systemd Developer Lennart Poettering lennart@poettering.net machinectl 1 machinectl Control the systemd machine manager machinectl OPTIONS COMMAND NAME machinectl may be used to introspect and control the state of the systemd1 virtual machine and container registration manager systemd-machined.service8. machinectl may be used to execute operations on machines and images. Machines in this sense are considered running instances of: Virtual Machines (VMs) that virtualize hardware to run full operating system (OS) instances (including their kernels) in a virtualized environment on top of the host OS. Containers that share the hardware and OS kernel with the host OS, in order to run OS userspace instances on top the host OS. The host system itself. Machines are identified by names that follow the same rules as UNIX and DNS host names. For details, see below. Machines are instantiated from disk or file system images that frequently — but not necessarily — carry the same name as machines running from them. Images in this sense may be: Directory trees containing an OS, including the top-level directories /usr, /etc, and so on. btrfs subvolumes containing OS trees, similar to normal directory trees. Binary "raw" disk images containing MBR or GPT partition tables and Linux file system partitions. The file system tree of the host OS itself. The following options are understood: -p --property= When showing machine or image properties, limit the output to certain properties as specified by the argument. If not specified, all set properties are shown. The argument should be a property name, such as Name. If specified more than once, all properties with the specified names are shown. -a --all When showing machine or image properties, show all properties regardless of whether they are set or not. When listing VM or container images, do not suppress images beginning in a dot character (.). When cleaning VM or container images, remove all images, not just hidden ones. --value When printing properties with show, only print the value, and skip the property name and =. -l --full Do not ellipsize process tree entries. --kill-who= When used with kill, choose which processes to kill. Must be one of leader, or all to select whether to kill only the leader process of the machine or all processes of the machine. If omitted, defaults to all. -s --signal= When used with kill, choose which signal to send to selected processes. Must be one of the well-known signal specifiers, such as SIGTERM, SIGINT or SIGSTOP. If omitted, defaults to SIGTERM. --uid= When used with the shell command, chooses the user ID to open the interactive shell session as. If the argument to the shell command also specifies a user name, this option is ignored. If the name is not specified in either way, root will be used by default. Note that this switch is not supported for the login command (see below). -E NAME=VALUE --setenv=NAME=VALUE When used with the shell command, sets an environment variable to pass to the executed shell. Takes an environment variable name and value, separated by =. This switch may be used multiple times to set multiple environment variables. Note that this switch is not supported for the login command (see below). --mkdir When used with bind, creates the destination file or directory before applying the bind mount. Note that even though the name of this option suggests that it is suitable only for directories, this option also creates the destination file node to mount over if the object to mount is not a directory, but a regular file, device node, socket or FIFO. --read-only When used with bind, creates a read-only bind mount. When used with clone, import-raw or import-tar a read-only container or VM image is created. -n --lines= When used with status, controls the number of journal lines to show, counting from the most recent ones. Takes a positive integer argument. Defaults to 10. -o --output= When used with status, controls the formatting of the journal entries that are shown. For the available choices, see journalctl1. Defaults to short. --verify= When downloading a container or VM image, specify whether the image shall be verified before it is made available. Takes one of no, checksum and signature. If no, no verification is done. If checksum is specified, the download is checked for integrity after the transfer is complete, but no signatures are verified. If signature is specified, the checksum is verified and the image's signature is checked against a local keyring of trustable vendors. It is strongly recommended to set this option to signature if the server and protocol support this. Defaults to signature. --force When downloading a container or VM image, and a local copy by the specified local machine name already exists, delete it first and replace it by the newly downloaded image. --format= When used with the export-tar or export-raw commands, specifies the compression format to use for the resulting file. Takes one of uncompressed, xz, gzip, bzip2. By default, the format is determined automatically from the image file name passed. --max-addresses= When used with the list-machines command, limits the number of ip addresses output for every machine. Defaults to 1. All addresses can be requested with all as argument to --max-addresses . If the argument to --max-addresses is less than the actual number of addresses, ...follows the last address. If multiple addresses are to be written for a given machine, every address except the first one is on a new line and is followed by , if another address will be output afterwards. -q --quiet Suppresses additional informational output while running. -M --machine= Connect to systemd-machined.service8 