1 systemd System and Service Manager
7 git@github.com:systemd/systemd.git
8 https://github.com/systemd/systemd
11 https://lists.freedesktop.org/mailman/listinfo/systemd-devel
14 #systemd on irc.libera.chat
17 https://github.com/systemd/systemd/issues
21 http://0pointer.de/blog/projects/systemd.html
22 https://www.freedesktop.org/wiki/Software/systemd
30 LGPL-2.1-or-later for all code, exceptions noted in LICENSES/README.md
34 ≥ 4.5 for pids controller in cgroup v2
35 ≥ 4.6 for cgroup namespaces
36 ≥ 4.9 for RENAME_NOREPLACE support in vfat
37 ≥ 4.10 for cgroup-bpf egress and ingress hooks
38 ≥ 4.15 for cgroup-bpf device hook and cpu controller in cgroup v2
39 ≥ 4.17 for cgroup-bpf socket address hooks
40 ≥ 5.3 for bounded loops in BPF program
41 ≥ 5.4 for signed Verity images
42 ≥ 5.7 for BPF links and the BPF LSM hook
44 Kernel versions below 4.15 have significant gaps in functionality and
45 are not recommended for use with this version of systemd. Taint flag
46 'old-kernel' will be set. Systemd will most likely still function, but
47 upstream support and testing are limited.
49 Kernel Config Options:
51 CONFIG_CGROUPS (it is OK to disable all controllers)
56 CONFIG_UNIX (it requires CONFIG_NET, but every other flag in it is not necessary)
59 CONFIG_FHANDLE (libudev, mount and bind mount handling)
61 Kernel crypto/hash API
62 CONFIG_CRYPTO_USER_API_HASH
66 udev will fail to work with the legacy sysfs layout:
67 CONFIG_SYSFS_DEPRECATED=n
69 Legacy hotplug slows down the system and confuses udev:
70 CONFIG_UEVENT_HELPER_PATH=""
72 Userspace firmware loading is not supported and should
73 be disabled in the kernel:
74 CONFIG_FW_LOADER_USER_HELPER=n
76 Some udev rules and virtualization detection relies on it:
79 Support for some SCSI devices serial number retrieval, to
80 create additional symlinks in /dev/disk/ and /dev/tape:
83 Required for PrivateNetwork= in service units:
85 Note that systemd-localed.service and other systemd units use
86 PrivateNetwork so this is effectively required.
88 Required for PrivateUsers= in service units:
91 Optional but strongly recommended:
95 CONFIG_{TMPFS,EXT4_FS,XFS,BTRFS_FS,...}_POSIX_ACL
97 CONFIG_SECCOMP_FILTER (required for seccomp support)
98 CONFIG_KCMP (for the kcmp() syscall, used to be under CONFIG_CHECKPOINT_RESTORE before ~5.12)
100 Required for CPUShares= in resource control unit settings
102 CONFIG_FAIR_GROUP_SCHED
104 Required for CPUQuota= in resource control unit settings
107 Required for IPAddressDeny=, IPAddressAllow=, IPIngressFilterPath=,
108 IPEgressFilterPath= in resource control unit settings
116 Required for SocketBind{Allow|Deny}=, RestrictNetworkInterfaces= in
117 resource control unit settings
128 Required for signed Verity images support:
129 CONFIG_DM_VERITY_VERIFY_ROOTHASH_SIG
131 Required for RestrictFileSystems= in service units:
135 CONFIG_DEBUG_INFO_BTF
136 CONFIG_LSM="...,bpf" or kernel booted with lsm="...,bpf".
138 We recommend to turn off Real-Time group scheduling in the
139 kernel when using systemd. RT group scheduling effectively
140 makes RT scheduling unavailable for most userspace, since it
141 requires explicit assignment of RT budgets to each unit whose
142 processes making use of RT. As there's no sensible way to
143 assign these budgets automatically this cannot really be
144 fixed, and it's best to disable group scheduling hence.
145 CONFIG_RT_GROUP_SCHED=n
147 It's a good idea to disable the implicit creation of networking bonding
148 devices by the kernel networking bonding module, so that the
149 automatically created "bond0" interface doesn't conflict with any such
150 device created by systemd-networkd (or other tools). Ideally there
151 would be a kernel compile-time option for this, but there currently
152 isn't. The next best thing is to make this change through a modprobe.d
153 drop-in. This is shipped by default, see modprobe.d/systemd.conf.
