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 LGPLv2.1+ for all code
29 - except src/basic/MurmurHash2.c which is Public Domain
30 - except src/basic/siphash24.c which is CC0 Public Domain
31 - except src/journal/lookup3.c which is Public Domain
32 - except src/udev/* which is (currently still) GPLv2, GPLv2+
33 - except tools/chromiumos/* which is BSD-style
37 Linux kernel >= 4.2 for unified cgroup hierarchy support
38 Linux kernel >= 4.10 for cgroup-bpf egress and ingress hooks
39 Linux kernel >= 4.15 for cgroup-bpf device hook
40 Linux kernel >= 4.17 for cgroup-bpf socket address hooks
41 Linux kernel >= 5.3 for bounded-loops in BPF program
42 Linux kernel >= 5.4 for signed Verity images support
43 Linux kernel >= 5.7 for BPF links
45 Kernel Config Options:
47 CONFIG_CGROUPS (it is OK to disable all controllers)
52 CONFIG_UNIX (it requires CONFIG_NET, but every other flag in it is not necessary)
55 CONFIG_FHANDLE (libudev, mount and bind mount handling)
57 Kernel crypto/hash API
58 CONFIG_CRYPTO_USER_API_HASH
62 udev will fail to work with the legacy sysfs layout:
63 CONFIG_SYSFS_DEPRECATED=n
65 Legacy hotplug slows down the system and confuses udev:
66 CONFIG_UEVENT_HELPER_PATH=""
68 Userspace firmware loading is not supported and should
69 be disabled in the kernel:
70 CONFIG_FW_LOADER_USER_HELPER=n
72 Some udev rules and virtualization detection relies on it:
75 Support for some SCSI devices serial number retrieval, to
76 create additional symlinks in /dev/disk/ and /dev/tape:
79 Required for PrivateNetwork= in service units:
81 Note that systemd-localed.service and other systemd units use
82 PrivateNetwork so this is effectively required.
84 Required for PrivateUsers= in service units:
87 Optional but strongly recommended:
91 CONFIG_{TMPFS,EXT4_FS,XFS,BTRFS_FS,...}_POSIX_ACL
93 CONFIG_SECCOMP_FILTER (required for seccomp support)
94 CONFIG_CHECKPOINT_RESTORE (for the kcmp() syscall)
96 Required for CPUShares= in resource control unit settings
98 CONFIG_FAIR_GROUP_SCHED
100 Required for CPUQuota= in resource control unit settings
103 Required for IPAddressDeny=, IPAddressAllow=, IPIngressFilterPath=,
104 IPEgressFilterPath= in resource control unit settings
112 Required for SocketBind{Allow|Deny}=, RestrictNetworkInterfaces= in
113 resource control unit settings
124 Required for signed Verity images support:
125 CONFIG_DM_VERITY_VERIFY_ROOTHASH_SIG
127 We recommend to turn off Real-Time group scheduling in the
128 kernel when using systemd. RT group scheduling effectively
129 makes RT scheduling unavailable for most userspace, since it
130 requires explicit assignment of RT budgets to each unit whose
131 processes making use of RT. As there's no sensible way to
132 assign these budgets automatically this cannot really be
133 fixed, and it's best to disable group scheduling hence.
134 CONFIG_RT_GROUP_SCHED=n
136 It's a good idea to disable the implicit creation of networking bonding
137 devices by the kernel networking bonding module, so that the
138 automatically created "bond0" interface doesn't conflict with any such
139 device created by systemd-networkd (or other tools). Ideally there
140 would be a kernel compile-time option for this, but there currently
141 isn't. The next best thing is to make this change through a modprobe.d
142 drop-in. This is shipped by default, see modprobe.d/systemd.conf.