running in a local container, to perform the specified operation within the container. The following commands are understood: list List currently running (online) virtual machines and containers. To enumerate machine images that can be started, use list-images (see below). Note that this command hides the special .host machine by default. Use the --all switch to show it. status NAME… Show runtime status information about one or more virtual machines and containers, followed by the most recent log data from the journal. This function is intended to generate human-readable output. If you are looking for computer-parsable output, use show instead. Note that the log data shown is reported by the virtual machine or container manager, and frequently contains console output of the machine, but not necessarily journal contents of the machine itself. show [NAME…] Show properties of one or more registered virtual machines or containers or the manager itself. If no argument is specified, properties of the manager will be shown. If a NAME is specified, properties of this virtual machine or container are shown. By default, empty properties are suppressed. Use --all to show those too. To select specific properties to show, use --property=. This command is intended to be used whenever computer-parsable output is required, and does not print the control group tree or journal entries. Use status if you are looking for formatted human-readable output. start NAME… Start a container as a system service, using systemd-nspawn1. This starts systemd-nspawn@.service, instantiated for the specified machine name, similar to the effect of systemctl start on the service name. systemd-nspawn looks for a container image by the specified name in /var/lib/machines/ (and other search paths, see below) and runs it. Use list-images (see below) for listing available container images to start. Note that systemd-machined.service8 also interfaces with a variety of other container and VM managers, systemd-nspawn is just one implementation of it. Most of the commands available in machinectl may be used on containers or VMs controlled by other managers, not just systemd-nspawn. Starting VMs and container images on those managers requires manager-specific tools. To interactively start a container on the command line with full access to the container's console, please invoke systemd-nspawn directly. To stop a running container use machinectl poweroff. login [NAME] Open an interactive terminal login session in a container or on the local host. If an argument is supplied, it refers to the container machine to connect to. If none is specified, or the container name is specified as the empty string, or the special machine name .host (see below) is specified, the connection is made to the local host instead. This will create a TTY connection to a specific container or the local host and asks for the execution of a getty on it. Note that this is only supported for containers running systemd1 as init system. This command will open a full login prompt on the container or the local host, which then asks for username and password. Use shell (see below) or systemd-run1 with the --machine= switch to directly invoke a single command, either interactively or in the background. shell [[NAME@]NAME [PATH [ARGUMENTS…]]] Open an interactive shell session in a container or on the local host. The first argument refers to the container machine to connect to. If none is specified, or the machine name is specified as the empty string, or the special machine name .host (see below) is specified, the connection is made to the local host instead. This works similar to login but immediately invokes a user process. This command runs the specified executable with the specified arguments, or the default shell for the user if none is specified, or /bin/sh if no default shell is found. By default, --uid=, or by prefixing the machine name with a username and an @ character, a different user may be selected. Use --setenv= to set environment variables for the executed process. Note that machinectl shell does not propagate the exit code/status of the invoked shell process. Use systemd-run instead if that information is required (see below). When using the shell command without arguments, (thus invoking the executed shell or command on the local host), it is in many ways similar to a su1 session, but, unlike su, completely isolates the new session from the originating session, so that it shares no process or session properties, and is in a clean and well-defined state. It will be tracked in a new utmp, login, audit, security and keyring session, and will not inherit any environment variables or resource limits, among other properties. Note that systemd-run1 with its --machine= switch may be used in place of the machinectl shell command, and allows non-interactive operation, more detailed and low-level configuration of the invoked unit, as well as access to runtime and exit code/status information of the invoked shell process. In particular, use systemd-run's --wait switch to propagate exit status information of the invoked process. Use systemd-run's --pty switch for acquiring an interactive shell, similar to machinectl shell. In general, systemd-run is preferable for scripting purposes. However, note that systemd-run might require higher privileges than machinectl shell. enable NAME… disable NAME… Enable or disable a container as a system service to start at system boot, using systemd-nspawn1. This enables or disables systemd-nspawn@.service, instantiated for the specified machine name, similar to the effect of systemctl enable or systemctl disable on the service name. poweroff NAME… Power off one or