shorewall_code/docs/LXC.xml

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.4//EN"
"http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd">
<article>
  <!--$Id$-->

  <articleinfo>
    <title>LXC and Shorewall</title>

    <authorgroup>
      <author>
        <firstname>Tom</firstname>

        <surname>Eastep</surname>
      </author>
    </authorgroup>

    <pubdate><?dbtimestamp format="Y/m/d"?></pubdate>

    <copyright>
      <year>2011</year>

      <holder>Thomas M. Eastep</holder>
    </copyright>

    <legalnotice>
      <para>Permission is granted to copy, distribute and/or modify this
      document under the terms of the GNU Free Documentation License, Version
      1.2 or any later version published by the Free Software Foundation; with
      no Invariant Sections, with no Front-Cover, and with no Back-Cover
      Texts. A copy of the license is included in the section entitled
      <quote><ulink url="GnuCopyright.htm">GNU Free Documentation
      License</ulink></quote>.</para>
    </legalnotice>
  </articleinfo>

  <section>
    <title>Background</title>

    <para>LXC (<ulink
    url="http://lxc.sourceforge.net/">http://lxc.sourceforge.net/</ulink>) is
    a set of user-space tools for managing the container capabilities that
    have been in the Linux Kernel since 2.6.27.</para>

    <para>This short article describes how I've implemented LXC here at
    shorewall.net, with emphasis on the networking and firewall
    aspects.</para>
  </section>

  <section>
    <title>Overview of a Working Configuration</title>

    <para>The following diagram shows the network at shorewall.net in the
    spring of 2011.</para>

    <graphic align="center" fileref="images/Network2011a.png" />

    <para>As shown in that diagram, the LXC containers are bridged to br0.
    Here are the relevant configuration entries.</para>

    <para><filename>/etc/network/interfaces:</filename></para>

    <programlisting>#
# LXC bridge
#
auto br0
iface br0 inet static
      bridge_ports none
      bridge_fd 0
      address 70.90.191.121
      broadcast 0.0.0.0
      netmask 255.255.255.255
      post-up ip route add 70.90.191.124/31 dev br0

iface br0 inet6 static
      address 2001:470:b:227::41
      netmask 124
</programlisting>

    <para><filename>/etc/lxc/mail.conf</filename><programlisting>lxc.network.type=veth
lxc.network.link=br0
lxc.network.flags=up

lxc.network.ipv4=70.90.191.124/29
lxc.network.ipv6=2001:470:b:227::42/124

…</programlisting></para>

    <para><filename>/etc/lxc/server.conf</filename><programlisting>lxc.network.type=veth
lxc.network.link=br0
lxc.network.flags=up

lxc.network.ipv4=70.90.191.125/29
lxc.network.ipv6=2001:470:b:227::43/124

…</programlisting></para>

    <para>Note that I have subnetted 2001:470:b:227::/64 with a /124
    (2001:470:b:227::40/124) assigned to the bridge. To make those addresses
    accessible from the LOC zone, the following entries are required in
    /etc/shorewall6/proxyndp:</para>

    <programlisting>#ADDRESS		INTERFACE	EXTERNAL	HAVEROUTE	PERSISTENT
2001:470:b:227::41	-		eth1		Yes		Yes
2001:470:b:227::42	-		eth1		Yes		Yes
2001:470:b:227::43	-		eth1		Yes		Yes
</programlisting>

    <para>The entries in the LXC .conf files are expected to configure eth0 in
    the LXC containers; they do, <emphasis>sort of</emphasis>. In both of the
    containers, no ipv6 default route was assigned. I corrected that by adding
    this entry in <filename>/etc/sysctl.conf</filename> in both
    containers:</para>

    <programlisting>net.ipv6.conf.all.forwarding=0
</programlisting>

    <para>I then added this stanza to <filename>/etc/radvd.conf</filename> on
    the host:</para>

    <programlisting>interface br0{
	AdvSendAdvert on;
	MinRtrAdvInterval 300;
	MaxRtrAdvInterval 505;
	AdvDefaultLifetime 9000;

	route ::/0 {
		AdvRouteLifetime infinity;
	};
};
</programlisting>

    <para>Curiosly, LXC gives container mail's eth0 this somewhat odd
    configuration, and fails to add a default ipv4 route:</para>

    <programlisting>14: eth0: &lt;BROADCAST,MULTICAST,UP,LOWER_UP&gt; mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 4e:56:66:11:3c:6b brd ff:ff:ff:ff:ff:ff
    inet 70.90.191.124/29 <emphasis role="bold">brd 70.90.191.120</emphasis> scope global eth0
    inet6 2001:470:b:227::42/124 scope global 
       valid_lft forever preferred_lft forever
    inet6 fe80::4c56:66ff:fe11:3c6b/64 scope link 
       valid_lft forever preferred_lft forever
</programlisting>

    <para>So in that container's<filename> /etc/rc.local</filename>, I also
    have:</para>

    <programlisting>ip route add default via 70.90.191.121</programlisting>

    <para>With the exception of the entries in
    <filename>/etc/shorewall6/proxyndp</filename>. the Shorewall and
    Shorewall6 configurations are fairly conventional three-interface setups.
    In both configurations, the <filename>interfaces</filename> file entry for
    br0 has the <option>routeback</option> option specified.</para>
  </section>
</article>