144 Required for systemd-nspawn:
145 CONFIG_DEVPTS_MULTIPLE_INSTANCES or Linux kernel >= 4.7
147 Required for systemd-oomd:
150 Note that kernel auditing is broken when used with systemd's
151 container code. When using systemd in conjunction with
152 containers, please make sure to either turn off auditing at
153 runtime using the kernel command line option "audit=0", or
154 turn it off at kernel compile time using:
156 If systemd is compiled with libseccomp support on
157 architectures which do not use socketcall() and where seccomp
158 is supported (this effectively means x86-64 and ARM, but
159 excludes 32-bit x86!), then nspawn will now install a
160 work-around seccomp filter that makes containers boot even
161 with audit being enabled. This works correctly only on kernels
162 3.14 and newer though. TL;DR: turn audit off, still.
166 libmount >= 2.30 (from util-linux)
167 (util-linux *must* be built without --enable-libmount-support-mtab)
168 libseccomp >= 2.3.1 (optional)
169 libblkid >= 2.24 (from util-linux) (optional)
170 libkmod >= 15 (optional)
171 PAM >= 1.1.2 (optional)
172 libcryptsetup (optional), >= 2.3.0 required for signed Verity images support
175 libbpf >= 0.2.0 (optional)
176 libfdisk >= 2.33 (from util-linux) (optional)
177 libselinux (optional)
179 liblz4 >= 1.3.0 / 130 (optional)
180 libzstd >= 1.4.0 (optional)
182 libqrencode (optional)
183 libmicrohttpd (optional)
185 libidn2 or libidn (optional)
186 gnutls >= 3.1.4 (optional, >= 3.6.0 is required to support DNS-over-TLS with gnutls)
187 openssl >= 1.1.0 (optional, required to support DNS-over-TLS with openssl)
188 elfutils >= 158 (optional)
190 tzdata >= 2014f (optional)
193 docbook-xsl (optional, required for documentation)
194 xsltproc (optional, required for documentation)
196 python-lxml (optional, required to build the indices)
198 meson >= 0.47 (>= 0.49 is required to build position-independent executables)
200 gcc, awk, sed, grep, and similar tools
201 clang >= 10.0, llvm >= 10.0 (optional, required to build BPF programs
202 from source code in C)
204 During runtime, you need the following additional
207 util-linux >= v2.27.1 required
208 dbus >= 1.4.0 (strictly speaking optional, but recommended)
209 NOTE: If using dbus < 1.9.18, you should override the default
210 policy directory (--with-dbuspolicydir=/etc/dbus-1/system.d).
214 To build in directory build/:
215 meson setup build/ && meson compile -C build/
217 Any configuration options can be specified as -Darg=value... arguments
218 to meson. After the build directory is initially configured, meson will
219 refuse to run again, and options must be changed with:
220 meson configure -Darg=value build/
221 meson configure without any arguments will print out available options and
222 their current values.
225 meson compile -v -C build/ some/target
227 sudo meson install -C build/
228 DESTDIR=... meson install -C build/
230 A tarball can be created with:
231 git archive --format=tar --prefix=systemd-222/ v222 | xz > systemd-222.tar.xz
233 When systemd-hostnamed is used, it is strongly recommended to
234 install nss-myhostname to ensure that, in a world of
235 dynamically changing hostnames, the hostname stays resolvable
236 under all circumstances. In fact, systemd-hostnamed will warn
237 if nss-myhostname is not installed.
239 nss-systemd must be enabled on systemd systems, as that's required for
240 DynamicUser= to work. Note that we ship services out-of-the-box that
241 make use of DynamicUser= now, hence enabling nss-systemd is not
244 Note that the build prefix for systemd must be /usr. (Moreover,
245 packages systemd relies on — such as D-Bus — really should use the same
246 prefix, otherwise you are on your own.) -Dsplit-usr=false (which is the
247 default and does not need to be specified) is the recommended setting.
248 -Dsplit-usr=true can be used to give a semblance of support for systems
249 with programs installed split between / and /usr. Moving everything
250 under /usr is strongly encouraged.