more containers. This will trigger a reboot by sending SIGRTMIN+4 to the container's init process, which causes systemd-compatible init systems to shut down cleanly. Use stop as alias for poweroff. This operation does not work on containers that do not run a systemd1-compatible init system, such as sysvinit. Use terminate (see below) to immediately terminate a container or VM, without cleanly shutting it down. reboot NAME… Reboot one or more containers. This will trigger a reboot by sending SIGINT to the container's init process, which is roughly equivalent to pressing Ctrl+Alt+Del on a non-containerized system, and is compatible with containers running any system manager. terminate NAME… Immediately terminates a virtual machine or container, without cleanly shutting it down. This kills all processes of the virtual machine or container and deallocates all resources attached to that instance. Use poweroff to issue a clean shutdown request. kill NAME… Send a signal to one or more processes of the virtual machine or container. This means processes as seen by the host, not the processes inside the virtual machine or container. Use --kill-who= to select which process to kill. Use --signal= to select the signal to send. bind NAME PATH [PATH] Bind mounts a file or directory from the host into the specified container. The first path argument is the source file or directory on the host, the second path argument is the destination file or directory in the container. When the latter is omitted, the destination path in the container is the same as the source path on the host. When combined with the --read-only switch, a ready-only bind mount is created. When combined with the --mkdir switch, the destination path is first created before the mount is applied. Note that this option is currently only supported for systemd-nspawn1 containers, and only if user namespacing (--private-users) is not used. This command supports bind mounting directories, regular files, device nodes, AF_UNIX socket nodes, as well as FIFOs. copy-to NAME PATH [PATH] Copies files or directories from the host system into a running container. Takes a container name, followed by the source path on the host and the destination path in the container. If the destination path is omitted, the same as the source path is used. If host and container share the same user and group namespace, file ownership by numeric user ID and group ID is preserved for the copy, otherwise all files and directories in the copy will be owned by the root user and group (UID/GID 0). copy-from NAME PATH [PATH] Copies files or directories from a container into the host system. Takes a container name, followed by the source path in the container the destination path on the host. If the destination path is omitted, the same as the source path is used. If host and container share the same user and group namespace, file ownership by numeric user ID and group ID is preserved for the copy, otherwise all files and directories in the copy will be owned by the root user and group (UID/GID 0). list-images Show a list of locally installed container and VM images. This enumerates all raw disk images and container directories and subvolumes in /var/lib/machines/ (and other search paths, see below). Use start (see above) to run a container off one of the listed images. Note that, by default, containers whose name begins with a dot (.) are not shown. To show these too, specify --all. Note that a special image .host always implicitly exists and refers to the image the host itself is booted from. image-status [NAME…] Show terse status information about one or more container or VM images. This function is intended to generate human-readable output. Use show-image (see below) to generate computer-parsable output instead. show-image [NAME…] Show properties of one or more registered virtual machine or container images, or the manager itself. If no argument is specified, properties of the manager will be shown. If a NAME is specified, properties of this virtual machine or container image are shown. By default, empty properties are suppressed. Use --all to show those too. To select specific properties to show, use --property=. This command is intended to be used whenever computer-parsable output is required. Use image-status if you are looking for formatted human-readable output. clone NAME NAME Clones a container or VM image. The arguments specify the name of the image to clone and the name of the newly cloned image. Note that plain directory container images are cloned into btrfs subvolume images with this command, if the underlying file system supports this. Note that cloning a container or VM image is optimized for file systems that support copy-on-write, and might not be efficient on others, due to file system limitations. Note that this command leaves host name, machine ID and all other settings that could identify the instance unmodified. The original image and the cloned copy will hence share these credentials, and it might be necessary to manually change them in the copy. If combined with the --read-only switch a read-only cloned image is created. rename NAME NAME Renames a container or VM image. The arguments specify the name of the image to rename and the new name of the image. read-only NAME [BOOL] Marks or (unmarks) a container or VM image read-only. Takes a VM or container image name, followed by a boolean as arguments. If the boolean is omitted, positive is implied, i.e. the image is marked read-only. remove NAME… Removes one or more container or VM images. The special image .host, which refers to the host's own directory tree, may not be removed. set-limit [NAME] BYTES Sets the maximum size in bytes