systemd userdbctl 1 userdbctl Inspect users, groups and group memberships userdbctl OPTIONS COMMAND NAME userdbctl may be used to inspect user and groups (as well as group memberships) of the system. This client utility inquires user/group information provided by various system services, both operating on JSON user/group records (as defined by the JSON User Records and JSON Group Records definitions), and classic UNIX NSS/glibc user and group records. This tool is primarily a client to the User/Group Record Lookup API via Varlink, and may also pick up drop-in JSON user and group records from /etc/userdb/, /run/userdb/, /run/host/userdb/, /usr/lib/userdb/. The following options are understood: --output=MODE Choose the output mode, takes one of classic, friendly, table, json. If classic, an output very close to the format of /etc/passwd or /etc/group is generated. If friendly a more comprehensive and user friendly, human readable output is generated; if table a minimal, tabular output is generated; if json a JSON formatted output is generated. Defaults to friendly if a user/group is specified on the command line, table otherwise. Note that most output formats do not show all available information. In particular, classic and table show only the most important fields. Various modes also do not show password hashes. Use json to view all fields, including any authentication fields. --json=FORMAT Selects JSON output mode (like --output=json) and selects the precise display mode. Takes one of pretty or short. If pretty, human-friendly whitespace and newlines are inserted in the output to make the JSON data more readable. If short, all superfluous whitespace is suppressed. --service=SERVICE:SERVICE… -s SERVICE:SERVICE… Controls which services to query for users/groups. Takes a list of one or more service names, separated by :. See below for a list of well-known service names. If not specified all available services are queried at once. --with-nss=BOOL Controls whether to include classic glibc/NSS user/group lookups in the output. If --with-nss=no is used any attempts to resolve or enumerate users/groups provided only via glibc NSS is suppressed. If --with-nss=yes is specified such users/groups are included in the output (which is the default). --with-varlink=BOOL Controls whether to include Varlink user/group lookups in the output, i.e. those done via the User/Group Record Lookup API via Varlink. If --with-varlink=no is used any attempts to resolve or enumerate users/groups provided only via Varlink are suppressed. If --with-varlink=yes is specified such users/groups are included in the output (which is the default). --with-dropin=BOOL Controls whether to include user/group lookups in the output that are defined using drop-in files in /etc/userdb/, /run/userdb/, /run/host/userdb/, /usr/lib/userdb/. If --with-dropin=no is used these records are suppressed. If --with-dropin=yes is specified such users/groups are included in the output (which is the default). --synthesize=BOOL Controls whether to synthesize records for the root and nobody users/groups if they aren't defined otherwise. By default (or yes) such records are implicitly synthesized if otherwise missing since they have special significance to the OS. When no this synthesizing is turned off. -N This option is short for --with-nss=no --synthesize=no. Use this option to show only records that are natively defined as JSON user or group records, with all NSS/glibc compatibility and all implicit synthesis turned off. --multiplexer=BOOL Controls whether to do lookups via the multiplexer service (if specified as true, the default) or do lookups in the client (if specified as false). Using the multiplexer service is typically preferable, since it runs in a locked down sandbox. --chain When used with the ssh-authorized-keys command, this will allow passing an additional command line after the user name that is chain executed after the lookup completed. This allows chaining multiple tools that show SSH authorized keys. --fuzzy -z When used with the user or group command, do a fuzzy string search. Any specified arguments will be matched against the user name, the real name of the user record, the email address, and other descriptive strings of the user or group record. Moreover, instead of precise matching, a substring match or a match allowing slight deviations in spelling is applied. --disposition= When used with the user or group command, filters by disposition of the record. Takes one of intrinsic, system, regular, dynamic, container. May be used multiple times, in which case only users matching any of the specified dispositions are shown. -I -S -R Shortcuts for --disposition=intrinsic, --disposition=system, --disposition=regular, respectively. --uid-min= --uid-max= When used with the user or group command, filters the output by UID/GID ranges. Takes numeric minimum resp. maximum UID/GID values. Shows only records