systemd-nspawn — Spawn a command or OS in a lightweight container
systemd-nspawn
[OPTIONS...] [COMMAND
[ARGS...]
]
systemd-nspawn
--boot [OPTIONS...] [ARGS...]
systemd-nspawn may be used to run a command or OS in a lightweight namespace container. In many ways it is similar to chroot(1), but more powerful since it virtualizes the file system hierarchy, as well as the process tree, the various IPC subsystems, and the host and domain names.
systemd-nspawn may be invoked on any directory tree containing an operating system tree,
using the --directory=
command line option. By using the --machine=
option an OS
tree is automatically searched for in a couple of locations, most importantly in
/var/lib/machines/
, the suggested directory to place OS container images installed on the
system.
In contrast to chroot(1) systemd-nspawn may be used to boot full Linux-based operating systems in a container.
systemd-nspawn limits access to various kernel interfaces in the container to
read-only, such as /sys/
, /proc/sys/
, or
/sys/fs/selinux/
. The host's network interfaces and the system clock may not be
changed from within the container. Device nodes may not be created. The host system cannot be rebooted
and kernel modules may not be loaded from within the container. This sandbox can easily be
circumvented from within the container if user namespaces are not used. This means that
untrusted code must always be run in a user namespace, see the discussion of the
--private-users=
option below.
Use a tool like dnf(8), debootstrap(8), or pacman(8) to set up an OS directory tree suitable as file system hierarchy for systemd-nspawn containers. See the Examples section below for details on suitable invocation of these commands.
As a safety check systemd-nspawn will verify the existence of
/usr/lib/os-release
or /etc/os-release
in the container tree before
booting a container (see
os-release(5)). It might be
necessary to add this file to the container tree manually if the OS of the container is too old to contain this
file out-of-the-box.
systemd-nspawn may be invoked directly from the interactive command line or run as system
service in the background. In this mode each container instance runs as its own service instance; a default
template unit file systemd-nspawn@.service
is provided to make this easy, taking the container
name as instance identifier. Note that different default options apply when systemd-nspawn is
invoked by the template unit file than interactively on the command line. Most importantly the template unit file
makes use of the --boot
option which is not the default in case systemd-nspawn
is invoked from the interactive command line. Further differences with the defaults are documented along with the
various supported options below.
The machinectl(1) tool may
be used to execute a number of operations on containers. In particular it provides easy-to-use commands to run
containers as system services using the systemd-nspawn@.service
template unit
file.
Along with each container a settings file with the .nspawn
suffix may exist, containing
additional settings to apply when running the container. See
systemd.nspawn(5) for
details. Settings files override the default options used by the systemd-nspawn@.service
template unit file, making it usually unnecessary to alter this template file directly.
Note that systemd-nspawn will mount file systems private to the container to
/dev/
, /run/
, and similar. These will not be visible outside of
the container, and their contents will be lost when the container exits.
Note that running two systemd-nspawn containers from the same directory tree will not make processes in them see each other. The PID namespace separation of the two containers is complete and the containers will share very few runtime objects except for the underlying file system. Rather use machinectl(1)'s login or shell commands to request an additional login session in a running container.
systemd-nspawn implements the Container Interface specification.
While running, containers invoked with systemd-nspawn are registered with the systemd-machined(8) service that keeps track of running containers, and provides programming interfaces to interact with them.
systemd-nspawn may be invoked with or without privileges. The full functionality is currently only available when invoked with privileges. When invoked without privileges, various limitations apply, including, but not limited to:
Only disk image based containers are supported (i.e. --image=
).
Directory based ones (i.e. --directory=
) are not supported.
Machine registration via --machine=
is not supported.
Only --private-network
and --network-veth
networking modes are supported.
When running in unprivileged mode, some needed functionality is provided via systemd-mountfsd.service(8) and systemd-nsresourced.service(8)
If option --boot
is specified, the arguments
are used as arguments for the init program. Otherwise,
COMMAND
specifies the program to launch
in the container, and the remaining arguments are used as
arguments for this program. If --boot
is not used and
no arguments are specified, a shell is launched in the
container.
The following options are understood:
-q
, --quiet
¶Turns off any status output by the tool itself. When this switch is used, the only output from nspawn will be the console output of the container OS itself.
--settings=MODE
¶Controls whether
systemd-nspawn shall search for and use
additional per-container settings from
.nspawn
files. Takes a boolean or the
special values override
or
trusted
.
If enabled (the default), a settings file named after the
machine (as specified with the --machine=
setting, or derived from the directory or image file name)
with the suffix .nspawn
is searched in
/etc/systemd/nspawn/
and
/run/systemd/nspawn/
. If it is found
there, its settings are read and used. If it is not found
there, it is subsequently searched in the same directory as the
image file or in the immediate parent of the root directory of
the container. In this case, if the file is found, its settings
will be also read and used, but potentially unsafe settings
are ignored. Note that in both these cases, settings on the
command line take precedence over the corresponding settings
from loaded .nspawn
files, if both are
specified. Unsafe settings are considered all settings that
elevate the container's privileges or grant access to
additional resources such as files or directories of the
host. For details about the format and contents of
.nspawn
files, consult
systemd.nspawn(5).
If this option is set to override
, the
file is searched, read and used the same way, however, the order of
precedence is reversed: settings read from the
.nspawn
file will take precedence over
the corresponding command line options, if both are
specified.
If this option is set to trusted
, the
file is searched, read and used the same way, but regardless
of being found in /etc/systemd/nspawn/
,
/run/systemd/nspawn/
or next to the image
file or container root directory, all settings will take
effect, however, command line arguments still take precedence
over corresponding settings.
If disabled, no .nspawn
file is read
and no settings except the ones on the command line are in
effect.
-D
, --directory=
¶Directory to use as file system root for the container.
If neither --directory=
, nor --image=
is specified the
directory is determined by searching for a directory named the same as the machine name specified
with --machine=
. See
machinectl(1)
section "Files and Directories" for the precise search path.
In place of the directory path a ".v/
" versioned directory may be specified,
see systemd.v(7) for
details.
If neither --directory=
, --image=
, nor
--machine=
are specified, the current directory will be used. May not be specified
together with --image=
.
--template=
¶Directory or "btrfs
" subvolume to use as template for the
container's root directory. If this is specified and the container's root directory (as configured by
--directory=
) does not yet exist it is created as "btrfs
" snapshot
(if supported) or plain directory (otherwise) and populated from this template tree. Ideally, the
specified template path refers to the root of a "btrfs
" subvolume, in which case a
simple copy-on-write snapshot is taken, and populating the root directory is instant. If the
specified template path does not refer to the root of a "btrfs
" subvolume (or not
even to a "btrfs
" file system at all), the tree is copied (though possibly in a
'reflink' copy-on-write scheme — if the file system supports that), which can be substantially more
time-consuming. Note that the snapshot taken is of the specified directory or subvolume, including
all subdirectories and subvolumes below it, but excluding any sub-mounts. May not be specified
together with --image=
or --ephemeral
.
Note that this switch leaves hostname, machine ID and all other settings that could identify the instance unmodified.
-x
, --ephemeral
¶If specified, the container is run with a temporary snapshot of its file system that is removed
immediately when the container terminates. May not be specified together with
--template=
.
Note that this switch leaves hostname, machine ID and all other settings that could identify
the instance unmodified. Please note that — as with --template=
— taking the
temporary snapshot is more efficient on file systems that support subvolume snapshots or 'reflinks'
natively ("btrfs
" or new "xfs
") than on more traditional file
systems that do not ("ext4
"). Note that the snapshot taken is of the specified
directory or subvolume, including all subdirectories and subvolumes below it, but excluding any
sub-mounts.
With this option no modifications of the container image are retained. Use
--volatile=
(described below) for other mechanisms to restrict persistency of
container images during runtime.
-i
, --image=
¶Disk image to mount the root directory for the container from. Takes a path to a regular file or to a block device node. The file or block device must contain either:
An MBR partition table with a single partition of type 0x83 that is marked bootable.
A GUID partition table (GPT) with a single partition of type 0fc63daf-8483-4772-8e79-3d69d8477de4.
A GUID partition table (GPT) with a marked root partition which is mounted as the root directory of the container. Optionally, GPT images may contain a home and/or a server data partition which are mounted to the appropriate places in the container. All these partitions must be identified by the partition types defined by the Discoverable Partitions Specification.
