systemd/man/systemd-repart.xml

<?xml version='1.0'?> <!--*-nxml-*-->
<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
  "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
<!-- SPDX-License-Identifier: LGPL-2.1+ -->

<refentry id="systemd-repart" conditional='ENABLE_REPART'
    xmlns:xi="http://www.w3.org/2001/XInclude">

  <refentryinfo>
    <title>systemd-repart</title>
    <productname>systemd</productname>
  </refentryinfo>

  <refmeta>
    <refentrytitle>systemd-repart</refentrytitle>
    <manvolnum>8</manvolnum>
  </refmeta>

  <refnamediv>
    <refname>systemd-repart</refname>
    <refname>systemd-repart.service</refname>
    <refpurpose>Automatically grow and add partitions</refpurpose>
  </refnamediv>

  <refsynopsisdiv>
    <cmdsynopsis>
      <command>systemd-repart</command>
      <arg choice="opt" rep="repeat">OPTIONS</arg>
      <arg choice="opt" rep="repeat"><replaceable><optional>BLOCKDEVICE</optional></replaceable></arg>
    </cmdsynopsis>

    <para><filename>systemd-repart.service</filename></para>
  </refsynopsisdiv>

  <refsect1>
    <title>Description</title>

    <para><command>systemd-repart</command> grows and adds partitions to a partition table, based on the
    configuration files described in
    <citerefentry><refentrytitle>repart.d</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
    </para>

    <para>If invoked with no arguments, it operates on the block device backing the root file system partition
    of the OS, thus growing and adding partitions of the booted OS image itself. When called in the initial
    RAM disk it operates on the block device backing <filename>/sysroot/</filename> instead, i.e. on the
    block device the system will soon transition into. The <filename>systemd-repart.service</filename>
    service is generally run at boot in the initial RAM disk, in order to augment the partition table of the
    OS before its partitions are mounted. <command>systemd-repart</command> (mostly) operates in a purely
    incremental mode: it only grows existing and adds new partitions; it does not shrink, delete or move
    existing partitions. The service is intended to be run on every boot, but when it detects that the
    partition table already matches the installed <filename>repart.d/*.conf</filename> configuration
    files, it executes no operation.</para>

    <para><command>systemd-repart</command> is intended to be used when deploying OS images, to automatically
    adjust them to the system they are running on, during first boot. This way the deployed image can be
    minimal in size and may be augmented automatically at boot when needed, taking possession of disk space
    available but not yet used. Specifically the following use cases are among those covered:</para>

    <itemizedlist>
      <listitem><para>The root partition may be grown to cover the whole available disk space.</para></listitem>
      <listitem><para>A <filename>/home/</filename>, swap or <filename>/srv/</filename> partition can be
      added.</para></listitem>
      <listitem><para>A second (or third, …) root partition may be added, to cover A/B style setups
      where a second version of the root file system is alternatingly used for implementing update
      schemes. The deployed image would carry only a single partition ("A") but on first boot a second
      partition ("B") for this purpose is automatically created.</para></listitem>
    </itemizedlist>

    <para>The algorithm executed by <command>systemd-repart</command> is roughly as follows:</para>

    <orderedlist>
      <listitem><para>The <filename>repart.d/*.conf</filename> configuration files are loaded and parsed,
      and ordered by filename (without the directory prefix).</para></listitem>

      <listitem><para>The partition table already existing on the block device is loaded and
      parsed.</para></listitem>

      <listitem><para>The existing partitions in the partition table are matched up with the
      <filename>repart.d/*.conf</filename> files by GPT partition type UUID. The first existing partition
      of a specific type is assigned the first configuration file declaring the same type. The second
      existing partition of a specific type is then assigned the second configuration file declaring the same
      type, and so on. After this iterative assigning is complete any left-over existing partitions that have
      no matching configuration file are considered "foreign" and left as they are. And any configuration
      files for which no partition currently exists are understood as a request to create such a
      partition.</para></listitem>