252 Additional packages are necessary to run some tests:
253 - busybox (used by test/TEST-13-NSPAWN-SMOKE)
254 - nc (used by test/TEST-12-ISSUE-3171)
256 - python3-evdev (used by hwdb parsing tests)
257 - strace (used by test/test-functions)
258 - capsh (optional, used by test-execute)
261 Default udev rules use the following standard system group
262 names, which need to be resolvable by getgrnam() at any time,
263 even in the very early boot stages, where no other databases
264 and network are available:
266 audio, cdrom, dialout, disk, input, kmem, kvm, lp, render, tape, tty, video
268 During runtime, the journal daemon requires the
269 "systemd-journal" system group to exist. New journal files will
270 be readable by this group (but not writable), which may be used
271 to grant specific users read access. In addition, system
272 groups "wheel" and "adm" will be given read-only access to
273 journal files using systemd-tmpfiles.service.
275 The journal remote daemon requires the
276 "systemd-journal-remote" system user and group to
277 exist. During execution this network facing service will drop
278 privileges and assume this uid/gid for security reasons.
280 Similarly, the network management daemon requires the
281 "systemd-network" system user and group to exist.
283 Similarly, the name resolution daemon requires the
284 "systemd-resolve" system user and group to exist.
286 Similarly, the coredump support requires the
287 "systemd-coredump" system user and group to exist.
290 systemd ships with four glibc NSS modules:
292 nss-myhostname resolves the local hostname to locally configured IP
293 addresses, as well as "localhost" to 127.0.0.1/::1.
295 nss-resolve enables DNS resolution via the systemd-resolved DNS/LLMNR
296 caching stub resolver "systemd-resolved".
298 nss-mymachines enables resolution of all local containers registered
299 with machined to their respective IP addresses.
301 nss-systemd enables resolution of users/group registered via the
302 User/Group Record Lookup API (https://systemd.io/USER_GROUP_API),
303 including all dynamically allocated service users. (See the
304 DynamicUser= setting in unit files.)
306 To make use of these NSS modules, please add them to the "hosts:",
307 "passwd:" and "group:" lines in /etc/nsswitch.conf. The "resolve"
308 module should replace the glibc "dns" module in this file (and don't
309 worry, it chain-loads the "dns" module if it can't talk to resolved).
311 The four modules should be used in the following order:
313 passwd: compat systemd
314 group: compat systemd
315 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
318 When calling "systemctl enable/disable/is-enabled" on a unit which is a
319 SysV init.d script, it calls /usr/lib/systemd/systemd-sysv-install;
320 this needs to translate the action into the distribution specific
321 mechanism such as chkconfig or update-rc.d. Packagers need to provide
322 this script if you need this functionality (you don't if you disabled
325 Please see src/systemctl/systemd-sysv-install.SKELETON for how this
326 needs to look like, and provide an implementation at the marked places.
329 systemd will warn during early boot if /usr is not already mounted at
330 this point (that means: either located on the same file system as / or
331 already mounted in the initrd). While in systemd itself very little
332 will break if /usr is on a separate, late-mounted partition, many of
333 its dependencies very likely will break sooner or later in one form or
334 another. For example, udev rules tend to refer to binaries in /usr,
335 binaries that link to libraries in /usr or binaries that refer to data
336 files in /usr. Since these breakages are not always directly visible,
337 systemd will warn about this, since this kind of file system setup is
338 not really supported anymore by the basic set of Linux OS components.
340 systemd requires that the /run mount point exists. systemd also
341 requires that /var/run is a symlink to /run.
343 For more information on this issue consult
344 https://www.freedesktop.org/wiki/Software/systemd/separate-usr-is-broken
346 To run systemd under valgrind, compile with meson option
347 -Dvalgrind=true and have valgrind development headers installed
348 (i.e. valgrind-devel or equivalent). Otherwise, false positives will be
349 triggered by code which violates some rules but is actually safe. Note
350 that valgrind generates nice output only on exit(), hence on shutdown
351 we don't execve() systemd-shutdown.
353 STABLE BRANCHES AND BACKPORTS:
354 Stable branches with backported patches are available in the
355 systemd-stable repo at https://github.com/systemd/systemd-stable.
357 Stable branches are started for certain releases of systemd and named
358 after them, e.g. v238-stable. Stable branches are managed by
359 distribution maintainers on an as needed basis. See
360 https://www.freedesktop.org/wiki/Software/systemd/Backports/ for some
361 more information and examples.