No partition table, and a single file system spanning the whole image.
On GPT images, if an EFI System Partition (ESP) is discovered, it is automatically mounted to
/efi
(or /boot
as fallback) in case a directory by this name exists
and is empty.
Partitions encrypted with LUKS are automatically decrypted. Also, on GPT images dm-verity data integrity
hash partitions are set up if the root hash for them is specified using the --root-hash=
option.
Single file system images (i.e. file systems without a surrounding partition table) can be opened using
dm-verity if the integrity data is passed using the --root-hash=
and
--verity-data=
(and optionally --root-hash-sig=
) options.
Any other partitions, such as foreign partitions or swap partitions are not mounted. May not be specified
together with --directory=
, --template=
.
In place of the image path a ".v/
" versioned directory may be specified, see
systemd.v(7) for
details.
--image-policy=policy
¶Takes an image policy string as argument, as per
systemd.image-policy(7). The
policy is enforced when operating on the disk image specified via --image=
, see
above. If not specified defaults to
"root=verity+signed+encrypted+unprotected+absent:usr=verity+signed+encrypted+unprotected+absent:home=encrypted+unprotected+absent:srv=encrypted+unprotected+absent:esp=unprotected+absent:xbootldr=unprotected+absent:tmp=encrypted+unprotected+absent:var=encrypted+unprotected+absent
",
i.e. all recognized file systems in the image are used, but not the swap partition.
--oci-bundle=
¶Takes the path to an OCI runtime bundle to invoke, as specified in the OCI Runtime Specification. In
this case no .nspawn
file is loaded, and the root directory and various settings are read
from the OCI runtime JSON data (but data passed on the command line takes precedence).
--read-only
¶Mount the container's root file system (and any other file systems contained in the container
image) read-only. This has no effect on additional mounts made with --bind=
,
--tmpfs=
and similar options. This mode is implied if the container image file or directory is
marked read-only itself. It is also implied if --volatile=
is used. In this case the container
image on disk is strictly read-only, while changes are permitted but kept non-persistently in memory only. For
further details, see below.
--volatile
, --volatile=MODE
¶Boots the container in volatile mode. When no mode parameter is passed or when mode is
specified as yes
, full volatile mode is enabled. This means the root directory is mounted as a
mostly unpopulated "tmpfs
" instance, and /usr/
from the OS tree is
mounted into it in read-only mode (the system thus starts up with read-only OS image, but pristine state and
configuration, any changes are lost on shutdown). When the mode parameter is specified as
state
, the OS tree is mounted read-only, but /var/
is mounted as a
writable "tmpfs
" instance into it (the system thus starts up with read-only OS resources and
configuration, but pristine state, and any changes to the latter are lost on shutdown). When the mode parameter
is specified as overlay
the read-only root file system is combined with a writable
tmpfs
instance through "overlayfs
", so that it appears at it normally
would, but any changes are applied to the temporary file system only and lost when the container is
terminated. When the mode parameter is specified as no
(the default), the whole OS tree is
made available writable (unless --read-only
is specified, see above).
Note that if one of the volatile modes is chosen, its effect is limited to the root file system
(or /var/
in case of state
), and any other mounts placed in the
hierarchy are unaffected — regardless if they are established automatically (e.g. the EFI system
partition that might be mounted to /efi/
or /boot/
) or
explicitly (e.g. through an additional command line option such as --bind=
, see
below). This means, even if --volatile=overlay
is used changes to
/efi/
or /boot/
are prohibited in case such a partition
exists in the container image operated on, and even if --volatile=state
is used the
hypothetical file /etc/foobar
is potentially writable if
--bind=/etc/foobar
is used to mount it from outside the read-only container
/etc/
directory.
The --ephemeral
option is closely related to this setting, and provides similar
behaviour by making a temporary, ephemeral copy of the whole OS image and executing that. For further details,
see above.
The --tmpfs=
and --overlay=
options provide similar functionality, but
for specific sub-directories of the OS image only. For details, see below.
This option provides similar functionality for containers as the "systemd.volatile=
"
kernel command line switch provides for host systems. See
kernel-command-line(7) for
details.
Note that setting this option to yes
or state
will only work
correctly with operating systems in the container that can boot up with only
/usr/
mounted, and are able to automatically populate /var/
(and /etc/
in case of "--volatile=yes
"). Specifically, this
means that operating systems that follow the historic split of /bin/
and
/lib/
(and related directories) from /usr/
(i.e. where the
former are not symlinks into the latter) are not supported by "--volatile=yes
" as
container payload. The overlay
option does not require any particular preparations
in the OS, but do note that "overlayfs
" behaviour differs from regular file systems
in a number of ways, and hence compatibility is limited.
--root-hash=
¶Takes a data integrity (dm-verity) root hash specified in hexadecimal. This option enables data
integrity checks using dm-verity, if the used image contains the appropriate integrity data (see above). The
specified hash must match the root hash of integrity data, and is usually at least 256 bits (and hence 64
formatted hexadecimal characters) long (in case of SHA256 for example). If this option is not specified, but
the image file carries the "user.verity.roothash
" extended file attribute (see xattr(7)), then the root
hash is read from it, also as formatted hexadecimal characters. If the extended file attribute is not found (or
is not supported by the underlying file system), but a file with the .roothash
suffix is
found next to the image file, bearing otherwise the same name (except if the image has the
.raw
suffix, in which case the root hash file must not have it in its name), the root hash
is read from it and automatically used, also as formatted hexadecimal characters.
Note that this configures the root hash for the root file system. Disk images may also contain
separate file systems for the /usr/
hierarchy, which may be Verity protected as
well. The root hash for this protection may be configured via the
"user.verity.usrhash
" extended file attribute or via a .usrhash
file adjacent to the disk image, following the same format and logic as for the root hash for the
root file system described here. Note that there's currently no switch to configure the root hash for
the /usr/
from the command line.
Also see the RootHash=
option in
systemd.exec(5).
--root-hash-sig=
¶Takes a PKCS7 signature of the --root-hash=
option.
The semantics are the same as for the RootHashSignature=
option, see
systemd.exec(5).
--verity-data=
¶Takes the path to a data integrity (dm-verity) file. This option enables data integrity checks
using dm-verity, if a root-hash is passed and if the used image itself does not contain the integrity data.
The integrity data must be matched by the root hash. If this option is not specified, but a file with the
.verity
suffix is found next to the image file, bearing otherwise the same name (except if
the image has the .raw
suffix, in which case the verity data file must not have it in its name),
the verity data is read from it and automatically used.
--pivot-root=
¶Pivot the specified directory to /
inside the container, and either unmount the
container's old root, or pivot it to another specified directory. Takes one of: a path argument — in which case the
specified path will be pivoted to /
and the old root will be unmounted; or a colon-separated pair
of new root path and pivot destination for the old root. The new root path will be pivoted to /
,
and the old /
will be pivoted to the other directory. Both paths must be absolute, and are resolved
in the container's file system namespace.
This is for containers which have several bootable directories in them; for example, several OSTree deployments. It emulates the behavior of the boot loader and the initrd which normally select which directory to mount as the root and start the container's PID 1 in.
-a
, --as-pid2
¶Invoke the shell or specified program as process ID (PID) 2 instead of PID 1 (init). By
default, if neither this option nor --boot
is used, the selected program is run as the process
with PID 1, a mode only suitable for programs that are aware of the special semantics that the process with
PID 1 has on UNIX. For example, it needs to reap all processes reparented to it, and should implement
sysvinit compatible signal handling (specifically: it needs to reboot on SIGINT, reexecute
on SIGTERM, reload configuration on SIGHUP, and so on). With --as-pid2
a minimal stub init
process is run as PID 1 and the selected program is executed as PID 2 (and hence does not need to implement any
special semantics). The stub init process will reap processes as necessary and react appropriately to
signals. It is recommended to use this mode to invoke arbitrary commands in containers, unless they have been
modified to run correctly as PID 1. Or in other words: this switch should be used for pretty much all commands,
except when the command refers to an init or shell implementation, as these are generally capable of running
correctly as PID 1. This option may not be combined with --boot
.