155 Required for systemd-nspawn:
156 CONFIG_DEVPTS_MULTIPLE_INSTANCES or Linux kernel >= 4.7
158 Required for systemd-oomd:
161 Note that kernel auditing is broken when used with systemd's
162 container code. When using systemd in conjunction with
163 containers, please make sure to either turn off auditing at
164 runtime using the kernel command line option "audit=0", or
165 turn it off at kernel compile time using:
167 If systemd is compiled with libseccomp support on
168 architectures which do not use socketcall() and where seccomp
169 is supported (this effectively means x86-64 and ARM, but
170 excludes 32-bit x86!), then nspawn will now install a
171 work-around seccomp filter that makes containers boot even
172 with audit being enabled. This works correctly only on kernels
173 3.14 and newer though. TL;DR: turn audit off, still.
177 libmount >= 2.30 (from util-linux)
178 (util-linux *must* be built without --enable-libmount-support-mtab)
179 libseccomp >= 2.3.1 (optional)
180 libblkid >= 2.24 (from util-linux) (optional)
181 libkmod >= 15 (optional)
182 PAM >= 1.1.2 (optional)
183 libcryptsetup (optional), >= 2.3.0 required for signed Verity images support
186 libbpf >= 0.2.0 (optional)
187 libfdisk >= 2.32 (from util-linux) (optional)
188 libselinux (optional)
190 liblz4 >= 1.3.0 / 130 (optional)
191 libzstd >= 1.4.0 (optional)
193 libqrencode (optional)
194 libmicrohttpd (optional)
196 libidn2 or libidn (optional)
197 gnutls >= 3.1.4 (optional, >= 3.6.0 is required to support DNS-over-TLS with gnutls)
198 openssl >= 1.1.0 (optional, required to support DNS-over-TLS with openssl)
199 elfutils >= 158 (optional)
201 tzdata >= 2014f (optional)
204 docbook-xsl (optional, required for documentation)
205 xsltproc (optional, required for documentation)
207 python-lxml (optional, required to build the indices)
209 meson >= 0.53.2 (>= 0.54.0 is required to build with 'meson compile')
211 gcc, awk, sed, grep, and similar tools
212 clang >= 10.0, llvm >= 10.0 (optional, required to build BPF programs
213 from source code in C)
214 gnu-efi >= 3.0.5 (optional, required for systemd-boot)
216 During runtime, you need the following additional
219 util-linux >= v2.27.1 required
220 dbus >= 1.4.0 (strictly speaking optional, but recommended)
221 NOTE: If using dbus < 1.9.18, you should override the default
222 policy directory (--with-dbuspolicydir=/etc/dbus-1/system.d).
226 To build in directory build/:
227 meson setup build/ && meson compile -C build/
229 Any configuration options can be specified as -Darg=value... arguments
230 to meson. After the build directory is initially configured, meson will
231 refuse to run again, and options must be changed with:
232 meson configure -Darg=value build/
233 meson configure without any arguments will print out available options and
234 their current values.
237 meson compile -v -C build/ some/target
239 sudo meson install -C build/
240 DESTDIR=... meson install -C build/
242 A tarball can be created with:
243 git archive --format=tar --prefix=systemd-222/ v222 | xz > systemd-222.tar.xz
245 When systemd-hostnamed is used, it is strongly recommended to
246 install nss-myhostname to ensure that, in a world of
247 dynamically changing hostnames, the hostname stays resolvable
248 under all circumstances. In fact, systemd-hostnamed will warn
249 if nss-myhostname is not installed.
251 nss-systemd must be enabled on systemd systems, as that's required for
252 DynamicUser= to work. Note that we ship services out-of-the-box that
253 make use of DynamicUser= now, hence enabling nss-systemd is not
256 Note that the build prefix for systemd must be /usr. (Moreover,
257 packages systemd relies on — such as D-Bus — really should use the same
258 prefix, otherwise you are on your own.) -Dsplit-usr=false (which is the
259 default and does not need to be specified) is the recommended setting.
260 -Dsplit-usr=true can be used to give a semblance of support for systems
261 with programs installed split between / and /usr. Moving everything
262 under /usr is strongly encouraged.
264 Additional packages are necessary to run some tests:
265 - busybox (used by test/TEST-13-NSPAWN-SMOKE)
266 - nc (used by test/TEST-12-ISSUE-3171)
268 - python3-evdev (used by hwdb parsing tests)
269 - strace (used by test/test-functions)
270 - capsh (optional, used by test-execute)
273 Default udev rules use the following standard system group
274 names, which need to be resolvable by getgrnam() at any time,
275 even in the very early boot stages, where no other databases
276 and network are available:
278 audio, cdrom, dialout, disk, input, kmem, kvm, lp, render, tape, tty, video
280 During runtime, the journal daemon requires the
281 "systemd-journal" system group to exist. New journal files will
282 be readable by this group (but not writable), which may be used
283 to grant specific users read access. In addition, system
284 groups "wheel" and "adm" will be given read-only access to
285 journal files using systemd-tmpfiles.service.
