1 systemd System and Service Manager
4 http://0pointer.de/blog/projects/systemd.html
7 https://www.freedesktop.org/wiki/Software/systemd
10 git@github.com:systemd/systemd.git
11 https://github.com/systemd/systemd
14 https://lists.freedesktop.org/mailman/listinfo/systemd-devel
17 #systemd on irc.libera.chat
20 https://github.com/systemd/systemd/issues
28 LGPL-2.1-or-later for all code, exceptions noted in LICENSES/README.md
32 Linux kernel >= 4.2 for unified cgroup hierarchy support
33 Linux kernel >= 4.10 for cgroup-bpf egress and ingress hooks
34 Linux kernel >= 4.15 for cgroup-bpf device hook
35 Linux kernel >= 4.17 for cgroup-bpf socket address hooks
36 Linux kernel >= 5.3 for bounded-loops in BPF program
37 Linux kernel >= 5.4 for signed Verity images support
38 Linux kernel >= 5.7 for BPF links and the BPF LSM hook
40 Kernel Config Options:
42 CONFIG_CGROUPS (it is OK to disable all controllers)
47 CONFIG_UNIX (it requires CONFIG_NET, but every other flag in it is not necessary)
50 CONFIG_FHANDLE (libudev, mount and bind mount handling)
52 Kernel crypto/hash API
53 CONFIG_CRYPTO_USER_API_HASH
57 udev will fail to work with the legacy sysfs layout:
58 CONFIG_SYSFS_DEPRECATED=n
60 Legacy hotplug slows down the system and confuses udev:
61 CONFIG_UEVENT_HELPER_PATH=""
63 Userspace firmware loading is not supported and should
64 be disabled in the kernel:
65 CONFIG_FW_LOADER_USER_HELPER=n
67 Some udev rules and virtualization detection relies on it:
70 Support for some SCSI devices serial number retrieval, to
71 create additional symlinks in /dev/disk/ and /dev/tape:
74 Required for PrivateNetwork= in service units:
76 Note that systemd-localed.service and other systemd units use
77 PrivateNetwork so this is effectively required.
79 Required for PrivateUsers= in service units:
82 Optional but strongly recommended:
86 CONFIG_{TMPFS,EXT4_FS,XFS,BTRFS_FS,...}_POSIX_ACL
88 CONFIG_SECCOMP_FILTER (required for seccomp support)
89 CONFIG_CHECKPOINT_RESTORE (for the kcmp() syscall)
91 Required for CPUShares= in resource control unit settings
93 CONFIG_FAIR_GROUP_SCHED
95 Required for CPUQuota= in resource control unit settings
98 Required for IPAddressDeny=, IPAddressAllow=, IPIngressFilterPath=,
99 IPEgressFilterPath= in resource control unit settings
107 Required for SocketBind{Allow|Deny}=, RestrictNetworkInterfaces= in
108 resource control unit settings
119 Required for signed Verity images support:
120 CONFIG_DM_VERITY_VERIFY_ROOTHASH_SIG
122 Required for RestrictFileSystems= in service units:
126 CONFIG_DEBUG_INFO_BTF
127 CONFIG_LSM="...,bpf" or kernel booted with lsm="...,bpf".
129 We recommend to turn off Real-Time group scheduling in the
130 kernel when using systemd. RT group scheduling effectively
131 makes RT scheduling unavailable for most userspace, since it
132 requires explicit assignment of RT budgets to each unit whose
133 processes making use of RT. As there's no sensible way to
134 assign these budgets automatically this cannot really be
135 fixed, and it's best to disable group scheduling hence.
136 CONFIG_RT_GROUP_SCHED=n
138 It's a good idea to disable the implicit creation of networking bonding
139 devices by the kernel networking bonding module, so that the
140 automatically created "bond0" interface doesn't conflict with any such
141 device created by systemd-networkd (or other tools). Ideally there
142 would be a kernel compile-time option for this, but there currently
143 isn't. The next best thing is to make this change through a modprobe.d
144 drop-in. This is shipped by default, see modprobe.d/systemd.conf.