-b
, --boot
¶Automatically search for an init program and invoke it as PID 1, instead of a shell or a user
supplied program. If this option is used, arguments specified on the command line are used as arguments for the
init program. This option may not be combined with --as-pid2
.
The following table explains the different modes of invocation and relationship to
--as-pid2
(see above):
Table 1. Invocation Mode
Switch | Explanation |
---|---|
Neither --as-pid2 nor --boot specified | The passed parameters are interpreted as the command line, which is executed as PID 1 in the container. |
--as-pid2 specified | The passed parameters are interpreted as the command line, which is executed as PID 2 in the container. A stub init process is run as PID 1. |
--boot specified | An init program is automatically searched for and run as PID 1 in the container. The passed parameters are used as invocation parameters for this process. |
Note that --boot
is the default mode of operation if the
systemd-nspawn@.service
template unit file is used.
--chdir=
¶Change to the specified working directory before invoking the process in the container. Expects an absolute path in the container's file system namespace.
-E NAME
[=VALUE
]
, --setenv=NAME
[=VALUE
]
¶Specifies an environment variable to pass to the init process in the container. This
may be used to override the default variables or to set additional variables. It may be used more
than once to set multiple variables. When "=
" and VALUE
are omitted, the value of the variable with the same name in the program environment will be used.
-u
, --user=
¶After transitioning into the container, change to the specified user defined in the container's user database. Like all other systemd-nspawn features, this is not a security feature and provides protection against accidental destructive operations only.
Note that if credentials are used in combination with a non-root --user=
(e.g.: --set-credential=
, --load-credential=
or
--import-credential=
), then --no-new-privileges=yes
must be used, and
--boot
or --as-pid2
must not be used, as the credentials would
otherwise be unreadable by the container due to missing privileges after switching to the specified
user.
--kill-signal=
¶Specify the process signal to send to the container's PID 1 when nspawn itself receives
SIGTERM
, in order to trigger an orderly shutdown of the container. Defaults to
SIGRTMIN+3
if --boot
is used (on systemd-compatible init systems
SIGRTMIN+3
triggers an orderly shutdown). If --boot
is not used and this
option is not specified the container's processes are terminated abruptly via SIGKILL
. For
a list of valid signals, see signal(7).
--notify-ready=
¶Configures support for notifications from the container's init process.
--notify-ready=
takes a boolean (no
and yes
).
With option no
systemd-nspawn notifies systemd
with a "READY=1
" message when the init process is created.
With option yes
systemd-nspawn waits for the
"READY=1
" message from the init process in the container
before sending its own to systemd. For more details about notifications
see sd_notify(3).
--suppress-sync=
¶Expects a boolean argument. If true, turns off any form of on-disk file system
synchronization for the container payload. This means all system calls such as sync(2),
fsync()
, syncfs()
, … will execute no operation, and the
O_SYNC
/O_DSYNC
flags to open(2) and
related calls will be made unavailable. This is potentially dangerous, as assumed data integrity
guarantees to the container payload are not actually enforced (i.e. data assumed to have been written
to disk might be lost if the system is shut down abnormally). However, this can dramatically improve
container runtime performance – as long as these guarantees are not required or desirable, for
example because any data written by the container is of temporary, redundant nature, or just an
intermediary artifact that will be further processed and finalized by a later step in a
pipeline. Defaults to false.
-M
, --machine=
¶Sets the machine name for this container. This
name may be used to identify this container during its runtime
(for example in tools like
machinectl(1)
and similar), and is used to initialize the container's
hostname (which the container can choose to override,
however). If not specified, the last component of the root
directory path of the container is used, possibly suffixed
with a random identifier in case --ephemeral
mode is selected. If the root directory selected is the host's
root directory the host's hostname is used as default
instead.
--hostname=
¶Controls the hostname to set within the container, if different from the machine name. Expects
a valid hostname as argument. If this option is used, the kernel hostname of the container will be set to this
value, otherwise it will be initialized to the machine name as controlled by the --machine=
option described above. The machine name is used for various aspect of identification of the container from the
outside, the kernel hostname configurable with this option is useful for the container to identify itself from
the inside. It is usually a good idea to keep both forms of identification synchronized, in order to avoid
confusion. It is hence recommended to avoid usage of this option, and use --machine=
exclusively. Note that regardless whether the container's hostname is initialized from the name set with
--hostname=
or the one set with --machine=
, the container can later override
its kernel hostname freely on its own as well.
--uuid=
¶Set the specified UUID for the container. The
init system will initialize
/etc/machine-id
from this if this file is
not set yet. Note that this option takes effect only if
/etc/machine-id
in the container is
unpopulated.
-S
, --slice=
¶Make the container part of the specified slice, instead of the default
machine.slice
. This applies only if the machine is run in its own scope unit, i.e. if
--keep-unit
isn't used.
--property=
¶Set a unit property on the scope unit to register for the machine. This applies only if the
machine is run in its own scope unit, i.e. if --keep-unit
isn't used. Takes unit property
assignments in the same format as systemctl set-property. This is useful to set memory
limits and similar for the container.
--register=
¶Controls whether the container is registered with
systemd-machined(8). Takes a
boolean argument, which defaults to "yes
". This option should be enabled when the container
runs a full Operating System (more specifically: a system and service manager as PID 1), and is useful to
ensure that the container is accessible via
machinectl(1) and shown by
tools such as ps(1). If the container
does not run a service manager, it is recommended to set this option to
"no
".
--keep-unit
¶Instead of creating a transient scope unit to run the container in, simply use the service or
scope unit systemd-nspawn has been invoked in. If --register=yes
is set
this unit is registered with
systemd-machined(8). This
switch should be used if systemd-nspawn is invoked from within a service unit, and the
service unit's sole purpose is to run a single systemd-nspawn container. This option is not
available if run from a user session.
Note that passing --keep-unit
disables the effect of --slice=
and
--property=
. Use --keep-unit
and --register=no
in
combination to disable any kind of unit allocation or registration with
systemd-machined.
--private-users=
¶Controls user namespacing. If enabled, the container will run with its own private set of UNIX user and group ids (UIDs and GIDs). This involves mapping the private UIDs/GIDs used in the container (starting with the container's root user 0 and up) to a range of UIDs/GIDs on the host that are not used for other purposes (usually in the range beyond the host's UID/GID 65536). The parameter may be specified as follows:
If one or two colon-separated numbers are specified, user namespacing is turned on. The first parameter specifies the first host UID/GID to assign to the container, the second parameter specifies the number of host UIDs/GIDs to assign to the container. If the second parameter is omitted, 65536 UIDs/GIDs are assigned.
If the parameter is "yes
", user namespacing is turned on. The
UID/GID range to use is determined automatically from the file ownership of the root directory of
the container's directory tree. To use this option, make sure to prepare the directory tree in
advance, and ensure that all files and directories in it are owned by UIDs/GIDs in the range you'd
like to use. Also, make sure that used file ACLs exclusively reference UIDs/GIDs in the appropriate
range. In this mode, the number of UIDs/GIDs assigned to the container is 65536, and the owner
UID/GID of the root directory must be a multiple of 65536.
The special value "pick
" turns on user namespacing. In this case
the UID/GID range is automatically chosen. As first step, the file owner UID/GID of the root
directory of the container's directory tree is read, and it is checked that no other container is
currently using it. If this check is successful, the UID/GID range determined this way is used,
similarly to the behavior if "yes
" is specified. If the check is not successful
(and thus the UID/GID range indicated in the root directory's file owner is already used elsewhere)
a new – currently unused – UID/GID range of 65536 UIDs/GIDs is randomly chosen between the host
UID/GIDs of 524288 and 1878982656, always starting at a multiple of 65536, and, if possible,
consistently hashed from the machine name. This setting implies
--private-users-ownership=auto
(see below), which possibly has the effect that the
files and directories in the container's directory tree will be owned by the appropriate users of
the range picked. Using this option makes user namespace behavior fully automatic. Note that the
first invocation of a previously unused container image might result in picking a new UID/GID range
for it, and thus in the (possibly expensive) file ownership adjustment operation. However,
subsequent invocations of the container will be cheap (unless of course the picked UID/GID range is
assigned to a different use by then).
If the parameter is "no
", user namespacing is turned off. This is
the default when systemd-nspawn is invoked directly. (Note that the
systemd-nspawn@.service
unit enables private users.) This option is not
secure and must not be used to run untrusted code.