      <listitem><para>Taking the size constraints and weights declared in the configuration files into
      account, all partitions that shall be created are now allocated to the disk, taking up all free space,
      always respecting the size and padding requests. Similar, existing partitions that are determined to
      grow are grown. New partitions are always appended to the end of the existing partition table, taking
      the first partition table slot whose index is greater than the indexes of all existing
      partitions. Partition table slots are never reordered and thus partition numbers are ensured to remain
      stable. Note that this allocation happens in RAM only, the partition table on disk is not updated
      yet.</para></listitem>

      <listitem><para>All existing partitions for which configuration files exist and which currently have no
      GPT partition label set will be assigned a label, either explicitly configured in the configuration or
      (if that's missing) derived automatically from the partition type. The same is done for all partitions
      that are newly created. These assignments are done in RAM only, too, the disk is not updated
      yet.</para></listitem>

      <listitem><para>Similarly, all existing partitions for which configuration files exist and which
      currently have an all-zero identifying UUID will be assigned a new UUID. This UUID is cryptographically
      hashed from a common seed value together with the partition type UUID (and a counter in case multiple
      partitions of the same type are defined), see below. The same is done for all partitions that are
      created anew. These assignments are done in RAM only, too, the disk is not updated
      yet.</para></listitem>

      <listitem><para>Similarly, if the disk's volume UUID is all zeroes it is also initialized, also
      cryptographically hashed from the same common seed value. Also, in RAM only, too.</para></listitem>

      <listitem><para>The disk space assigned to new partitions (i.e. what was previously considered free
      space but is no longer) is now erased. Specifically, all file system signatures are removed, and if the
      device supports it the <constant>BLKDISCARD</constant> I/O control command is issued to inform the
      hardware that the space is empty now. In addition any "padding" between partitions and at the end of
      the device is similarly erased.</para></listitem>

      <listitem><para>The new partition table is finally written to disk. The kernel is asked to reread the
      partition table.</para></listitem>
    </orderedlist>

    <para>As exception to the normally strictly incremental operation, when called in a special "factory
    reset" mode, <command>systemd-repart</command> may also be used to erase existing partitions to
    reset an installation back to vendor defaults. This mode of operation is used when either the
    <option>--factory-reset=yes</option> switch is passed on the tool's command line, or the
    <option>systemd.factory_reset=yes</option> option specified on the kernel command line, or the
    <varname>FactoryReset</varname> EFI variable (vendor UUID
    <constant>8cf2644b-4b0b-428f-9387-6d876050dc67</constant>) is set to "yes". It alters the algorithm above
    slightly: between the 3rd and the 4th step above any partition marked explicitly via the
    <varname>FactoryReset=</varname> boolean is deleted, and the algorithm restarted, thus immediately
    re-creating these partitions anew empty.</para>

    <para>Note that <command>systemd-repart</command> only changes partition tables, it does not create or
    resize any file systems within these partitions. A separate mechanism should be used for that, for
    example
    <citerefentry><refentrytitle>systemd-growfs</refentrytitle><manvolnum>8</manvolnum></citerefentry> and
    <command>systemd-makefs</command>.</para>

    <para>The UUIDs identifying the new partitions created (or assigned to existing partitions that have no
    UUID yet), as well as the disk as a whole are hashed cryptographically from a common seed value. This
    seed value is usually the
    <citerefentry><refentrytitle>machine-id</refentrytitle><manvolnum>5</manvolnum></citerefentry> of the
    system, so that the machine ID reproducibly determines the UUIDs assigned to all partitions. If the
    machine ID cannot be read (or the user passes <option>--seed=random</option>, see below) the seed is
    generated randomly instead, so that the partition UUIDs are also effectively random. The seed value may
    also be set explicitly, formatted as UUID via the <option>--seed=</option> option. By hashing these UUIDs
    from a common seed images prepared with this tool become reproducible and the result of the algorithm
    above deterministic.</para>

