man/systemd-nspawn.xml

   1 <?xml version='1.0'?> <!--*-nxml-*-->
   2 <!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
   3         "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
   4
   5 <!--
   6   This file is part of systemd.
   7
   8   Copyright 2010 Lennart Poettering
   9
  10   systemd is free software; you can redistribute it and/or modify it
  11   under the terms of the GNU Lesser General Public License as published by
  12   the Free Software Foundation; either version 2.1 of the License, or
  13   (at your option) any later version.
  14
  15   systemd is distributed in the hope that it will be useful, but
  16   WITHOUT ANY WARRANTY; without even the implied warranty of
  17   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  18   Lesser General Public License for more details.
  19
  20   You should have received a copy of the GNU Lesser General Public License
  21   along with systemd; If not, see <http://www.gnu.org/licenses/>.
  22 -->
  23
  24 <refentry id="systemd-nspawn">
  25
  26         <refentryinfo>
  27                 <title>systemd-nspawn</title>
  28                 <productname>systemd</productname>
  29
  30                 <authorgroup>
  31                         <author>
  32                                 <contrib>Developer</contrib>
  33                                 <firstname>Lennart</firstname>
  34                                 <surname>Poettering</surname>
  35                                 <email>lennart@poettering.net</email>
  36                         </author>
  37                 </authorgroup>
  38         </refentryinfo>
  39
  40         <refmeta>
  41                 <refentrytitle>systemd-nspawn</refentrytitle>
  42                 <manvolnum>1</manvolnum>
  43         </refmeta>
  44
  45         <refnamediv>
  46                 <refname>systemd-nspawn</refname>
  47                 <refpurpose>Spawn a namespace container for debugging, testing and building</refpurpose>
  48         </refnamediv>
  49
  50         <refsynopsisdiv>
  51                 <cmdsynopsis>
  52                         <command>systemd-nspawn</command>
  53                         <arg choice="opt" rep="repeat">OPTIONS</arg>
  54                         <arg choice="opt">COMMAND</arg>
  55                         <arg choice="opt" rep="repeat">ARGS</arg>
  56                 </cmdsynopsis>
  57         </refsynopsisdiv>
  58
  59         <refsect1>
  60                 <title>Description</title>
  61
  62                 <para><command>systemd-nspawn</command> may be used to
  63                 run a command or OS in a light-weight namespace
  64                 container. In many ways it is similar to
  65                 <citerefentry><refentrytitle>chroot</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
  66                 but more powerful since it fully virtualizes the file
  67                 system hierarchy, as well as the process tree, the
  68                 various IPC subsystems and the host and domain
  69                 name.</para>
  70
  71                 <para><command>systemd-nspawn</command> limits access
  72                 to various kernel interfaces in the container to
  73                 read-only, such as <filename>/sys</filename>,
  74                 <filename>/proc/sys</filename> or
  75                 <filename>/sys/fs/selinux</filename>. Network
  76                 interfaces and the system clock may not be changed
  77                 from within the container. Device nodes may not be
  78                 created. The host system cannot be rebooted and kernel
  79                 modules may not be loaded from within the
  80                 container.</para>
  81
  82                 <para>Note that even though these security precautions
  83                 are taken <command>systemd-nspawn</command> is not
  84                 suitable for secure container setups. Many of the
  85                 security features may be circumvented and are hence
  86                 primarily useful to avoid accidental changes to the
  87                 host system from the container. The intended use of
  88                 this program is debugging and testing as well as
  89                 building of packages, distributions and software
  90                 involved with boot and systems management.</para>
  91
  92                 <para>In contrast to
  93                 <citerefentry><refentrytitle>chroot</refentrytitle><manvolnum>1</manvolnum></citerefentry>
  94                 <command>systemd-nspawn</command> may be used to boot
  95                 full Linux-based operating systems in a
  96                 container.</para>
  97
  98                 <para>Use a tool like
  99                 <citerefentry><refentrytitle>yum</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
 100                 <citerefentry><refentrytitle>debootstrap</refentrytitle><manvolnum>8</manvolnum></citerefentry>
 101                 or
 102                 <citerefentry><refentrytitle>pacman</refentrytitle><manvolnum>8</manvolnum></citerefentry>
 103                 to set up an OS directory tree suitable as file system
 104                 hierarchy for <command>systemd-nspawn</command>
 105                 containers.</para>
 106
 107                 <para>Note that <command>systemd-nspawn</command> will
 108                 mount file systems private to the container to
 109                 <filename>/dev</filename>,
 110                 <filename>/run</filename> and similar. These will
 111                 not be visible outside of the container, and their
 112                 contents will be lost when the container exits.</para>
 113
 114                 <para>Note that running two
 115                 <command>systemd-nspawn</command> containers from the
 116                 same directory tree will not make processes in them
 117                 see each other. The PID namespace separation of the
 118                 two containers is complete and the containers will
 119                 share very few runtime objects except for the
 120                 underlying file system. It is however possible to
 121                 enter an existing container, see
 122                 <link linkend='example-nsenter'>Example 4</link> below.