If the parameter is "identity
", user namespacing is employed with
an identity mapping for the first 65536 UIDs/GIDs. This is mostly equivalent to
--private-users=0:65536
. While it does not provide UID/GID isolation, since all
host and container UIDs/GIDs are chosen identically it does provide process capability isolation,
but may be useful if proper user namespacing with distinct UID maps is not possible. This option is
not secure and must not be used to run untrusted code.
It is recommended to assign at least 65536 UIDs/GIDs to each container, so that the usable
UID/GID range in the container covers 16 bits. For best security, do not assign overlapping UID/GID
ranges to multiple containers. It is hence a good idea to use the upper 16 bit of the host 32-bit
UIDs/GIDs as container identifier, while the lower 16 bits encode the container UID/GID used. This is
in fact the behavior enforced by the --private-users=pick
option.
When user namespaces are used, the GID range assigned to each container is always chosen identical to the UID range.
In most cases, using --private-users=pick
is the recommended option as user
namespacing is required for security, and this option massively enhances container security while
operating fully automatically in most cases.
Note that the picked UID/GID range is not written to /etc/passwd
or
/etc/group
. In fact, the allocation of the range is not stored persistently,
except in the file ownership of the files and directories of the container.
Note that when user namespacing is used file ownership on disk reflects this, and all of the container's files and directories are owned by the container's effective user and group IDs. This means that copying files from and to the container image requires correction of the numeric UID/GID values, according to the UID/GID shift applied.
--private-users-ownership=
¶Controls how to adjust the container image's UIDs and GIDs to match the UID/GID range
chosen with --private-users=
, see above. Takes one of "off
" (to
leave the image as is), "chown
" (to recursively chown()
the
container's directory tree as needed), "map
" (in order to use transparent ID mapping
mounts) or "auto
" for automatically using "map
" where available and
"chown
" where not.
If "chown
" is selected, all files and directories in the container's directory
tree will be adjusted so that they are owned by the appropriate UIDs/GIDs selected for the container
(see above). This operation is potentially expensive, as it involves iterating through the full
directory tree of the container. Besides actual file ownership, file ACLs are adjusted as
well.
Typically "map
" is the best choice, since it transparently maps UIDs/GIDs in
memory as needed without modifying the image, and without requiring an expensive recursive adjustment
operation. However, it is not available for all file systems, currently.
The --private-users-ownership=auto
option is implied if
--private-users=pick
is used. This option has no effect if user namespacing is not
used.
-U
¶If the kernel supports the user namespaces feature, equivalent to
--private-users=pick --private-users-ownership=auto
, otherwise equivalent to
--private-users=no
.
Note that -U
is the default if the
systemd-nspawn@.service
template unit file is used.
Note: it is possible to undo the effect of --private-users-ownership=chown
(or
-U
) on the file system by redoing the operation with the first UID of 0:
systemd-nspawn … --private-users=0 --private-users-ownership=chown
--private-network
¶Disconnect networking of the container from
the host. This makes all network interfaces unavailable in the
container, with the exception of the loopback device and those
specified with --network-interface=
and
configured with --network-veth
. If this
option is specified, the CAP_NET_ADMIN
capability will be
added to the set of capabilities the container retains. The
latter may be disabled by using --drop-capability=
.
If this option is not specified (or implied by one of the options
listed below), the container will have full access to the host network.
--network-interface=
¶Assign the specified network interface to the container. Either takes a single
interface name, referencing the name on the host, or a colon-separated pair of interfaces, in which
case the first one references the name on the host, and the second one the name in the container.
When the container terminates, the interface is moved back to the calling namespace and renamed to
its original name. Note that --network-interface=
implies
--private-network
. This option may be used more than once to add multiple network
interfaces to the container.
Note that any network interface specified this way must already exist at the time the container
is started. If the container shall be started automatically at boot via a
systemd-nspawn@.service
unit file instance, it might hence make sense to add a
unit file drop-in to the service instance
(e.g. /etc/systemd/system/systemd-nspawn@foobar.service.d/50-network.conf
) with
contents like the following:
[Unit] Wants=sys-subsystem-net-devices-ens1.device After=sys-subsystem-net-devices-ens1.device
This will make sure that activation of the container service will be delayed until the
"ens1
" network interface has shown up. This is required since hardware probing is
fully asynchronous, and network interfaces might be discovered only later during the boot process,
after the container would normally be started without these explicit dependencies.
--network-macvlan=
¶Create a "macvlan
" interface of the specified Ethernet network
interface and add it to the container. Either takes a single interface name, referencing the name
on the host, or a colon-separated pair of interfaces, in which case the first one references the name
on the host, and the second one the name in the container. A "macvlan
" interface is
a virtual interface that adds a second MAC address to an existing physical Ethernet link. If the
container interface name is not defined, the interface in the container will be named after the
interface on the host, prefixed with "mv-
". Note that
--network-macvlan=
implies --private-network
. This option may be
used more than once to add multiple network interfaces to the container.
As with --network-interface=
, the underlying Ethernet network interface must
already exist at the time the container is started, and thus similar unit file drop-ins as described
above might be useful.
--network-ipvlan=
¶Create an "ipvlan
" interface of the specified Ethernet network
interface and add it to the container. Either takes a single interface name, referencing the name on
the host, or a colon-separated pair of interfaces, in which case the first one references the name
on the host, and the second one the name in the container. An "ipvlan
" interface is
a virtual interface,
similar to a "macvlan
" interface, which uses the same MAC address as the underlying
interface. If the container interface name is not defined, the interface in the container will be
named after the interface on the host, prefixed
with "iv-
". Note that --network-ipvlan=
implies
--private-network
. This option may be used more than once to add multiple network
interfaces to the container.
As with --network-interface=
, the underlying Ethernet network interface must
already exist at the time the container is started, and thus similar unit file drop-ins as described
above might be useful.
-n
, --network-veth
¶Create a virtual Ethernet link ("veth
") between host and container. The host
side of the Ethernet link will be available as a network interface named after the container's name (as
specified with --machine=
), prefixed with "ve-
". The container side of the
Ethernet link will be named "host0
". The --network-veth
option implies
--private-network
.
Note that
systemd-networkd.service(8)
includes by default a network file /usr/lib/systemd/network/80-container-ve.network
matching the host-side interfaces created this way, which contains settings to enable automatic address
provisioning on the created virtual link via DHCP, as well as automatic IP routing onto the host's external
network interfaces. It also contains /usr/lib/systemd/network/80-container-host0.network
matching the container-side interface created this way, containing settings to enable client side address
assignment via DHCP. In case systemd-networkd
is running on both the host and inside the
container, automatic IP communication from the container to the host is thus available, with further
connectivity to the external network.
Note that --network-veth
is the default if the
systemd-nspawn@.service
template unit file is used.
Note that on Linux network interface names may have a length of 15 characters at maximum, while
container names may have a length up to 64 characters. As this option derives the host-side interface
name from the container name the name is possibly truncated. Thus, care needs to be taken to ensure
that interface names remain unique in this case, or even better container names are generally not
chosen longer than 12 characters, to avoid the truncation. If the name is truncated,
systemd-nspawn will automatically append a 4-digit hash value to the name to
reduce the chance of collisions. However, the hash algorithm is not collision-free. (See
systemd.net-naming-scheme(7)
for details on older naming algorithms for this interface). Alternatively, the
--network-veth-extra=
option may be used, which allows free configuration of the
host-side interface name independently of the container name — but might require a bit more
additional configuration in case bridging in a fashion similar to --network-bridge=
is desired.
--network-veth-extra=
¶Adds an additional virtual Ethernet link
between host and container. Takes a colon-separated pair of
host interface name and container interface name. The latter
may be omitted in which case the container and host sides will
be assigned the same name. This switch is independent of
--network-veth
, and — in contrast — may be
used multiple times, and allows configuration of the network
interface names. Note that --network-bridge=
has no effect on interfaces created with
--network-veth-extra=
.
--network-bridge=
¶Adds the host side of the Ethernet link created with --network-veth
to the specified Ethernet bridge interface. Expects a valid network interface name of a bridge device
as argument. Note that --network-bridge=
implies --network-veth
. If
this option is used, the host side of the Ethernet link will use the "vb-
" prefix
instead of "ve-
". Regardless of the used naming prefix the same network interface
name length limits imposed by Linux apply, along with the complications this creates (for details see
above).