    <para>The positional argument should specify the block device to operate on. Instead of a block device
    node path a regular file may be specified too, in which case the command operates on it like it would if
    a loopback block device node was specified with the file attached. If <option>--empty=create</option> is
    specified the specified path is created as regular file, which is useful for generating disk images from
    scratch.</para>
  </refsect1>

  <refsect1>
    <title>Options</title>

    <para>The following options are understood:</para>

    <variablelist>
      <varlistentry>
        <term><option>--dry-run=</option></term>
        <listitem><para>Takes a boolean. If this switch is not specified <option>--dry-run=yes</option> is
        the implied default. Controls whether <filename>systemd-repart</filename> executes the requested
        re-partition operations or whether it should only show what it would do. Unless
        <option>--dry-run=no</option> is specified <filename>systemd-repart</filename> will not actually
        touch the device's partition table.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--empty=</option></term>
        <listitem><para>Takes one of <literal>refuse</literal>, <literal>allow</literal>,
        <literal>require</literal>, <literal>force</literal> or <literal>create</literal>. Controls how to
        operate on block devices that are entirely empty, i.e. carry no partition table/disk label yet. If
        this switch is not specified the implied default is <literal>refuse</literal>.</para>

        <para>If <literal>refuse</literal> <command>systemd-repart</command> requires that the block device
        it shall operate on already carries a partition table and refuses operation if none is found. If
        <literal>allow</literal> the command will extend an existing partition table or create a new one if
        none exists. If <literal>require</literal> the command will create a new partition table if none
        exists so far, and refuse operation if one already exists. If <literal>force</literal> it will create
        a fresh partition table unconditionally, erasing the disk fully in effect. If
        <literal>force</literal> no existing partitions will be taken into account or survive the
        operation. Hence: use with care, this is a great way to lose all your data. If
        <literal>create</literal> a new loopback file is create under the path passed via the device node
        parameter, of the size indicated with <option>--size=</option>, see below.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--discard=</option></term>

        <listitem><para>Takes a boolean. If this switch is not specified <option>--discard=yes</option> is
        the implied default. Controls whether to issue the <constant>BLKDISCARD</constant> I/O control
        command on the space taken up by any added partitions or on the space in between them. Usually, it's
        a good idea to issue this request since it tells the underlying hardware that the covered blocks
        shall be considered empty, improving performance. If operating on a regular file instead of a block
        device node, a sparse file is generated.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--size=</option></term>

        <listitem><para>Takes a size in bytes, using the usual K, M, G, T suffixes, or the special value
        <literal>auto</literal>. If used the specified device node path must refer to a regular file, which
        is then grown to the specified size if smaller, before any change is made to the partition table. If
        specified as <literal>auto</literal> the minimal size for the disk image is automatically determined
        (i.e. the minimal sizes of all partitions are summed up, taking space for additional metadata into
        account). This switch is not supported if the specified node is a block device. This switch has no
        effect if the file is already as large as the specified size or larger. The specified size is
        implicitly rounded up to multiples of 4096. When used with <option>--empty=create</option> this
        specifies the initial size of the loopback file to create.</para>

        <para>The <option>--size=auto</option> option takes the sizes of pre-existing partitions into
        account. However, it does not accommodate for partition tables that are not tightly packed: the
        configured partitions might still not fit into the backing device if empty space exists between
        pre-existing partitions (or before the first partition) that cannot be fully filled by partitions to
        grow or create.</para>

        <para>Also note that the automatic size determination does not take files or directories specified
        with <option>CopyFiles=</option> into account: operation might fail if the specified files or
        directories require more disk space then the configured per-partition minimal size
        limit.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--factory-reset=</option></term>