 123                 </para>
 124
 125                 <para><command>systemd-nspawn</command> implements the
 126                 <ulink
 127                 url="http://www.freedesktop.org/wiki/Software/systemd/ContainerInterface">Container
 128                 Interface</ulink> specification.</para>
 129         </refsect1>
 130
 131         <refsect1>
 132                 <title>Options</title>
 133
 134                 <para>If no arguments are passed the container is set
 135                 up and a shell started in it, otherwise the passed
 136                 command and arguments are executed in it. The
 137                 following options are understood:</para>
 138
 139                 <variablelist>
 140                         <varlistentry>
 141                                 <term><option>-h</option></term>
 142                                 <term><option>--help</option></term>
 143
 144                                 <listitem><para>Prints a short help
 145                                 text and exits.</para></listitem>
 146                         </varlistentry>
 147
 148                         <varlistentry>
 149                                 <term><option>--version</option></term>
 150
 151                                 <listitem><para>Prints a version string
 152                                 and exits.</para></listitem>
 153                         </varlistentry>
 154
 155                         <varlistentry>
 156                                 <term><option>-D</option></term>
 157                                 <term><option>--directory=</option></term>
 158
 159                                 <listitem><para>Directory to use as
 160                                 file system root for the namespace
 161                                 container. If omitted the current
 162                                 directory will be
 163                                 used.</para></listitem>
 164                         </varlistentry>
 165
 166                         <varlistentry>
 167                                 <term><option>-b</option></term>
 168                                 <term><option>--boot</option></term>
 169
 170                                 <listitem><para>Automatically search
 171                                 for an init binary and invoke it
 172                                 instead of a shell or a user supplied
 173                                 program.</para></listitem>
 174                         </varlistentry>
 175
 176                         <varlistentry>
 177                                 <term><option>-u</option></term>
 178                                 <term><option>--user=</option></term>
 179
 180                                 <listitem><para>Run the command
 181                                 under specified user, create home
 182                                 directory and cd into it. As rest
 183                                 of systemd-nspawn, this is not
 184                                 the security feature and limits
 185                                 against accidental changes only.
 186                                 </para></listitem>
 187                         </varlistentry>
 188
 189                         <varlistentry>
 190                                 <term><option>--uuid=</option></term>
 191
 192                                 <listitem><para>Set the specified uuid
 193                                 for the container. The init system
 194                                 will initialize
 195                                 <filename>/etc/machine-id</filename>
 196                                 from this if this file is not set yet.
 197                                 </para></listitem>
 198                         </varlistentry>
 199
 200                         <varlistentry>
 201                                 <term><option>-C</option></term>
 202                                 <term><option>--controllers=</option></term>
 203
 204                                 <listitem><para>Makes the container appear in
 205                                 other hierarchies than the name=systemd:/ one.
 206                                 Takes a comma-separated list of controllers.
 207                                 </para></listitem>
 208                         </varlistentry>
 209
 210                         <varlistentry>
 211                                 <term><option>--private-network</option></term>
 212
 213                                 <listitem><para>Turn off networking in
 214                                 the container. This makes all network
 215                                 interfaces unavailable in the
 216                                 container, with the exception of the
 217                                 loopback device.</para></listitem>
 218                         </varlistentry>
 219
 220                         <varlistentry>
 221                                 <term><option>--read-only</option></term>
 222
 223                                 <listitem><para>Mount the root file
 224                                 system read only for the
 225                                 container.</para></listitem>
 226                         </varlistentry>
 227
 228                         <varlistentry>
 229                                 <term><option>--capability=</option></term>
 230
 231                                 <listitem><para>List one or more
 232                                 additional capabilities to grant the
 233                                 container. Takes a comma separated
 234                                 list of capability names, see
 235                                 <citerefentry><refentrytitle>capabilities</refentrytitle><manvolnum>7</manvolnum></citerefentry>
 236                                 for more information. Note that the
 237                                 following capabilities will be granted
 238                                 in any way: CAP_CHOWN,
 239                                 CAP_DAC_OVERRIDE, CAP_DAC_READ_SEARCH,
 240                                 CAP_FOWNER, CAP_FSETID, CAP_IPC_OWNER,
 241                                 CAP_KILL, CAP_LEASE,
 242                                 CAP_LINUX_IMMUTABLE,
 243                                 CAP_NET_BIND_SERVICE,
 244                                 CAP_NET_BROADCAST, CAP_NET_RAW,
 245                                 CAP_SETGID, CAP_SETFCAP, CAP_SETPCAP,
 246                                 CAP_SETUID, CAP_SYS_ADMIN,
 247                                 CAP_SYS_CHROOT, CAP_SYS_NICE,
 248                                 CAP_SYS_PTRACE, CAP_SYS_TTY_CONFIG,
 249                                 CAP_SYS_RESOURCE, CAP_SYS_BOOT,
 250                                 CAP_AUDIT_WRITE,
 251                                 CAP_AUDIT_CONTROL.</para></listitem>