As with --network-interface=
, the underlying bridge network interface must
already exist at the time the container is started, and thus similar unit file drop-ins as described
above might be useful.
--network-zone=
¶Creates a virtual Ethernet link ("veth
") to the container and adds it to an
automatically managed Ethernet bridge interface. The bridge interface is named after the passed argument,
prefixed with "vz-
". The bridge interface is automatically created when the first container
configured for its name is started, and is automatically removed when the last container configured for its
name exits. Hence, each bridge interface configured this way exists only as long as there's at least one
container referencing it running. This option is very similar to --network-bridge=
, besides
this automatic creation/removal of the bridge device.
This setting makes it easy to place multiple related containers on a common, virtual Ethernet-based
broadcast domain, here called a "zone". Each container may only be part of one zone, but each zone may contain
any number of containers. Each zone is referenced by its name. Names may be chosen freely (as long as they form
valid network interface names when prefixed with "vz-
"), and it is sufficient to pass the same
name to the --network-zone=
switch of the various concurrently running containers to join
them in one zone.
Note that
systemd-networkd.service(8)
includes by default a network file /usr/lib/systemd/network/80-container-vz.network
matching the bridge interfaces created this way, which contains settings to enable automatic address
provisioning on the created virtual network via DHCP, as well as automatic IP routing onto the host's external
network interfaces. Using --network-zone=
is hence in most cases fully automatic and
sufficient to connect multiple local containers in a joined broadcast domain to the host, with further
connectivity to the external network.
--network-namespace-path=
¶Takes the path to a file representing a kernel
network namespace that the container shall run in. The specified path
should refer to a (possibly bind-mounted) network namespace file, as
exposed by the kernel below /proc/$PID/ns/net
.
This makes the container enter the given network namespace. One of the
typical use cases is to give a network namespace under
/run/netns
created by ip-netns(8),
for example, --network-namespace-path=/run/netns/foo
.
Note that this option cannot be used together with other
network-related options, such as --private-network
or --network-interface=
.
-p
, --port=
¶If private networking is enabled, maps an IP
port on the host onto an IP port on the container. Takes a
protocol specifier (either "tcp
" or
"udp
"), separated by a colon from a host port
number in the range 1 to 65535, separated by a colon from a
container port number in the range from 1 to 65535. The
protocol specifier and its separating colon may be omitted, in
which case "tcp
" is assumed. The container
port number and its colon may be omitted, in which case the
same port as the host port is implied. This option is only
supported if private networking is used, such as with
--network-veth
, --network-zone=
--network-bridge=
.
--capability=
¶List one or more additional capabilities to grant the container. Takes a
comma-separated list of capability names, see capabilities(7)
for more information. Note that the following capabilities will be granted in any way:
CAP_AUDIT_CONTROL
, CAP_AUDIT_WRITE
,
CAP_CHOWN
, CAP_DAC_OVERRIDE
,
CAP_DAC_READ_SEARCH
, CAP_FOWNER
,
CAP_FSETID
, CAP_IPC_OWNER
, CAP_KILL
,
CAP_LEASE
, CAP_LINUX_IMMUTABLE
,
CAP_MKNOD
, CAP_NET_BIND_SERVICE
,
CAP_NET_BROADCAST
, CAP_NET_RAW
,
CAP_SETFCAP
, CAP_SETGID
, CAP_SETPCAP
,
CAP_SETUID
, CAP_SYS_ADMIN
,
CAP_SYS_BOOT
, CAP_SYS_CHROOT
,
CAP_SYS_NICE
, CAP_SYS_PTRACE
,
CAP_SYS_RESOURCE
, CAP_SYS_TTY_CONFIG
. Also
CAP_NET_ADMIN
is retained if --private-network
is specified.
If the special value "all
" is passed, all capabilities are retained.
If the special value of "help
" is passed, the program will print known
capability names and exit.
This option sets the bounding set of capabilities which
also limits the ambient capabilities as given with the
--ambient-capability=
.
--drop-capability=
¶Specify one or more additional capabilities to drop for the container. This allows running the container with fewer capabilities than the default (see above).
If the special value of "help
" is passed, the program will print known
capability names and exit.
This option sets the bounding set of capabilities which
also limits the ambient capabilities as given with the
--ambient-capability=
.
--ambient-capability=
¶Specify one or more additional capabilities to
pass in the inheritable and ambient set to the program started
within the container. The value "all
" is not
supported for this setting.
All capabilities specified here must be in the set
allowed with the --capability=
and
--drop-capability=
options. Otherwise, an
error message will be shown.
This option cannot be combined with the boot mode of the
container (as requested via --boot
).
If the special value of "help
" is
passed, the program will print known capability names and
exit.
--no-new-privileges=
¶Takes a boolean argument. Specifies the value of the
PR_SET_NO_NEW_PRIVS
flag for the container payload. Defaults to off. When turned
on the payload code of the container cannot acquire new privileges, i.e. the "setuid" file bit as
well as file system capabilities will not have an effect anymore. See prctl(2) for
details about this flag.
--system-call-filter=
¶Alter the system call filter
applied to containers. Takes a space-separated list of system call names or group names (the latter
prefixed with "@
", as listed by the syscall-filter command of
systemd-analyze(1)). Passed
system calls will be permitted. The list may optionally be prefixed by "~
", in which
case all listed system calls are prohibited. If this command line option is used multiple times the
configured lists are combined. If both a positive and a negative list (that is one system call list
without and one with the "~
" prefix) are configured, the negative list takes
precedence over the positive list. Note that systemd-nspawn always implements a
system call allow list (as opposed to a deny list!), and this command line option hence adds or
removes entries from the default allow list, depending on the "~
" prefix. Note that
the applied system call filter is also altered implicitly if additional capabilities are passed using
the --capabilities=.
-Z
, --selinux-context=
¶Sets the SELinux security context to be used to label processes in the container.
-L
, --selinux-apifs-context=
¶Sets the SELinux security context to be used to label files in the virtual API file systems in the container.
--rlimit=
¶Sets the specified POSIX resource limit for the container payload. Expects an assignment of the
form
"
"
or "LIMIT
=SOFT
:HARD
", where
LIMIT
=VALUE
LIMIT
should refer to a resource limit type, such as
RLIMIT_NOFILE
or RLIMIT_NICE
. The SOFT
and
HARD
fields should refer to the numeric soft and hard resource limit values. If the
second form is used, VALUE
may specify a value that is used both as soft and hard
limit. In place of a numeric value the special string "infinity
" may be used to turn off
resource limiting for the specific type of resource. This command line option may be used multiple times to
control limits on multiple limit types. If used multiple times for the same limit type, the last use
wins. For details about resource limits see setrlimit(2). By default
resource limits for the container's init process (PID 1) are set to the same values the Linux kernel originally
passed to the host init system. Note that some resource limits are enforced on resources counted per user, in
particular RLIMIT_NPROC
. This means that unless user namespacing is deployed
(i.e. --private-users=
is used, see above), any limits set will be applied to the resource
usage of the same user on all local containers as well as the host. This means particular care needs to be
taken with these limits as they might be triggered by possibly less trusted code. Example:
"--rlimit=RLIMIT_NOFILE=8192:16384
".
--oom-score-adjust=
¶Changes the OOM ("Out Of Memory") score adjustment value for the container payload. This controls
/proc/self/oom_score_adj
which influences the preference with which this container is
terminated when memory becomes scarce. For details see proc(5). Takes an
integer in the range -1000…1000.
--cpu-affinity=
¶Controls the CPU affinity of the container payload. Takes a comma separated list of CPU numbers or number ranges (the latter's start and end value separated by dashes). See sched_setaffinity(2) for details.
--personality=
¶Control the architecture ("personality")
reported by
uname(2)
in the container. Currently, only "x86
" and
"x86-64
" are supported. This is useful when
running a 32-bit container on a 64-bit host. If this setting
is not used, the personality reported in the container is the
same as the one reported on the host.
