/mandos/trunk

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"

"http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd" [

<!ENTITY TIMESTAMP "2020-09-16">

<!ENTITY % common SYSTEM "common.ent">

%common;

]>

<refentry xmlns:xi="http://www.w3.org/2001/XInclude">

   <refentryinfo>

    <title>Mandos Manual</title>

    <!-- NWalsh’s docbook scripts use this to generate the footer: -->

    <productname>Mandos</productname>

    <productnumber>&version;</productnumber>

    <date>&TIMESTAMP;</date>

    <authorgroup>

      <author>

	<firstname>Björn</firstname>

	<surname>Påhlsson</surname>

	<address>

	  <email>belorn@recompile.se</email>

	</address>

      </author>

      <author>

	<firstname>Teddy</firstname>

	<surname>Hogeborn</surname>

	<address>

	  <email>teddy@recompile.se</email>

	</address>

      </author>

    </authorgroup>

    <copyright>

      <year>2011</year>

      <year>2012</year>

      <year>2013</year>

      <year>2014</year>

      <year>2015</year>

      <year>2016</year>

      <year>2017</year>

      <year>2018</year>

      <year>2019</year>

      <year>2020</year>

      <holder>Teddy Hogeborn</holder>

      <holder>Björn Påhlsson</holder>

    </copyright>

    <xi:include href="legalnotice.xml"/>

  </refentryinfo>

  <refmeta>

    <refentrytitle>intro</refentrytitle>

    <manvolnum>8mandos</manvolnum>

  </refmeta>

  <refnamediv>

    <refname>intro</refname>

    <refpurpose>

      Introduction to the Mandos system

    </refpurpose>

  </refnamediv>

  <refsect1 id="description">

    <title>DESCRIPTION</title>

    <para>

      This is the the Mandos system, which allows computers to have

      encrypted root file systems and at the same time be capable of

      remote and/or unattended reboots.

    </para>

    <para>

      The computers run a small client program in the initial RAM disk

      environment which will communicate with a server over a network.

      All network communication is encrypted using TLS.  The clients

      are identified by the server using a TLS public key; each client

      has one unique to it.  The server sends the clients an encrypted

      password.  The encrypted password is decrypted by the clients

      using a separate OpenPGP key, and the password is then used to

      unlock the root file system, whereupon the computers can

      continue booting normally.

    </para>

  </refsect1>

  <refsect1 id="introduction">

    <title>INTRODUCTION</title>

    <para>

      <!-- This paragraph is a combination and paraphrase of two

           quotes from the 1995 movie “The Usual Suspects”. -->

      You know how it is.  You’ve heard of it happening.  The Man

      comes and takes away your servers, your friends’ servers, the

      servers of everybody in the same hosting facility. The servers

      of their neighbors, and their neighbors’ friends.  The servers

      of people who owe them money.  And like

      <emphasis>that</emphasis>, they’re gone.  And you doubt you’ll

      ever see them again.

    </para>

    <para>

      That is why your servers have encrypted root file systems.

      However, there’s a downside.  There’s no going around it:

      rebooting is a pain.  Dragging out that rarely-used keyboard and

      screen and unraveling cables behind your servers to plug them in

      to type in that password is messy, especially if you have many

      servers.  There are some people who do clever things like using

      serial line consoles and daisy-chain it to the next server, and

      keep all the servers connected in a ring with serial cables,

      which will work, if your servers are physically close enough.

      There are also other out-of-band management solutions, but with

      <emphasis>all</emphasis> these, you still have to be on hand and

      manually type in the password at boot time.  Otherwise the

      server just sits there, waiting for a password.

    </para>

    <para>

      Wouldn’t it be great if you could have the security of encrypted

      root file systems and still have servers that could boot up

      automatically if there was a short power outage while you were

      asleep?  That you could reboot at will, without having someone

      run over to the server to type in the password?