        <listitem><para>Takes boolean. If this switch is not specified <option>--factory=reset=no</option> is
        the implied default. Controls whether to operate in "factory reset" mode, see above. If set to true
        this will remove all existing partitions marked with <varname>FactoryReset=</varname> set to yes
        early while executing the re-partitioning algorithm. Use with care, this is a great way to lose all
        your data. Note that partition files need to explicitly turn <varname>FactoryReset=</varname> on, as
        the option defaults to off. If no partitions are marked for factory reset this switch has no
        effect. Note that there are two other methods to request factory reset operation: via the kernel
        command line and via an EFI variable, see above.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--can-factory-reset</option></term>

        <listitem><para>If this switch is specified the disk is not re-partitioned. Instead it is determined
        if any existing partitions are marked with <varname>FactoryReset=</varname>. If there are the tool
        will exit with exit status zero, otherwise non-zero. This switch may be used to quickly determine
        whether the running system supports a factory reset mechanism built on
        <command>systemd-repart</command>.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--root=</option></term>

        <listitem><para>Takes a path to a directory to use as root file system when searching for
        <filename>repart.d/*.conf</filename> files and for the machine ID file to use as seed. By default
        when invoked on the regular system this defaults to the host's root file system
        <filename>/</filename>. If invoked from the initial RAM disk this defaults to
        <filename>/sysroot/</filename>, so that the tool operates on the configuration and machine ID stored
        in the root file system later transitioned into itself.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--seed=</option></term>

        <listitem><para>Takes a UUID as argument or the special value <constant>random</constant>. If a UUID
        is specified the UUIDs to assign to partitions and the partition table itself are derived via
        cryptographic hashing from it. If not specified it is attempted to read the machine ID from the host
        (or more precisely, the root directory configured via <option>--root=</option>) and use it as seed
        instead, falling back to a randomized seed otherwise. Use <option>--seed=random</option> to force a
        randomized seed. Explicitly specifying the seed may be used to generated strictly reproducible
        partition tables.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--pretty=</option></term>

        <listitem><para>Takes a boolean argument. If this switch is not specified, it defaults to on when
        called from an interactive terminal and off otherwise. Controls whether to show a user friendly table
        and graphic illustrating the changes applied.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--json=</option><replaceable>MODE</replaceable></term>

        <listitem><para>Shows output formatted as JSON. Expects one of <literal>short</literal> (for the
        shortest possible output without any redundant whitespace or line breaks), <literal>pretty</literal>
        (for a pretty version of the same, with indentation and line breaks) or <literal>off</literal> (to turn
        off json output).</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--definitions=</option></term>

        <listitem><para>Takes a file system path. If specified the <filename>*.conf</filename> files are read
        from the specified directory instead of searching in <filename>/usr/lib/repart.d/*.conf</filename>,
        <filename>/etc/repart.d/*.conf</filename>,
        <filename>/run/repart.d/*.conf</filename>.</para></listitem>
      </varlistentry>

      <varlistentry>
        <term><option>--key-file=</option></term>

        <listitem><para>Takes a file system path. Configures the encryption key to use when setting up LUKS2
        volumes configured with the <varname>Encrypt=</varname> setting in partition files. Should refer to a
        regular file containing the key, or an <constant>AF_UNIX</constant> stream socket in the file
        system. In the latter case a connection is made to it and the key read from it. If this switch is not
        specified the empty key (i.e. zero length key) is used. This behaviour is useful for setting up encrypted
        partitions during early first boot that receive their user-supplied password only in a later setup
        step.</para></listitem>
      </varlistentry>

      <xi:include href="standard-options.xml" xpointer="help" />
      <xi:include href="standard-options.xml" xpointer="version" />
    </variablelist>
  </refsect1>

  <refsect1>
    <title>See Also</title>
    <para>
      <citerefentry><refentrytitle>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
      <citerefentry><refentrytitle>repart.d</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
      <citerefentry><refentrytitle>machine-id</refentrytitle><manvolnum>5</manvolnum></citerefentry>
    </para>
  </refsect1>

</refentry>