287 The journal remote daemon requires the
288 "systemd-journal-remote" system user and group to
289 exist. During execution this network facing service will drop
290 privileges and assume this uid/gid for security reasons.
292 Similarly, the network management daemon requires the
293 "systemd-network" system user and group to exist.
295 Similarly, the name resolution daemon requires the
296 "systemd-resolve" system user and group to exist.
298 Similarly, the coredump support requires the
299 "systemd-coredump" system user and group to exist.
302 systemd ships with four glibc NSS modules:
304 nss-myhostname resolves the local hostname to locally configured IP
305 addresses, as well as "localhost" to 127.0.0.1/::1.
307 nss-resolve enables DNS resolution via the systemd-resolved DNS/LLMNR
308 caching stub resolver "systemd-resolved".
310 nss-mymachines enables resolution of all local containers registered
311 with machined to their respective IP addresses.
313 nss-systemd enables resolution of users/group registered via the
314 User/Group Record Lookup API (https://systemd.io/USER_GROUP_API),
315 including all dynamically allocated service users. (See the
316 DynamicUser= setting in unit files.)
318 To make use of these NSS modules, please add them to the "hosts:",
319 "passwd:" and "group:" lines in /etc/nsswitch.conf. The "resolve"
320 module should replace the glibc "dns" module in this file (and don't
321 worry, it chain-loads the "dns" module if it can't talk to resolved).
323 The four modules should be used in the following order:
325 passwd: compat systemd
326 group: compat systemd
327 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
330 When calling "systemctl enable/disable/is-enabled" on a unit which is a
331 SysV init.d script, it calls /usr/lib/systemd/systemd-sysv-install;
332 this needs to translate the action into the distribution specific
333 mechanism such as chkconfig or update-rc.d. Packagers need to provide
334 this script if you need this functionality (you don't if you disabled
337 Please see src/systemctl/systemd-sysv-install.SKELETON for how this
338 needs to look like, and provide an implementation at the marked places.
340 WARNINGS and TAINT FLAGS:
341 systemd will warn during early boot if /usr is not already mounted at
342 this point (that means: either located on the same file system as / or
343 already mounted in the initrd). While in systemd itself very little
344 will break if /usr is on a separate late-mounted partition, many of its
345 dependencies very likely will break sooner or later in one form or
346 another. For example, udev rules tend to refer to binaries in /usr,
347 binaries that link to libraries in /usr or binaries that refer to data
348 files in /usr. Since these breakages are not always directly visible,
349 systemd will warn about this. Such setups are not really supported by
350 the basic set of Linux OS components. Taint flag 'split-usr' will be
351 set when this condition is detected.
353 For more information on this issue consult
354 https://www.freedesktop.org/wiki/Software/systemd/separate-usr-is-broken
356 systemd requires that the /run mount point exists. systemd also
357 requires that /var/run is a symlink to /run. Taint flag 'var-run-bad'
358 will be set when this condition is detected.
360 Systemd will also warn when the cgroup support is unavailable in the
361 kernel (taint flag 'cgroups-missing'), the system is using the old
362 cgroup hierarchy (taint flag 'cgroupsv1'), the hardware clock is
363 running in non-UTC mode (taint flag 'local-hwclock'), the kernel
364 overflow UID or GID are not 65534 (taint flags 'overflowuid-not-65534'
365 and 'overflowgid-not-65534'), the UID or GID range assigned to the
366 running systemd instance covers less than 0…65534 (taint flags
367 'short-uid-range' and 'short-gid-range').
369 Taint conditions are logged during boot, but may also be checked at any
372 busctl get-property org.freedesktop.systemd1 /org/freedesktop/systemd1 org.freedesktop.systemd1.Manager Tainted
375 To run systemd under valgrind, compile with meson option
376 -Dvalgrind=true and have valgrind development headers installed
377 (i.e. valgrind-devel or equivalent). Otherwise, false positives will be
378 triggered by code which violates some rules but is actually safe. Note
379 that valgrind generates nice output only on exit(), hence on shutdown
380 we don't execve() systemd-shutdown.
382 STABLE BRANCHES AND BACKPORTS:
383 Stable branches with backported patches are available in the
384 systemd-stable repo at https://github.com/systemd/systemd-stable.
386 Stable branches are started for certain releases of systemd and named
387 after them, e.g. v238-stable. Stable branches are managed by
388 distribution maintainers on an as needed basis. See
389 https://www.freedesktop.org/wiki/Software/systemd/Backports/ for some
390 more information and examples.