146 Required for systemd-nspawn:
147 CONFIG_DEVPTS_MULTIPLE_INSTANCES or Linux kernel >= 4.7
149 Required for systemd-oomd:
152 Note that kernel auditing is broken when used with systemd's
153 container code. When using systemd in conjunction with
154 containers, please make sure to either turn off auditing at
155 runtime using the kernel command line option "audit=0", or
156 turn it off at kernel compile time using:
158 If systemd is compiled with libseccomp support on
159 architectures which do not use socketcall() and where seccomp
160 is supported (this effectively means x86-64 and ARM, but
161 excludes 32-bit x86!), then nspawn will now install a
162 work-around seccomp filter that makes containers boot even
163 with audit being enabled. This works correctly only on kernels
164 3.14 and newer though. TL;DR: turn audit off, still.
168 libmount >= 2.30 (from util-linux)
169 (util-linux *must* be built without --enable-libmount-support-mtab)
170 libseccomp >= 2.3.1 (optional)
171 libblkid >= 2.24 (from util-linux) (optional)
172 libkmod >= 15 (optional)
173 PAM >= 1.1.2 (optional)
174 libcryptsetup (optional), >= 2.3.0 required for signed Verity images support
177 libbpf >= 0.2.0 (optional)
178 libfdisk >= 2.33 (from util-linux) (optional)
179 libselinux (optional)
181 liblz4 >= 1.3.0 / 130 (optional)
182 libzstd >= 1.4.0 (optional)
184 libqrencode (optional)
185 libmicrohttpd (optional)
187 libidn2 or libidn (optional)
188 gnutls >= 3.1.4 (optional, >= 3.6.0 is required to support DNS-over-TLS with gnutls)
189 openssl >= 1.1.0 (optional, required to support DNS-over-TLS with openssl)
190 elfutils >= 158 (optional)
192 tzdata >= 2014f (optional)
195 docbook-xsl (optional, required for documentation)
196 xsltproc (optional, required for documentation)
198 python-lxml (optional, required to build the indices)
200 meson >= 0.53.2 (>= 0.54.0 is required to build with 'meson compile')
202 gcc, awk, sed, grep, and similar tools
203 clang >= 10.0, llvm >= 10.0 (optional, required to build BPF programs
204 from source code in C)
205 gnu-efi >= 3.0.5 (optional, required for systemd-boot)
207 During runtime, you need the following additional
210 util-linux >= v2.27.1 required
211 dbus >= 1.4.0 (strictly speaking optional, but recommended)
212 NOTE: If using dbus < 1.9.18, you should override the default
213 policy directory (--with-dbuspolicydir=/etc/dbus-1/system.d).
217 To build in directory build/:
218 meson setup build/ && meson compile -C build/
220 Any configuration options can be specified as -Darg=value... arguments
221 to meson. After the build directory is initially configured, meson will
222 refuse to run again, and options must be changed with:
223 meson configure -Darg=value build/
224 meson configure without any arguments will print out available options and
225 their current values.
228 meson compile -v -C build/ some/target
230 sudo meson install -C build/
231 DESTDIR=... meson install -C build/
233 A tarball can be created with:
234 git archive --format=tar --prefix=systemd-222/ v222 | xz > systemd-222.tar.xz
236 When systemd-hostnamed is used, it is strongly recommended to
237 install nss-myhostname to ensure that, in a world of
238 dynamically changing hostnames, the hostname stays resolvable
239 under all circumstances. In fact, systemd-hostnamed will warn
240 if nss-myhostname is not installed.
242 nss-systemd must be enabled on systemd systems, as that's required for
243 DynamicUser= to work. Note that we ship services out-of-the-box that
244 make use of DynamicUser= now, hence enabling nss-systemd is not
247 Note that the build prefix for systemd must be /usr. (Moreover,
248 packages systemd relies on — such as D-Bus — really should use the same
249 prefix, otherwise you are on your own.) -Dsplit-usr=false (which is the
250 default and does not need to be specified) is the recommended setting.
251 -Dsplit-usr=true can be used to give a semblance of support for systems
252 with programs installed split between / and /usr. Moving everything
253 under /usr is strongly encouraged.