--resolv-conf=
¶Configures how /etc/resolv.conf
inside of the container shall be
handled (i.e. DNS configuration synchronization from host to container). Takes one of
"off
", "copy-host
", "copy-static
",
"copy-uplink
", "copy-stub
", "replace-host
",
"replace-static
", "replace-uplink
",
"replace-stub
", "bind-host
", "bind-static
",
"bind-uplink
", "bind-stub
", "delete
" or
"auto
".
If set to "off
" the /etc/resolv.conf
file in the
container is left as it is included in the image, and neither modified nor bind mounted over.
If set to "copy-host
", the /etc/resolv.conf
file from the
host is copied into the container, unless the file exists already and is not a regular file (e.g. a
symlink). Similarly, if "replace-host
" is used the file is copied, replacing any
existing inode, including symlinks. Similarly, if "bind-host
" is used, the file is
bind mounted from the host into the container.
If set to "copy-static
", "replace-static
" or
"bind-static
" the static resolv.conf
file supplied with
systemd-resolved.service(8)
(specifically: /usr/lib/systemd/resolv.conf
) is copied or bind mounted into the
container.
If set to "copy-uplink
", "replace-uplink
" or
"bind-uplink
" the uplink resolv.conf
file managed by
systemd-resolved.service
(specifically:
/run/systemd/resolve/resolv.conf
) is copied or bind mounted into the
container.
If set to "copy-stub
", "replace-stub
" or
"bind-stub
" the stub resolv.conf
file managed by
systemd-resolved.service
(specifically:
/run/systemd/resolve/stub-resolv.conf
) is copied or bind mounted into the
container.
If set to "delete
" the /etc/resolv.conf
file in the
container is deleted if it exists.
Finally, if set to "auto
" the file is left as it is if private networking is
turned on (see --private-network
). Otherwise, if
systemd-resolved.service
is running its stub resolv.conf
file is used, and if not the host's /etc/resolv.conf
file. In the latter cases
the file is copied if the image is writable, and bind mounted otherwise.
It's recommended to use "copy-…
" or "replace-…
" if the
container shall be able to make changes to the DNS configuration on its own, deviating from the
host's settings. Otherwise "bind
" is preferable, as it means direct changes to
/etc/resolv.conf
in the container are not allowed, as it is a read-only bind
mount (but note that if the container has enough privileges, it might simply go ahead and unmount the
bind mount anyway). Note that both if the file is bind mounted and if it is copied no further
propagation of configuration is generally done after the one-time early initialization (this is
because the file is usually updated through copying and renaming). Defaults to
"auto
".
--timezone=
¶Configures how /etc/localtime
inside of the container
(i.e. local timezone synchronization from host to container) shall be handled. Takes one of
"off
", "copy
", "bind
", "symlink
",
"delete
" or "auto
". If set to "off
" the
/etc/localtime
file in the container is left as it is included in the image, and
neither modified nor bind mounted over. If set to "copy
" the
/etc/localtime
file of the host is copied into the container. Similarly, if
"bind
" is used, the file is bind mounted from the host into the container. If set to
"symlink
", a symlink is created pointing from /etc/localtime
in
the container to the timezone file in the container that matches the timezone setting on the host. If
set to "delete
", the file in the container is deleted, should it exist. If set to
"auto
" and the /etc/localtime
file of the host is a symlink,
then "symlink
" mode is used, and "copy
" otherwise, except if the
image is read-only in which case "bind
" is used instead. Defaults to
"auto
".
--link-journal=
¶Control whether the container's journal shall
be made visible to the host system. If enabled, allows viewing
the container's journal files from the host (but not vice
versa). Takes one of "no
",
"host
", "try-host
",
"guest
", "try-guest
",
"auto
". If "no
", the journal
is not linked. If "host
", the journal files
are stored on the host file system (beneath
/var/log/journal/
)
and the subdirectory is bind-mounted into the container at the
same location. If "machine-id
guest
", the journal files
are stored on the guest file system (beneath
/var/log/journal/
)
and the subdirectory is symlinked into the host at the same
location. "machine-id
try-host
" and
"try-guest
" do the same but do not fail if
the host does not have persistent journaling enabled, or if
the container is in the --ephemeral
mode. If
"auto
" (the default), and the right
subdirectory of /var/log/journal
exists,
it will be bind mounted into the container. If the
subdirectory does not exist, no linking is performed.
Effectively, booting a container once with
"guest
" or "host
" will link
the journal persistently if further on the default of
"auto
" is used.
Note that --link-journal=try-guest
is the default if the
systemd-nspawn@.service
template unit file is used.
-j
¶Equivalent to
--link-journal=try-guest
.
--bind=
, --bind-ro=
¶Bind mount a file or directory from the host into the container. Takes one of: a path
argument — in which case the specified path will be mounted from the host to the same path in the container, or
a colon-separated pair of paths — in which case the first specified path is the source in the host, and the
second path is the destination in the container, or a colon-separated triple of source path, destination path
and mount options. The source path may optionally be prefixed with a "+
" character. If so, the
source path is taken relative to the image's root directory. This permits setting up bind mounts within the
container image. The source path may be specified as empty string, in which case a temporary directory below
the host's /var/tmp/
directory is used. It is automatically removed when the container is
shut down. If the source path is not absolute, it is resolved relative to the current working directory.
The --bind-ro=
option creates read-only bind mounts. Backslash escapes are interpreted,
so "\:
" may be used to embed colons in either path. This option may be specified
multiple times for creating multiple independent bind mount points.
Mount options are comma-separated. rbind
and norbind
control whether
to create a recursive or a regular bind mount. Defaults to rbind
. noidmap
,
idmap
, rootidmap
and owneridmap
control ID mapping.
Using idmap
, rootidmap
or owneridmap
requires support
by the source filesystem for user/group ID mapped mounts. Defaults to noidmap
. With
x
being the container's UID range offset, y
being the length of the
container's UID range, and p
being the owner UID of the bind mount source inode on the host:
If noidmap
is used, any user z
in the range
0 … y
seen from inside of the container is mapped to x + z
in the
x … x + y
range on the host. Other host users are mapped to
nobody
inside the container.
If idmap
is used, any user z
in the UID range
0 … y
as seen from inside the container is mapped to the same z
in the same 0 … y
range on the host. Other host users are mapped to
nobody
inside the container.
If rootidmap
is used, the user 0
seen from inside
of the container is mapped to p
on the host. Other host users are mapped to
nobody
inside the container.
If owneridmap
is used, the owner of the target directory inside of the
container is mapped to p
on the host. Other host users are mapped to
nobody
inside the container.
Whichever ID mapping option is used, the same mapping will be used for users and groups IDs. If
rootidmap
or owneridmap
are used, the group owning the bind mounted directory
will have no effect.
Note that when this option is used in combination with --private-users
, the resulting
mount points will be owned by the nobody
user. That's because the mount and its files and
directories continue to be owned by the relevant host users and groups, which do not exist in the container,
and thus show up under the wildcard UID 65534 (nobody). If such bind mounts are created, it is recommended to
make them read-only, using --bind-ro=
. Alternatively you can use the "idmap" mount option to
map the filesystem IDs.
--bind-user=
¶Binds the home directory of the specified user on the host into the container. Takes the name of an existing user on the host as argument. May be used multiple times to bind multiple users into the container. This does three things:
The user's home directory is bind mounted from the host into
/run/host/home/
.
An additional UID/GID mapping is added that maps the host user's UID/GID to a container UID/GID, allocated from the 60514…60577 range.
A JSON user and group record is generated in /run/userdb/
that
describes the mapped user. It contains a minimized representation of the host's user record,
adjusted to the UID/GID and home directory path assigned to the user in the container. The
nss-systemd(8)
glibc NSS module will pick up these records from there and make them available in the container's
user/group databases.
The combination of the three operations above ensures that it is possible to log into the
container using the same account information as on the host. The user is only mapped transiently,
while the container is running, and the mapping itself does not result in persistent changes to the
container (except maybe for log messages generated at login time, and similar). Note that in
particular the UID/GID assignment in the container is not made persistently. If the user is mapped
transiently, it is best to not allow the user to make persistent changes to the container. If the
user leaves files or directories owned by the user, and those UIDs/GIDs are reused during later
container invocations (possibly with a different --bind-user=
mapping), those files
and directories will be accessible to the "new" user.
The user/group record mapping only works if the container contains systemd 249 or newer, with
nss-systemd properly configured in nsswitch.conf
. See
nss-systemd(8) for
details.