within the specified range. When applied to the user command matches against UIDs, when applied to the group command against GIDs (despite the name of the switch). If unspecified defaults to 0 (for the minimum) and 4294967294 (for the maximum), i.e. by default no filtering is applied as the whole UID/GID range is covered. --boundaries= When used with the user or group command, controls whether to show relevant UID/GID range boundary information in the tabular output. Takes a boolean. Defaults to true. -B Shortcut for --boundaries=no. The following commands are understood: user USER… List all known users records or show details of one or more specified user records. Use --output= to tweak output mode. group GROUP… List all known group records or show details of one or more specified group records. Use --output= to tweak output mode. users-in-group GROUP… List users that are members of the specified groups. If no groups are specified list all user/group memberships defined. Use --output= to tweak output mode. groups-of-user USER… List groups that the specified users are members of. If no users are specified list all user/group memberships defined (in this case groups-of-user and users-in-group are equivalent). Use --output= to tweak output mode. services List all services currently providing user/group definitions to the system. See below for a list of well-known services providing user information. ssh-authorized-keys Show SSH authorized keys for this account. This command is intended to be used to allow the SSH daemon to pick up authorized keys from user records, see below. The userdbctl services command will list all currently running services that provide user or group definitions to the system. The following well-known services are shown among this list: io.systemd.DynamicUser This service is provided by the system service manager itself (i.e. PID 1) and makes all users (and their groups) synthesized through the DynamicUser= setting in service unit files available to the system (see systemd.exec5 for details about this setting). io.systemd.Home This service is provided by systemd-homed.service8 and makes all users (and their groups) belonging to home directories managed by that service available to the system. io.systemd.Machine This service is provided by systemd-machined.service8 and synthesizes records for all users/groups used by a container that employs user namespacing. io.systemd.Multiplexer This service is provided by systemd-userdbd.service8 and multiplexes user/group look-ups to all other running lookup services. This is the primary entry point for user/group record clients, as it simplifies client side implementation substantially since they can ask a single service for lookups instead of asking all running services in parallel. userdbctl uses this service preferably, too, unless --with-nss= or --service= are used, in which case finer control over the services to talk to is required. io.systemd.NameServiceSwitch This service is (also) provided by systemd-userdbd.service8 and converts classic NSS/glibc user and group records to JSON user/group records, providing full backwards compatibility. Use --with-nss=no to disable this compatibility, see above. Note that compatibility is actually provided in both directions: nss-systemd8 will automatically synthesize classic NSS/glibc user/group records from all JSON user/group records provided to the system, thus using both APIs is mostly equivalent and provides access to the same data, however the NSS/glibc APIs necessarily expose a more reduced set of fields only. io.systemd.DropIn This service is (also) provided by systemd-userdbd.service8 and picks up JSON user/group records from /etc/userdb/, /run/userdb/, /run/host/userdb/, /usr/lib/userdb/. Note that userdbctl has internal support for NSS-based lookups too. This means that if neither io.systemd.Multiplexer nor io.systemd.NameServiceSwitch are running look-ups into the basic user/group databases will still work. The userdbctl tool may be used to make the list of SSH authorized keys possibly contained in a user record available to the SSH daemon for authentication. For that configure the following in sshd_config5: … AuthorizedKeysCommand /usr/bin/userdbctl ssh-authorized-keys %u AuthorizedKeysCommandUser root … Sometimes it's useful to allow chain invocation of another program to list SSH authorized keys. By using the --chain such a tool may be chain executed by userdbctl ssh-authorized-keys once a lookup completes (regardless if an SSH key was found or not). Example: … AuthorizedKeysCommand /usr/bin/userdbctl ssh-authorized-keys %u --chain /usr/bin/othertool %u AuthorizedKeysCommandUser root … The above will first query the userdb database for SSH keys, and then chain execute /usr/bin/othertool to also be queried. On success, 0 is returned, a non-zero failure code otherwise. systemd1 systemd-userdbd.service8 systemd-homed.service8 nss-systemd8 getent1