that a specific container or VM image, or all images, may grow up to on disk (disk quota). Takes either one or two parameters. The first, optional parameter refers to a container or VM image name. If specified, the size limit of the specified image is changed. If omitted, the overall size limit of the sum of all images stored locally is changed. The final argument specifies the size limit in bytes, possibly suffixed by the usual K, M, G, T units. If the size limit shall be disabled, specify - as size. Note that per-container size limits are only supported on btrfs file systems. Also note that, if set-limit is invoked without an image parameter, and /var/lib/machines is empty, and the directory is not located on btrfs, a btrfs loopback file is implicitly created as /var/lib/machines.raw with the given size, and mounted to /var/lib/machines. The size of the loopback may later be readjusted with set-limit, as well. If such a loopback-mounted /var/lib/machines directory is used, set-limit without an image name alters both the quota setting within the file system as well as the loopback file and file system size itself. clean Remove hidden VM or container images (or all). This command removes all hidden machine images from /var/lib/machines, i.e. those whose name begins with a dot. Use machinectl list-images --all to see a list of all machine images, including the hidden ones. When combined with the --all switch removes all images, not just hidden ones. This command effectively empties /var/lib/machines. Note that commands such as machinectl pull-tar or machinectl pull-raw usually create hidden, read-only, unmodified machine images from the downloaded image first, before cloning a writable working copy of it, in order to avoid duplicate downloads in case of images that are reused multiple times. Use machinectl clean to remove old, hidden images created this way. pull-tar URL [NAME] Downloads a .tar container image from the specified URL, and makes it available under the specified local machine name. The URL must be of type http:// or https://, and must refer to a .tar, .tar.gz, .tar.xz or .tar.bz2 archive file. If the local machine name is omitted, it is automatically derived from the last component of the URL, with its suffix removed. The image is verified before it is made available, unless --verify=no is specified. Verification is done either via an inline signed file with the name of the image and the suffix .sha256 or via separate SHA256SUMS and SHA256SUMS.gpg files. The signature files need to be made available on the same web server, under the same URL as the .tar file. With --verify=checksum, only the SHA256 checksum for the file is verified, based on the .sha256 suffixed file or theSHA256SUMS file. With --verify=signature, the sha checksum file is first verified with the inline signature in the .sha256 file or the detached GPG signature file SHA256SUMS.gpg. The public key for this verification step needs to be available in /usr/lib/systemd/import-pubring.gpg or /etc/systemd/import-pubring.gpg. The container image will be downloaded and stored in a read-only subvolume in /var/lib/machines/ that is named after the specified URL and its HTTP etag. A writable snapshot is then taken from this subvolume, and named after the specified local name. This behavior ensures that creating multiple container instances of the same URL is efficient, as multiple downloads are not necessary. In order to create only the read-only image, and avoid creating its writable snapshot, specify - as local machine name. Note that the read-only subvolume is prefixed with .tar-, and is thus not shown by list-images, unless --all is passed. Note that pressing C-c during execution of this command will not abort the download. Use cancel-transfer, described below. pull-raw URL [NAME] Downloads a .raw container or VM disk image from the specified URL, and makes it available under the specified local machine name. The URL must be of type http:// or https://. The container image must either be a .qcow2 or raw disk image, optionally compressed as .gz, .xz, or .bz2. If the local machine name is omitted, it is automatically derived from the last component of the URL, with its suffix removed. Image verification is identical for raw and tar images (see above). If the downloaded image is in .qcow2 format it is converted into a raw image file before it is made available. Downloaded images of this type will be placed as read-only .raw file in /var/lib/machines/. A local, writable (reflinked) copy is then made under the specified local machine name. To omit creation of the local, writable copy pass - as local machine name. Similar to the behavior of pull-tar, the read-only image is prefixed with .raw-, and thus not shown by list-images, unless --all is passed. Note that pressing C-c during execution of this command will not abort the download. Use cancel-transfer, described below. import-tar FILE [NAME] import-raw FILE [NAME] Imports a TAR or RAW container or VM image, and places it under the specified name in /var/lib/machines/. When import-tar is used, the file specified as the first argument should be a tar archive, possibly compressed with xz, gzip or bzip2. It will then be unpacked into its own subvolume in /var/lib/machines. When import-raw is used, the file should be a qcow2 or raw disk image, possibly compressed with xz, gzip or bzip2. If the second argument (the resulting image name) is not specified, it is automatically derived from the file name. If the filename is passed as -, the image is read from standard input, in which case the second argument is mandatory. Both pull-tar and pull-raw will resize /var/lib/machines.raw and the filesystem therein