255 Additional packages are necessary to run some tests:
256 - busybox (used by test/TEST-13-NSPAWN-SMOKE)
257 - nc (used by test/TEST-12-ISSUE-3171)
259 - python3-evdev (used by hwdb parsing tests)
260 - strace (used by test/test-functions)
261 - capsh (optional, used by test-execute)
264 Default udev rules use the following standard system group
265 names, which need to be resolvable by getgrnam() at any time,
266 even in the very early boot stages, where no other databases
267 and network are available:
269 audio, cdrom, dialout, disk, input, kmem, kvm, lp, render, tape, tty, video
271 During runtime, the journal daemon requires the
272 "systemd-journal" system group to exist. New journal files will
273 be readable by this group (but not writable), which may be used
274 to grant specific users read access. In addition, system
275 groups "wheel" and "adm" will be given read-only access to
276 journal files using systemd-tmpfiles.service.
278 The journal remote daemon requires the
279 "systemd-journal-remote" system user and group to
280 exist. During execution this network facing service will drop
281 privileges and assume this uid/gid for security reasons.
283 Similarly, the network management daemon requires the
284 "systemd-network" system user and group to exist.
286 Similarly, the name resolution daemon requires the
287 "systemd-resolve" system user and group to exist.
289 Similarly, the coredump support requires the
290 "systemd-coredump" system user and group to exist.
293 systemd ships with four glibc NSS modules:
295 nss-myhostname resolves the local hostname to locally configured IP
296 addresses, as well as "localhost" to 127.0.0.1/::1.
298 nss-resolve enables DNS resolution via the systemd-resolved DNS/LLMNR
299 caching stub resolver "systemd-resolved".
301 nss-mymachines enables resolution of all local containers registered
302 with machined to their respective IP addresses.
304 nss-systemd enables resolution of users/group registered via the
305 User/Group Record Lookup API (https://systemd.io/USER_GROUP_API),
306 including all dynamically allocated service users. (See the
307 DynamicUser= setting in unit files.)
309 To make use of these NSS modules, please add them to the "hosts:",
310 "passwd:" and "group:" lines in /etc/nsswitch.conf. The "resolve"
311 module should replace the glibc "dns" module in this file (and don't
312 worry, it chain-loads the "dns" module if it can't talk to resolved).
314 The four modules should be used in the following order:
316 passwd: compat systemd
317 group: compat systemd
318 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
321 When calling "systemctl enable/disable/is-enabled" on a unit which is a
322 SysV init.d script, it calls /usr/lib/systemd/systemd-sysv-install;
323 this needs to translate the action into the distribution specific
324 mechanism such as chkconfig or update-rc.d. Packagers need to provide
325 this script if you need this functionality (you don't if you disabled
328 Please see src/systemctl/systemd-sysv-install.SKELETON for how this
329 needs to look like, and provide an implementation at the marked places.
332 systemd will warn during early boot if /usr is not already mounted at
333 this point (that means: either located on the same file system as / or
334 already mounted in the initrd). While in systemd itself very little
335 will break if /usr is on a separate, late-mounted partition, many of
336 its dependencies very likely will break sooner or later in one form or
337 another. For example, udev rules tend to refer to binaries in /usr,
338 binaries that link to libraries in /usr or binaries that refer to data
339 files in /usr. Since these breakages are not always directly visible,
340 systemd will warn about this, since this kind of file system setup is
341 not really supported anymore by the basic set of Linux OS components.
343 systemd requires that the /run mount point exists. systemd also
344 requires that /var/run is a symlink to /run.
346 For more information on this issue consult
347 https://www.freedesktop.org/wiki/Software/systemd/separate-usr-is-broken
349 To run systemd under valgrind, compile with meson option
350 -Dvalgrind=true and have valgrind development headers installed
351 (i.e. valgrind-devel or equivalent). Otherwise, false positives will be
352 triggered by code which violates some rules but is actually safe. Note
353 that valgrind generates nice output only on exit(), hence on shutdown
354 we don't execve() systemd-shutdown.
356 STABLE BRANCHES AND BACKPORTS:
357 Stable branches with backported patches are available in the
358 systemd-stable repo at https://github.com/systemd/systemd-stable.
360 Stable branches are started for certain releases of systemd and named
361 after them, e.g. v238-stable. Stable branches are managed by
362 distribution maintainers on an as needed basis. See
363 https://www.freedesktop.org/wiki/Software/systemd/Backports/ for some
364 more information and examples.