    </para>

    <para>

      Well, with Mandos, you (almost) can!  The gain in convenience

      will only be offset by a small loss in security.  The setup is

      as follows:

    </para>

    <para>

      The server will still have its encrypted root file system.  The

      password to this file system will be stored on another computer

      (henceforth known as the Mandos server) on the same local

      network.  The password will <emphasis>not</emphasis> be stored

      in plaintext, but encrypted with OpenPGP.  To decrypt this

      password, a key is needed.  This key (the Mandos client key)

      will not be stored there, but back on the original server

      (henceforth known as the Mandos client) in the initial RAM disk

      image.  Oh, and all network Mandos client/server communications

      will be encrypted, using TLS (SSL).

    </para>

    <para>

      So, at boot time, the Mandos client will ask for its encrypted

      data over the network, decrypt the data to get the password, use

      the password to decrypt the root file system, and the client can

      then continue booting.

    </para>

    <para>

      Now, of course the initial RAM disk image is not on the

      encrypted root file system, so anyone who had physical access

      could take the Mandos client computer offline and read the disk

      with their own tools to get the authentication keys used by a

      client.  <emphasis>But</emphasis>, by then the Mandos server

      should notice that the original server has been offline for too

      long, and will no longer give out the encrypted key.  The timing

      here is the only real weak point, and the method, frequency and

      timeout of the server’s checking can be adjusted to any desired

      level of paranoia.

    </para>

    <para>

      (The encrypted keys on the Mandos server is on its normal file

      system, so those are safe, provided the root file system of

      <emphasis>that</emphasis> server is encrypted.)

    </para>

  </refsect1>

  <refsect1 id="faq">

    <title>FREQUENTLY ASKED QUESTIONS</title>

    <para>

      Couldn’t the security be defeated by…

    </para>

    <refsect2 id="quick">

      <title>Grabbing the Mandos client key from the

      initrd <emphasis>really quickly</emphasis>?</title>

    <para>

      This, as mentioned above, is the only real weak point.  But if

      you set the timing values tight enough, this will be really

      difficult to do.  An attacker would have to physically

      disassemble the client computer, extract the key from the

      initial RAM disk image, and then connect to a <emphasis>still

      online</emphasis> Mandos server to get the encrypted key, and do

      all this <emphasis>before</emphasis> the Mandos server timeout

      kicks in and the Mandos server refuses to give out the key to

      anyone.

    </para>

    <para>

      Now, as the typical procedure seems to be to barge in and turn

      off and grab <emphasis>all</emphasis> computers, to maybe look

      at them months later, this is not likely.  If someone does that,

      the whole system <emphasis>will</emphasis> lock itself up

      completely, since Mandos servers are no longer running.

    </para>

    <para>

      For sophisticated attackers who <emphasis>could</emphasis> do

      the clever thing, <emphasis>and</emphasis> had physical access

      to the server for enough time, it would be simpler to get a key

      for an encrypted file system by using hardware memory scanners

      and reading it right off the memory bus.

    </para>

    </refsect2>

    <refsect2 id="replay">

      <title>Replay attacks?</title>

      <para>

	Nope, the network stuff is all done over TLS, which provides

	protection against that.

      </para>

    </refsect2>

    <refsect2 id="mitm">

      <title>Man-in-the-middle?</title>

      <para>

	No.  The server only gives out the passwords to clients which

	have <emphasis>in the TLS handshake</emphasis> proven that

	they do indeed hold the private key corresponding to that

	client.

      </para>

    </refsect2>

    <refsect2 id="sniff">

      <title>How about sniffing the network traffic and decrypting it

      later by physically grabbing the Mandos client and using its

      key?</title>

      <para>

	We only use <acronym>PFS</acronym> (Perfect Forward Security)

	key exchange algorithms in TLS, which protects against this.

      </para>

    </refsect2>

    <refsect2 id="physgrab">

      <title>Physically grabbing the Mandos server computer?</title>

      <para>

	You could protect <emphasis>that</emphasis> computer the

	old-fashioned way, with a must-type-in-the-password-at-boot

	method.  Or you could have two computers be the Mandos server

	for each other.

      </para>

      <para>

	Multiple Mandos servers can coexist on a network without any

	trouble.  They do not clash, and clients will try all

	available servers.  This means that if just one reboots then

	the other can bring it back up, but if both reboot at the same

	time they will stay down until someone types in the password

	on one of them.