Note that the user record propagated from the host into the container will contain the UNIX
password hash of the user, so that seamless logins in the container are possible. If the container is
less trusted than the host it's hence important to use a strong UNIX password hash function
(e.g. yescrypt or similar, with the "$y$
" hash prefix).
When binding a user from the host into the container checks are executed to ensure that the
username is not yet known in the container. Moreover, it is checked that the UID/GID allocated for it
is not currently defined in the user/group databases of the container. Both checks directly access
the container's /etc/passwd
and /etc/group
, and thus might
not detect existing accounts in other databases.
This operation is only supported in combination with
--private-users=
/-U
.
--inaccessible=
¶Make the specified path inaccessible in the container. This over-mounts the specified path (which must exist in the container) with a file node of the same type that is empty and has the most restrictive access mode supported. This is an effective way to mask files, directories and other file system objects from the container payload. This option may be used more than once in case all specified paths are masked.
--tmpfs=
¶Mount a tmpfs file system into the container. Takes a single absolute path argument that
specifies where to mount the tmpfs instance to (in which case the directory access mode will be chosen as 0755,
owned by root/root), or optionally a colon-separated pair of path and mount option string that is used for
mounting (in which case the kernel default for access mode and owner will be chosen, unless otherwise
specified). Backslash escapes are interpreted in the path, so "\:
" may be used to embed colons
in the path.
Note that this option cannot be used to replace the root file system of the container with a temporary
file system. However, the --volatile=
option described below provides similar
functionality, with a focus on implementing stateless operating system images.
--overlay=
, --overlay-ro=
¶Combine multiple directory trees into one overlay file system and mount it into the container. Takes a list of colon-separated paths to the directory trees to combine and the destination mount point.
Backslash escapes are interpreted in the paths, so "\:
" may be used to embed
colons in the paths.
If three or more paths are specified, then the last specified path is the destination mount
point in the container, all paths specified before refer to directory trees on the host and are
combined in the specified order into one overlay file system. The left-most path is hence the lowest
directory tree, the second-to-last path the highest directory tree in the stacking order. If
--overlay-ro=
is used instead of --overlay=
, a read-only overlay
file system is created. If a writable overlay file system is created, all changes made to it are
written to the highest directory tree in the stacking order, i.e. the second-to-last specified.
If only two paths are specified, then the second specified path is used both as the top-level directory tree in the stacking order as seen from the host, as well as the mount point for the overlay file system in the container. At least two paths have to be specified.
The source paths may optionally be prefixed with "+
" character. If so they are
taken relative to the image's root directory. The uppermost source path may also be specified as an
empty string, in which case a temporary directory below the host's /var/tmp/
is
used. The directory is removed automatically when the container is shut down. This behaviour is
useful in order to make read-only container directories writable while the container is running. For
example, use "--overlay=+/var::/var
" in order to automatically overlay a writable
temporary directory on a read-only /var/
directory. If a source path is not
absolute, it is resolved relative to the current working directory.
For details about overlay file systems, see Overlay Filesystem.
Note that the semantics of overlay file systems are substantially different from normal file systems,
in particular regarding reported device and inode information. Device and inode information may
change for a file while it is being written to, and processes might see out-of-date versions of files
at times. Note that this switch automatically derives the "workdir=
" mount option
for the overlay file system from the top-level directory tree, making it a sibling of it. It is hence
essential that the top-level directory tree is not a mount point itself (since the working directory
must be on the same file system as the top-most directory tree). Also note that the
"lowerdir=
" mount option receives the paths to stack in the opposite order of this
switch.
Note that this option cannot be used to replace the root file system of the container with an overlay
file system. However, the --volatile=
option described above provides similar functionality,
with a focus on implementing stateless operating system images.
--console=MODE
¶Configures how to set up standard input, output and error output for the container
payload, as well as the /dev/console
device for the container. Takes one of
interactive
, read-only
, passive
,
pipe
or autopipe
. If interactive
, a pseudo-TTY is
allocated and made available as /dev/console
in the container. It is then
bi-directionally connected to the standard input and output passed to
systemd-nspawn. read-only
is similar but only the output of the
container is propagated and no input from the caller is read. If passive
, a pseudo
TTY is allocated, but it is not connected anywhere. In pipe
mode no pseudo TTY is
allocated, but the standard input, output and error output file descriptors passed to
systemd-nspawn are passed on — as they are — to the container payload, see the
following paragraph. Finally, autopipe
mode operates like
interactive
when systemd-nspawn is invoked on a terminal, and
like pipe
otherwise. Defaults to interactive
if
systemd-nspawn is invoked from a terminal, and read-only
otherwise.
In pipe
mode, /dev/console
will not exist in the
container. This means that the container payload generally cannot be a full init system as init
systems tend to require /dev/console
to be available. On the other hand, in this
mode container invocations can be used within shell pipelines. This is because intermediary pseudo
TTYs do not permit independent bidirectional propagation of the end-of-file (EOF) condition, which is
necessary for shell pipelines to work correctly. Note that the pipe
mode
should be used carefully, as passing arbitrary file descriptors to less trusted container
payloads might open up unwanted interfaces for access by the container payload. For example, if a
passed file descriptor refers to a TTY of some form, APIs such as TIOCSTI
may be
used to synthesize input that might be used for escaping the container. Hence pipe
mode should only be used if the payload is sufficiently trusted or when the standard
input/output/error output file descriptors are known safe, for example pipes.
--pipe
, -P
¶Equivalent to --console=pipe
.
--background=COLOR
¶Change the terminal background color to the specified ANSI color as long as the
container runs. The color specified should be an ANSI X3.64 SGR background color, i.e. strings such
as "40
", "41
", …, "47
", "48;2;…
",
"48;5;…
". See ANSI
Escape Code (Wikipedia) for details. Assign an empty string to disable any coloring.
--load-credential=ID
:PATH
, --set-credential=ID
:VALUE
¶Pass a credential to the container. These two options correspond to the
LoadCredential=
and SetCredential=
settings in unit files. See
systemd.exec(5) for
details about these concepts, as well as the syntax of the option's arguments.
Note: when systemd-nspawn runs as systemd system service it can propagate
the credentials it received via LoadCredential=
/SetCredential=
to the container payload. A systemd service manager running as PID 1 in the container can further
propagate them to the services it itself starts. It is thus possible to easily propagate credentials
from a parent service manager to a container manager service and from there into its payload. This
can even be done recursively.
In order to embed binary data into the credential data for --set-credential=
,
use C-style escaping (i.e. "\n
" to embed a newline, or "\x00
" to
embed a NUL
byte). Note that the invoking shell might already apply unescaping
once, hence this might require double escaping!
The
systemd-sysusers.service(8)
and
systemd-firstboot(1)
services read credentials configured this way for the purpose of configuring the container's root
user's password and shell, as well as system locale, keymap and timezone during the first boot
process of the container. This is particularly useful in combination with
--volatile=yes
where every single boot appears as first boot, since configuration
applied to /etc/
is lost on container reboot cycles. See the respective man
pages for details. Example:
# systemd-nspawn -i image.raw \ --volatile=yes \ --set-credential=firstboot.locale:de_DE.UTF-8 \ --set-credential=passwd.hashed-password.root:'$y$j9T$yAuRJu1o5HioZAGDYPU5d.$F64ni6J2y2nNQve90M/p0ZP0ECP/qqzipNyaY9fjGpC' \ -b
The above command line will invoke the specified image file image.raw
in
volatile mode, i.e. with empty /etc/
and /var/
. The
container payload will recognize this as a first boot, and will invoke
systemd-firstboot.service
, which then reads the two passed credentials to
configure the system's initial locale and root password.
$SYSTEMD_LOG_LEVEL
¶The maximum log level of emitted messages (messages with a higher
log level, i.e. less important ones, will be suppressed). Takes a comma-separated list of values. A
value may be either one of (in order of decreasing importance) emerg
,
alert
, crit
, err
,
warning
, notice
, info
,
debug
, or an integer in the range 0…7. See
syslog(3)
for more information. Each value may optionally be prefixed with one of console
,
syslog
, kmsg
or journal
followed by a
colon to set the maximum log level for that specific log target (e.g.
SYSTEMD_LOG_LEVEL=debug,console:info
specifies to log at debug level except when
logging to the console which should be at info level). Note that the global maximum log level takes
priority over any per target maximum log levels.