as necessary. Optionally, the --read-only switch may be used to create a read-only container or VM image. No cryptographic validation is done when importing the images. Much like image downloads, ongoing imports may be listed with list-transfers and aborted with cancel-transfer. export-tar NAME [FILE] export-raw NAME [FILE] Exports a TAR or RAW container or VM image and stores it in the specified file. The first parameter should be a VM or container image name. The second parameter should be a file path the TAR or RAW image is written to. If the path ends in .gz, the file is compressed with gzip, if it ends in .xz, with xz, and if it ends in .bz2, with bzip2. If the path ends in neither, the file is left uncompressed. If the second argument is missing, the image is written to standard output. The compression may also be explicitly selected with the --format= switch. This is in particular useful if the second parameter is left unspecified. Much like image downloads and imports, ongoing exports may be listed with list-transfers and aborted with cancel-transfer. Note that, currently, only directory and subvolume images may be exported as TAR images, and only raw disk images as RAW images. list-transfers Shows a list of container or VM image downloads, imports and exports that are currently in progress. cancel-transfer ID… Aborts a download, import or export of the container or VM image with the specified ID. To list ongoing transfers and their IDs, use list-transfers. The machinectl tool operates on machines and images whose names must be chosen following strict rules. Machine names must be suitable for use as host names following a conservative subset of DNS and UNIX/Linux semantics. Specifically, they must consist of one or more non-empty label strings, separated by dots. No leading or trailing dots are allowed. No sequences of multiple dots are allowed. The label strings may only consist of alphanumeric characters as well as the dash and underscore. The maximum length of a machine name is 64 characters. A special machine with the name .host refers to the running host system itself. This is useful for execution operations or inspecting the host system as well. Note that machinectl list will not show this special machine unless the --all switch is specified. Requirements on image names are less strict, however, they must be valid UTF-8, must be suitable as file names (hence not be the single or double dot, and not include a slash), and may not contain control characters. Since many operations search for an image by the name of a requested machine, it is recommended to name images in the same strict fashion as machines. A special image with the name .host refers to the image of the running host system. It hence conceptually maps to the special .host machine name described above. Note that machinectl list-images will not show this special image either, unless --all is specified. Machine images are preferably stored in /var/lib/machines/, but are also searched for in /usr/local/lib/machines/ and /usr/lib/machines/. For compatibility reasons, the directory /var/lib/container/ is searched, too. Note that images stored below /usr are always considered read-only. It is possible to symlink machines images from other directories into /var/lib/machines/ to make them available for control with machinectl. Note that some image operations are only supported, efficient or atomic on btrfs file systems. Due to this, if the pull-tar, pull-raw, import-tar, import-raw and set-limit commands notice that /var/lib/machines is empty and not located on btrfs, they will implicitly set up a loopback file /var/lib/machines.raw containing a btrfs file system that is mounted to /var/lib/machines. The size of this loopback file may be controlled dynamically with set-limit. Disk images are understood by systemd-nspawn1 and machinectl in three formats: A simple directory tree, containing the files and directories of the container to boot. Subvolumes (on btrfs file systems), which are similar to the simple directories, described above. However, they have additional benefits, such as efficient cloning and quota reporting. "Raw" disk images, i.e. binary images of disks with a GPT or MBR partition table. Images of this type are regular files with the suffix .raw. See systemd-nspawn1 for more information on image formats, in particular its --directory= and --image= options. # machinectl pull-tar https://cloud-images.ubuntu.com/trusty/current/trusty-server-cloudimg-amd64-root.tar.gz # systemd-nspawn -M trusty-server-cloudimg-amd64-root This downloads and verifies the specified .tar image, and then uses systemd-nspawn1 to open a shell in it. # machinectl pull-raw --verify=no https://dl.fedoraproject.org/pub/fedora/linux/releases/27/CloudImages/x86_64/images/Fedora-Cloud-Base-27-1.6.x86_64.raw.xz # systemd-nspawn -M Fedora-Cloud-Base-27-1.6.x86_64 # passwd # exit # machinectl start Fedora-Cloud-Base-27-1.6.x86_64 # machinectl login Fedora-Cloud-Base-27-1.6.x86_64 This downloads the specified .raw image with verification disabled. Then, a shell is opened in it and a root password is set. Afterwards the shell is left, and the machine started as system service. With the last command a login prompt into the container is requested. # machinectl export-tar fedora myfedora.tar.xz Exports the container fedora as an xz-compressed tar file myfedora.tar.xz into the current directory. # machinectl shell --uid=lennart This creates a new shell session on the local host for the user ID lennart, in a su1-like fashion. On success, 0 is returned, a non-zero failure code otherwise. systemd1, systemd-machined.service8, systemd-nspawn1, systemd.special7, tar1, xz1, gzip1, bzip21