client has to be at least IE 7 (and Vista, not XP), or Firefox &gt; 2.0, or Opera &gt; 8.0 or Google Chrome or Safari 3.2.1&#8230;</p>
<p>And a lot of visitors still use IE 7 and XP, or IE &lt; 7&#8230;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: sven</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2587</link>
		<dc:creator>sven</dc:creator>
		<pubDate>Thu, 26 Mar 2009 22:27:48 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2587</guid>
		<description>Yes, there is the option to pass the intended server name in the SSL handshake, and some browsers support it (or are at least supposed to support that). However, this is not the issue at hand here, which is that Apache stores the SSL certificate and key based on ServerName and &lt;b&gt;not&lt;/b&gt; - as one might think - based on what is in the &lt;VirtualHost&gt; stanza.</description>
		<content:encoded><![CDATA[<p>Yes, there is the option to pass the intended server name in the SSL handshake, and some browsers support it (or are at least supposed to support that). However, this is not the issue at hand here, which is that Apache stores the SSL certificate and key based on ServerName and <b>not</b> &#8211; as one might think &#8211; based on what is in the &lt;VirtualHost&gt; stanza.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Michael Croes</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2586</link>
		<dc:creator>Michael Croes</dc:creator>
		<pubDate>Thu, 26 Mar 2009 18:42:09 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2586</guid>
		<description>Name-based virtual hosting *is* possible with apache (and ssl of course), but it&#039;s somewhat harder to configure and it seems people don&#039;t want you to know that it&#039;s possible. I happen to use cherokee, which will happily do name-based ssl virtual hosts. There&#039;s just one little issue, and that is that the client needs to support it (it needs to send the hostname in the ssl initiation or something), and some older versions of IE are not really good at this (of course). I can only suggest that you look for the solution yourself.
Regards,

Michael</description>
		<content:encoded><![CDATA[<p>Name-based virtual hosting *is* possible with apache (and ssl of course), but it&#8217;s somewhat harder to configure and it seems people don&#8217;t want you to know that it&#8217;s possible. I happen to use cherokee, which will happily do name-based ssl virtual hosts. There&#8217;s just one little issue, and that is that the client needs to support it (it needs to send the hostname in the ssl initiation or something), and some older versions of IE are not really good at this (of course). I can only suggest that you look for the solution yourself.<br />
Regards,</p>
<p>Michael</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: sven</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2585</link>
		<dc:creator>sven</dc:creator>
		<pubDate>Thu, 26 Mar 2009 18:33:30 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2585</guid>
		<description>Well, I admit that I didn&#039;t check my post for formatting before posting it, so the &quot;VirtualHost&quot; statements were missing, but the fact that I talked about the &quot;same ServerName&quot; was a giveaway in my opinion:
I was talking about port based virtual hosts (which work in principal, see my workaround) that share the same hostname (but use different ports), as opposed to name-based virtual hosts which share the same IP and port, but differ in ServerName (well, the VirtualHost section should also have the ServerName in it).

I&#039;m quite well understanding that SSL happens before Apache sees any data from the client (apart from IP address and portnumber) that identifies the server it wants to speak to. So the only way to do shared IP/port name-based virtual servers is to also share the SSL certificate/key pair (which will either have to be a wildcard certificate or a certificate listing all used server names in the SubjectAltNames x509 property).

But, as said, this was not what we needed. We needed to have the same IP, but different ports (which makes it possible to use different SSL key/certificate pairs, just like it makes it possible to use completely different applications). And on those different ports, we needed different key/certificate pairs, which is, as said, quite possible. But in Apache, you not only need different (non-overlapping) VirtualHost sections (each specifying their own SSLCertificateFile/SSLCertificateKeyFile), but these sections also needed to use different ServerName statements.</description>
		<content:encoded><![CDATA[<p>Well, I admit that I didn&#8217;t check my post for formatting before posting it, so the &#8220;VirtualHost&#8221; statements were missing, but the fact that I talked about the &#8220;same ServerName&#8221; was a giveaway in my opinion:<br />
I was talking about port based virtual hosts (which work in principal, see my workaround) that share the same hostname (but use different ports), as opposed to name-based virtual hosts which share the same IP and port, but differ in ServerName (well, the VirtualHost section should also have the ServerName in it).</p>
<p>I&#8217;m quite well understanding that SSL happens before Apache sees any data from the client (apart from IP address and portnumber) that identifies the server it wants to speak to. So the only way to do shared IP/port name-based virtual servers is to also share the SSL certificate/key pair (which will either have to be a wildcard certificate or a certificate listing all used server names in the SubjectAltNames x509 property).</p>
<p>But, as said, this was not what we needed. We needed to have the same IP, but different ports (which makes it possible to use different SSL key/certificate pairs, just like it makes it possible to use completely different applications). And on those different ports, we needed different key/certificate pairs, which is, as said, quite possible. But in Apache, you not only need different (non-overlapping) VirtualHost sections (each specifying their own SSLCertificateFile/SSLCertificateKeyFile), but these sections also needed to use different ServerName statements.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: goundoulf</title>
		<link>http://blog.incase.de/index.php/2009/03/26/apache-ssl-oddity/comment-page-1/#comment-2584</link>
		<dc:creator>goundoulf</dc:creator>
		<pubDate>Thu, 26 Mar 2009 16:34:07 +0000</pubDate>
		<guid isPermaLink="false">http://blog.incase.de/?p=65#comment-2584</guid>
		<description>This is normal, for this to work you have to use IP-based virtual hosting instead.

With name-based virtual hosting, Apache has no way to know which certificate to use, as the SSL transaction starts before Apache gets to know the virtual host name.</description>
		<content:encoded><![CDATA[<p>This is normal, for this to work you have to use IP-based virtual hosting instead.</p>
<p>With name-based virtual hosting, Apache has no way to know which certificate to use, as the SSL transaction starts before Apache gets to know the virtual host name.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