$SYSTEMD_LOG_COLOR
¶A boolean. If true, messages written to the tty will be colored according to priority.
This setting is only useful when messages are written directly to the terminal, because journalctl(1) and other tools that display logs will color messages based on the log level on their own.
$SYSTEMD_LOG_TIME
¶A boolean. If true, console log messages will be prefixed with a timestamp.
This setting is only useful when messages are written directly to the terminal or a file, because journalctl(1) and other tools that display logs will attach timestamps based on the entry metadata on their own.
$SYSTEMD_LOG_LOCATION
¶A boolean. If true, messages will be prefixed with a filename and line number in the source code where the message originates.
Note that the log location is often attached as metadata to journal entries anyway. Including it directly in the message text can nevertheless be convenient when debugging programs.
$SYSTEMD_LOG_TID
¶A boolean. If true, messages will be prefixed with the current numerical thread ID (TID).
Note that the this information is attached as metadata to journal entries anyway. Including it directly in the message text can nevertheless be convenient when debugging programs.
$SYSTEMD_LOG_TARGET
¶The destination for log messages. One of
console
(log to the attached tty), console-prefixed
(log to
the attached tty but with prefixes encoding the log level and "facility", see syslog(3),
kmsg
(log to the kernel circular log buffer), journal
(log to
the journal), journal-or-kmsg
(log to the journal if available, and to kmsg
otherwise), auto
(determine the appropriate log target automatically, the default),
null
(disable log output).
$SYSTEMD_LOG_RATELIMIT_KMSG
¶ Whether to ratelimit kmsg or not. Takes a boolean.
Defaults to "true
". If disabled, systemd will not ratelimit messages written to kmsg.
$SYSTEMD_PAGER
¶Pager to use when --no-pager
is not given; overrides
$PAGER
. If neither $SYSTEMD_PAGER
nor $PAGER
are set, a
set of well-known pager implementations are tried in turn, including
less(1) and
more(1), until one is found. If
no pager implementation is discovered no pager is invoked. Setting this environment variable to an empty string
or the value "cat
" is equivalent to passing --no-pager
.
Note: if $SYSTEMD_PAGERSECURE
is not set, $SYSTEMD_PAGER
(as well as $PAGER
) will be silently ignored.
$SYSTEMD_LESS
¶Override the options passed to less (by default
"FRSXMK
").
Users might want to change two options in particular:
K
¶This option instructs the pager to exit immediately when Ctrl+C is pressed. To allow less to handle Ctrl+C itself to switch back to the pager command prompt, unset this option.
If the value of $SYSTEMD_LESS
does not include "K
",
and the pager that is invoked is less,
Ctrl+C will be ignored by the
executable, and needs to be handled by the pager.
X
¶This option instructs the pager to not send termcap initialization and deinitialization strings to the terminal. It is set by default to allow command output to remain visible in the terminal even after the pager exits. Nevertheless, this prevents some pager functionality from working, in particular paged output cannot be scrolled with the mouse.
Note that setting the regular $LESS
environment variable has no effect
for less invocations by systemd tools.
See less(1) for more discussion.
$SYSTEMD_LESSCHARSET
¶Override the charset passed to less (by default "utf-8
", if
the invoking terminal is determined to be UTF-8 compatible).
Note that setting the regular $LESSCHARSET
environment variable has no effect
for less invocations by systemd tools.
$SYSTEMD_PAGERSECURE
¶Takes a boolean argument. When true, the "secure" mode of the pager is enabled; if
false, disabled. If $SYSTEMD_PAGERSECURE
is not set at all, secure mode is enabled
if the effective UID is not the same as the owner of the login session, see
geteuid(2)
and sd_pid_get_owner_uid(3).
In secure mode, LESSSECURE=1
will be set when invoking the pager, and the pager shall
disable commands that open or create new files or start new subprocesses. When
$SYSTEMD_PAGERSECURE
is not set at all, pagers which are not known to implement
secure mode will not be used. (Currently only
less(1)
implements secure mode.)
Note: when commands are invoked with elevated privileges, for example under sudo(8) or
pkexec(1), care
must be taken to ensure that unintended interactive features are not enabled. "Secure" mode for the
pager may be enabled automatically as describe above. Setting SYSTEMD_PAGERSECURE=0
or not removing it from the inherited environment allows the user to invoke arbitrary commands. Note
that if the $SYSTEMD_PAGER
or $PAGER
variables are to be
honoured, $SYSTEMD_PAGERSECURE
must be set too. It might be reasonable to completely
disable the pager using --no-pager
instead.
$SYSTEMD_COLORS
¶Takes a boolean argument. When true, systemd and related utilities
will use colors in their output, otherwise the output will be monochrome. Additionally, the variable can
take one of the following special values: "16
", "256
" to restrict the use
of colors to the base 16 or 256 ANSI colors, respectively. This can be specified to override the automatic
decision based on $TERM
and what the console is connected to.
$SYSTEMD_URLIFY
¶The value must be a boolean. Controls whether clickable links should be generated in
the output for terminal emulators supporting this. This can be specified to override the decision that
systemd makes based on $TERM
and other conditions.
Example 1. Download an Ubuntu TAR image and open a shell in it
# importctl pull-tar -mN https://cloud-images.ubuntu.com/jammy/current/jammy-server-cloudimg-amd64-root.tar.xz # systemd-nspawn -M jammy-server-cloudimg-amd64-root
This downloads and verifies the specified .tar
image, and then uses
systemd-nspawn(1) to
open a shell in it.
Example 2. Build and boot a minimal Fedora distribution in a container
# dnf -y --releasever=40 --installroot=/var/lib/machines/f40 \ --repo=fedora --repo=updates --setopt=install_weak_deps=False install \ passwd dnf fedora-release vim-minimal util-linux systemd systemd-networkd # systemd-nspawn -bD /var/lib/machines/f40
This installs a minimal Fedora distribution into the
directory /var/lib/machines/f40
and then boots that OS in a namespace container. Because the installation
is located underneath the standard /var/lib/machines/
directory, it is also possible to start the machine using
systemd-nspawn -M f40.
Example 3. Spawn a shell in a container of a minimal Debian unstable distribution
# debootstrap unstable ~/debian-tree/ # systemd-nspawn -D ~/debian-tree/
This installs a minimal Debian unstable distribution into
the directory ~/debian-tree/
and then
spawns a shell from this image in a namespace container.
debootstrap supports Debian, Ubuntu, and Tanglu out of the box, so the same command can be used to install any of those. For other distributions from the Debian family, a mirror has to be specified, see debootstrap(8).
Example 4. Boot a minimal Arch Linux distribution in a container
# pacstrap -c ~/arch-tree/ base # systemd-nspawn -bD ~/arch-tree/
This installs a minimal Arch Linux distribution into the
directory ~/arch-tree/
and then boots an OS
in a namespace container in it.
Example 5. Install the OpenSUSE Tumbleweed rolling distribution
# zypper --root=/var/lib/machines/tumbleweed ar -c \ https://download.opensuse.org/tumbleweed/repo/oss tumbleweed # zypper --root=/var/lib/machines/tumbleweed refresh # zypper --root=/var/lib/machines/tumbleweed install --no-recommends \ systemd shadow zypper openSUSE-release vim # systemd-nspawn -M tumbleweed passwd root # systemd-nspawn -M tumbleweed -b
Example 6. Boot into an ephemeral snapshot of the host system
# systemd-nspawn -D / -xb
This runs a copy of the host system in a snapshot which is removed immediately when the container exits. All file system changes made during runtime will be lost on shutdown, hence.
Example 7. Run a container with SELinux sandbox security contexts
# chcon system_u:object_r:svirt_sandbox_file_t:s0:c0,c1 -R /srv/container # systemd-nspawn -L system_u:object_r:svirt_sandbox_file_t:s0:c0,c1 \ -Z system_u:system_r:svirt_lxc_net_t:s0:c0,c1 -D /srv/container /bin/sh
Example 8. Run a container with an OSTree deployment
# systemd-nspawn -b -i ~/image.raw \ --pivot-root=/ostree/deploy/$OS/deploy/$CHECKSUM:/sysroot \ --bind=+/sysroot/ostree/deploy/$OS/var:/var