      </para>

    </refsect2>

    <refsect2 id="fakecheck">

      <title>Faking checker results?</title>

      <para>

	If the Mandos client does not have an SSH server, the default

	is for the Mandos server to use

	<quote><literal>fping</literal></quote>, the replies to which

	could be faked to eliminate the timeout.  But this could

	easily be changed to any shell command, with any security

	measures you like.  If the Mandos client

	<emphasis>has</emphasis> an SSH server, the default

	configuration (as generated by

	<command>mandos-keygen</command> with the

	<option>--password</option> option) is for the Mandos server

	to use an <command>ssh-keyscan</command> command with strict

	keychecking, which can not be faked.  Alternatively, IPsec

	could be used for the ping packets, making them secure.

      </para>

    </refsect2>

  </refsect1>

  <refsect1 id="security">

    <title>SECURITY</title>

    <para>

      So, in summary:  The only weakness in the Mandos system is from

      people who have:

    </para>

    <orderedlist>

      <listitem>

	<para>

	  The power to come in and physically take your servers,

	  <emphasis>and</emphasis>

	</para>

      </listitem>

      <listitem>

	<para>

	  The cunning and patience to do it carefully, one at a time,

	  and <emphasis>quickly</emphasis>, faking Mandos

	  client/server responses for each one before the timeout.

	</para>

      </listitem>

    </orderedlist>

    <para>

      While there are some who may be threatened by people who have

      <emphasis>both</emphasis> these attributes, they do not,

      probably, constitute the majority.

    </para>

    <para>

      If you <emphasis>do</emphasis> face such opponents, you must

      figure that they could just as well open your servers and read

      the file system keys right off the memory by running wires to

      the memory bus.

    </para>

    <para>

      What Mandos is designed to protect against is

      <emphasis>not</emphasis> such determined, focused, and competent

      attacks, but against the early morning knock on your door and

      the sudden absence of all the servers in your server room.

      Which it does nicely.

    </para>

  </refsect1>

  <refsect1 id="plugins">

    <title>PLUGINS</title>

    <para>

      In the early designs, the

      <citerefentry><refentrytitle>mandos-client</refentrytitle

      ><manvolnum>8mandos</manvolnum></citerefentry> program (which

      retrieves a password from the Mandos server) also prompted for a

      password on the terminal, in case a Mandos server could not be

      found.  Other ways of retrieving a password could easily be

      envisoned, but this multiplicity of purpose was seen to be too

      complex to be a viable way to continue.  Instead, the original

      program was separated into <citerefentry><refentrytitle

      >mandos-client</refentrytitle><manvolnum>8mandos</manvolnum

      ></citerefentry> and <citerefentry><refentrytitle

      >password-prompt</refentrytitle><manvolnum>8mandos</manvolnum

      ></citerefentry>, and a <citerefentry><refentrytitle

      >plugin-runner</refentrytitle><manvolnum>8mandos</manvolnum

      ></citerefentry> exist to run them both in parallel, allowing

      the first successful plugin to provide the password.  This

      opened up for any number of additional plugins to run, all

      competing to be the first to find a password and provide it to

      the plugin runner.

    </para>

    <para>

      Four additional plugins are provided:

    </para>

    <variablelist>

      <varlistentry>

	<term>

	  <citerefentry><refentrytitle>plymouth</refentrytitle>

	  <manvolnum>8mandos</manvolnum></citerefentry>

	</term>

	<listitem>

	  <para>

	    This prompts for a password when using <citerefentry>

	    <refentrytitle>plymouth</refentrytitle><manvolnum

	    >8</manvolnum></citerefentry>.

	  </para>

	</listitem>

      </varlistentry>

      <varlistentry>

	<term>

	  <citerefentry><refentrytitle>usplash</refentrytitle>

	  <manvolnum>8mandos</manvolnum></citerefentry>

	</term>

	<listitem>

	  <para>

	    This prompts for a password when using <citerefentry>

	    <refentrytitle>usplash</refentrytitle><manvolnum

	    >8</manvolnum></citerefentry>.

