1
1
<?xml version="1.0" encoding="UTF-8"?>
2
2
<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
3
3
"http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd" [
4
<!ENTITY TIMESTAMP "2016-11-27">
4
<!ENTITY TIMESTAMP "2020-09-16">
5
5
<!ENTITY % common SYSTEM "common.ent">
65
68
The computers run a small client program in the initial RAM disk
66
69
environment which will communicate with a server over a network.
67
70
All network communication is encrypted using TLS. The clients
68
are identified by the server using an OpenPGP key; each client
71
are identified by the server using a TLS public key; each client
69
72
has one unique to it. The server sends the clients an encrypted
70
73
password. The encrypted password is decrypted by the clients
71
using the same OpenPGP key, and the password is then used to
74
using a separate OpenPGP key, and the password is then used to
72
75
unlock the root file system, whereupon the computers can
73
76
continue booting normally.
130
133
So, at boot time, the Mandos client will ask for its encrypted
131
data over the network, decrypt it to get the password, use it to
132
decrypt the root file, and continue booting.
134
data over the network, decrypt the data to get the password, use
135
the password to decrypt the root file system, and the client can
136
then continue booting.
135
139
Now, of course the initial RAM disk image is not on the
141
145
long, and will no longer give out the encrypted key. The timing
142
146
here is the only real weak point, and the method, frequency and
143
147
timeout of the server’s checking can be adjusted to any desired
147
151
(The encrypted keys on the Mandos server is on its normal file
199
203
No. The server only gives out the passwords to clients which
200
204
have <emphasis>in the TLS handshake</emphasis> proven that
201
they do indeed hold the OpenPGP private key corresponding to
205
they do indeed hold the private key corresponding to that
380
384
plugin requirements.
388
<refsect1 id="systemd">
389
<title>SYSTEMD</title>
391
More advanced startup systems like <citerefentry><refentrytitle
392
>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
393
already have their own plugin-like mechanisms for allowing
394
multiple agents to independently retrieve a password and deliver
395
it to the subsystem requesting a password to unlock the root
396
file system. On these systems, it would make no sense to run
397
<citerefentry><refentrytitle>plugin-runner</refentrytitle
398
><manvolnum>8mandos</manvolnum></citerefentry>, the plugins of
399
which would largely duplicate the work of (and conflict with)
400
the existing systems prompting for passwords.
403
As for <citerefentry><refentrytitle>systemd</refentrytitle
404
><manvolnum>1</manvolnum></citerefentry> in particular, it has
406
url="https://systemd.io/PASSWORD_AGENTS/">Password
407
Agents</ulink> system. Mandos uses this via its
408
<citerefentry><refentrytitle>password-agent</refentrytitle
409
><manvolnum>8mandos</manvolnum></citerefentry> program, which is
410
run instead of <citerefentry><refentrytitle
411
>plugin-runner</refentrytitle><manvolnum>8mandos</manvolnum
412
></citerefentry> when <citerefentry><refentrytitle
413
>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>
414
is used during system startup.
384
417
<refsect1 id="bugs">
385
418
<title>BUGS</title>
386
419
<xi:include href="bugs.xml"/>
401
434
<manvolnum>8</manvolnum></citerefentry>,
402
435
<citerefentry><refentrytitle>plugin-runner</refentrytitle>
403
436
<manvolnum>8mandos</manvolnum></citerefentry>,
437
<citerefentry><refentrytitle>password-agent</refentrytitle>
438
<manvolnum>8mandos</manvolnum></citerefentry>,
404
439
<citerefentry><refentrytitle>mandos-client</refentrytitle>
405
440
<manvolnum>8mandos</manvolnum></citerefentry>,
406
441
<citerefentry><refentrytitle>password-prompt</refentrytitle>