 252                         </varlistentry>
 253
 254                         <varlistentry>
 255                                 <term><option>--link-journal=</option></term>
 256
 257                                 <listitem><para>Control whether the
 258                                 container's journal shall be made
 259                                 visible to the host system. If enabled
 260                                 allows viewing the container's journal
 261                                 files from the host (but not vice
 262                                 versa). Takes one of
 263                                 <literal>no</literal>,
 264                                 <literal>host</literal>,
 265                                 <literal>guest</literal>,
 266                                 <literal>auto</literal>. If
 267                                 <literal>no</literal>, the journal is
 268                                 not linked. If <literal>host</literal>,
 269                                 the journal files are stored on the
 270                                 host file system (beneath
 271                                 <filename>/var/log/journal/<replaceable>machine-id</replaceable></filename>)
 272                                 and the subdirectory is bind-mounted
 273                                 into the container at the same
 274                                 location. If <literal>guest</literal>,
 275                                 the journal files are stored on the
 276                                 guest file system (beneath
 277                                 <filename>/var/log/journal/<replaceable>machine-id</replaceable></filename>)
 278                                 and the subdirectory is symlinked into the host
 279                                 at the same location. If
 280                                 <literal>auto</literal> (the default),
 281                                 and the right subdirectory of
 282                                 <filename>/var/log/journal</filename>
 283                                 exists, it will be bind mounted
 284                                 into the container. If the
 285                                 subdirectory doesn't exist, no
 286                                 linking is performed. Effectively,
 287                                 booting a container once with
 288                                 <literal>guest</literal> or
 289                                 <literal>host</literal> will link the
 290                                 journal persistently if further on
 291                                 the default of <literal>auto</literal>
 292                                 is used.</para></listitem>
 293                         </varlistentry>
 294
 295                         <varlistentry>
 296                                 <term><option>-j</option></term>
 297
 298                                 <listitem><para>Equivalent to
 299                                 <option>--link-journal=guest</option>.</para></listitem>
 300                         </varlistentry>
 301                 </variablelist>
 302
 303         </refsect1>
 304
 305         <refsect1>
 306                 <title>Example 1</title>
 307
 308                 <programlisting># yum -y --releasever=19 --nogpg --installroot=/srv/mycontainer --disablerepo='*' --enablerepo=fedora install systemd passwd yum fedora-release vim-minimal
 309 # systemd-nspawn -bD /srv/mycontainer</programlisting>
 310
 311                 <para>This installs a minimal Fedora distribution into
 312                 the directory <filename>/srv/mycontainer/</filename> and
 313                 then boots an OS in a namespace container in
 314                 it.</para>
 315         </refsect1>
 316
 317         <refsect1>
 318                 <title>Example 2</title>
 319
 320                 <programlisting># debootstrap --arch=amd64 unstable ~/debian-tree/
 321 # systemd-nspawn -D ~/debian-tree/</programlisting>
 322
 323                 <para>This installs a minimal Debian unstable
 324                 distribution into the directory
 325                 <filename>~/debian-tree/</filename> and then spawns a
 326                 shell in a namespace container in it.</para>
 327         </refsect1>
 328
 329         <refsect1>
 330                 <title>Example 3</title>
 331
 332                 <programlisting># pacstrap -c -d ~/arch-tree/ base
 333 # systemd-nspawn -bD ~/arch-tree/</programlisting>
 334
 335                 <para>This installs a mimimal Arch Linux distribution into
 336                 the directory <filename>~/arch-tree/</filename> and then
 337                 boots an OS in a namespace container in it.</para>
 338         </refsect1>
 339
 340         <refsect1 id='example-nsenter'>
 341                 <title>Example 4</title>
 342
 343                 <para>To enter the container, PID of one of the
 344                 processes sharing the new namespaces must be used.
 345                 <command>systemd-nspawn</command> prints the PID
 346                 (as viewed from the outside) of the launched process,
 347                 and it can be used to enter the container.</para>
 348
 349                 <programlisting># nsenter -muinpt $PID</programlisting>
 350
 351                 <para><citerefentry><refentrytitle>nsenter</refentrytitle><manvolnum>1</manvolnum></citerefentry>
 352                 is part of
 353                 <ulink url="https://github.com/karelzak/util-linux">util-linux</ulink>.
 354                 Kernel support for entering namespaces was added in
 355                 Linux 3.8.</para>
 356         </refsect1>
 357
 358         <refsect1>
 359                 <title>Exit status</title>
 360
 361                 <para>The exit code of the program executed in the
 362                 container is returned.</para>
 363         </refsect1>
 364
 365         <refsect1>
 366                 <title>See Also</title>
 367                 <para>
 368                         <citerefentry><refentrytitle>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
 369                         <citerefentry><refentrytitle>chroot</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
 370                         <citerefentry><refentrytitle>unshare</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
 371                         <citerefentry><refentrytitle>yum</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
 372                         <citerefentry><refentrytitle>debootstrap</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
 373                         <citerefentry><refentrytitle>pacman</refentrytitle><manvolnum>8</manvolnum></citerefentry>
 374                 </para>
 375         </refsect1>
 376
 377 </refentry>