	  </para>

	</listitem>

      </varlistentry>

      <varlistentry>

	<term>

	  <citerefentry><refentrytitle>splashy</refentrytitle>

	  <manvolnum>8mandos</manvolnum></citerefentry>

	</term>

	<listitem>

	  <para>

	    This prompts for a password when using <citerefentry>

	    <refentrytitle>splashy</refentrytitle><manvolnum

	    >8</manvolnum></citerefentry>.

	  </para>

	</listitem>

      </varlistentry>

      <varlistentry>

	<term>

	  <citerefentry><refentrytitle>askpass-fifo</refentrytitle>

	  <manvolnum>8mandos</manvolnum></citerefentry>

	</term>

	<listitem>

	  <para>

	    To provide compatibility with the "askpass" program from

	    cryptsetup, this plugin listens to the same FIFO as

	    askpass would do.

	  </para>

	</listitem>

      </varlistentry>

    </variablelist>

    <para>

      More plugins can easily be written and added by the system

      administrator; see the section called "WRITING PLUGINS" in

      <citerefentry><refentrytitle>plugin-runner</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry> to learn the

      plugin requirements.

    </para>

  </refsect1>

  <refsect1 id="systemd">

    <title>SYSTEMD</title>

    <para>

      More advanced startup systems like <citerefentry><refentrytitle

      >systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,

      already have their own plugin-like mechanisms for allowing

      multiple agents to independently retrieve a password and deliver

      it to the subsystem requesting a password to unlock the root

      file system.  On these systems, it would make no sense to run

      <citerefentry><refentrytitle>plugin-runner</refentrytitle

      ><manvolnum>8mandos</manvolnum></citerefentry>, the plugins of

      which would largely duplicate the work of (and conflict with)

      the existing systems prompting for passwords.

    </para>

    <para>

      As for <citerefentry><refentrytitle>systemd</refentrytitle

      ><manvolnum>1</manvolnum></citerefentry> in particular, it has

      its own <ulink

      url="https://systemd.io/PASSWORD_AGENTS/">Password

      Agents</ulink> system.  Mandos uses this via its

      <citerefentry><refentrytitle>password-agent</refentrytitle

      ><manvolnum>8mandos</manvolnum></citerefentry> program, which is

      run instead of <citerefentry><refentrytitle

      >plugin-runner</refentrytitle><manvolnum>8mandos</manvolnum

      ></citerefentry> when <citerefentry><refentrytitle

      >systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>

      is used during system startup.

    </para>

  </refsect1>

  <refsect1 id="bugs">

    <title>BUGS</title>

    <xi:include href="bugs.xml"/>

  </refsect1>

  <refsect1 id="see_also">

    <title>SEE ALSO</title>

    <para>

      <citerefentry><refentrytitle>mandos</refentrytitle>

      <manvolnum>8</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>mandos.conf</refentrytitle>

      <manvolnum>5</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>mandos-clients.conf</refentrytitle>

      <manvolnum>5</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>mandos-ctl</refentrytitle>

      <manvolnum>8</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>mandos-monitor</refentrytitle>

      <manvolnum>8</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>plugin-runner</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>password-agent</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>mandos-client</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>password-prompt</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>plymouth</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>usplash</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>splashy</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>askpass-fifo</refentrytitle>

      <manvolnum>8mandos</manvolnum></citerefentry>,

      <citerefentry><refentrytitle>mandos-keygen</refentrytitle>

      <manvolnum>8</manvolnum></citerefentry>

    </para>

    <variablelist>

      <varlistentry>

	<term>

	  <ulink url="https://www.recompile.se/mandos">Mandos</ulink>

	</term>

	<listitem>

	  <para>

	    The Mandos home page.

	  </para>

	</listitem>

      </varlistentry>

    </variablelist>

  </refsect1>

</refentry>

<!-- Local Variables: -->

<!-- time-stamp-start: "<!ENTITY TIMESTAMP [\"']" -->

<!-- time-stamp-end: "[\"']>" -->

<!-- time-stamp-format: "%:y-%02m-%02d" -->

<!-- End: -->