1 systemd System and Service Manager
4 http://0pointer.de/blog/projects/systemd.html
7 http://www.freedesktop.org/wiki/Software/systemd
10 git@github.com:systemd/systemd.git
11 https://github.com/systemd/systemd.git
14 https://github.com/systemd/systemd
17 http://lists.freedesktop.org/mailman/listinfo/systemd-devel
20 #systemd on irc.freenode.org
23 https://github.com/systemd/systemd/issues
31 LGPLv2.1+ for all code
32 - except src/basic/MurmurHash2.c which is Public Domain
33 - except src/basic/siphash24.c which is CC0 Public Domain
34 - except src/journal/lookup3.c which is Public Domain
35 - except src/udev/* which is (currently still) GPLv2, GPLv2+
39 Linux kernel >= 4.2 for unified cgroup hierarchy support
41 Kernel Config Options:
43 CONFIG_CGROUPS (it is OK to disable all controllers)
51 CONFIG_FHANDLE (libudev, mount and bind mount handling)
53 udev will fail to work with the legacy sysfs layout:
54 CONFIG_SYSFS_DEPRECATED=n
56 Legacy hotplug slows down the system and confuses udev:
57 CONFIG_UEVENT_HELPER_PATH=""
59 Userspace firmware loading is not supported and should
60 be disabled in the kernel:
61 CONFIG_FW_LOADER_USER_HELPER=n
63 Some udev rules and virtualization detection relies on it:
66 Support for some SCSI devices serial number retrieval, to
67 create additional symlinks in /dev/disk/ and /dev/tape:
70 Required for PrivateNetwork and PrivateDevices in service units:
72 CONFIG_DEVPTS_MULTIPLE_INSTANCES
73 Note that systemd-localed.service and other systemd units use
74 PrivateNetwork and PrivateDevices so this is effectively required.
76 Optional but strongly recommended:
80 CONFIG_{TMPFS,EXT4,XFS,BTRFS_FS,...}_POSIX_ACL
82 CONFIG_SECCOMP_FILTER (required for seccomp support)
83 CONFIG_CHECKPOINT_RESTORE (for the kcmp() syscall)
85 Required for CPUShares= in resource control unit settings
87 CONFIG_FAIR_GROUP_SCHED
89 Required for CPUQuota= in resource control unit settings
96 We recommend to turn off Real-Time group scheduling in the
97 kernel when using systemd. RT group scheduling effectively
98 makes RT scheduling unavailable for most userspace, since it
99 requires explicit assignment of RT budgets to each unit whose
100 processes making use of RT. As there's no sensible way to
101 assign these budgets automatically this cannot really be
102 fixed, and it's best to disable group scheduling hence.
103 CONFIG_RT_GROUP_SCHED=n
105 Note that kernel auditing is broken when used with systemd's
106 container code. When using systemd in conjunction with
107 containers, please make sure to either turn off auditing at
108 runtime using the kernel command line option "audit=0", or
109 turn it off at kernel compile time using:
111 If systemd is compiled with libseccomp support on
112 architectures which do not use socketcall() and where seccomp
113 is supported (this effectively means x86-64 and ARM, but
114 excludes 32-bit x86!), then nspawn will now install a
115 work-around seccomp filter that makes containers boot even
116 with audit being enabled. This works correctly only on kernels
117 3.14 and newer though. TL;DR: turn audit off, still.
121 libmount >= 2.27.1 (from util-linux)
122 (util-linux *must* be built with --enable-libmount-force-mountinfo)
123 libseccomp >= 2.3.1 (optional)
124 libblkid >= 2.24 (from util-linux) (optional)
125 libkmod >= 15 (optional)
126 PAM >= 1.1.2 (optional)
127 libcryptsetup (optional)
130 libselinux (optional)
132 liblz4 >= 119 (optional)
134 libqrencode (optional)
135 libmicrohttpd (optional)
138 elfutils >= 158 (optional)
139 make, gcc, and similar tools
141 During runtime, you need the following additional
144 util-linux >= v2.27.1 required
145 dbus >= 1.4.0 (strictly speaking optional, but recommended)
149 When building from git, the following tools are needed:
160 python-lxml (optional, but required to build the indices)
162 The build system is initialized with ./autogen.sh. A tar ball
164 git archive --format=tar --prefix=systemd-222/ v222 | xz > systemd-222.tar.xz
166 When systemd-hostnamed is used, it is strongly recommended to
167 install nss-myhostname to ensure that, in a world of
168 dynamically changing hostnames, the hostname stays resolvable
169 under all circumstances. In fact, systemd-hostnamed will warn
170 if nss-myhostname is not installed.
172 Additional packages are necessary to run some tests:
173 - busybox (used by test/TEST-13-NSPAWN-SMOKE)
174 - nc (used by test/TEST-12-ISSUE-3171)
176 - python3-evdev (used by hwdb parsing tests)
177 - strace (used by test/test-functions)
180 Default udev rules use the following standard system group
181 names, which need to be resolvable by getgrnam() at any time,
182 even in the very early boot stages, where no other databases
183 and network are available:
185 audio, cdrom, dialout, disk, input, kmem, lp, tape, tty, video
187 During runtime, the journal daemon requires the
188 "systemd-journal" system group to exist. New journal files will
189 be readable by this group (but not writable), which may be used
190 to grant specific users read access. In addition, system
191 groups "wheel" and "adm" will be given read-only access to
192 journal files using systemd-tmpfiles.service.
194 The journal gateway daemon requires the
195 "systemd-journal-gateway" system user and group to
196 exist. During execution this network facing service will drop
197 privileges and assume this uid/gid for security reasons.
199 Similarly, the NTP daemon requires the "systemd-timesync" system
200 user and group to exist.
202 Similarly, the network management daemon requires the
203 "systemd-network" system user and group to exist.
205 Similarly, the name resolution daemon requires the
206 "systemd-resolve" system user and group to exist.
208 Similarly, the coredump support requires the
209 "systemd-coredump" system user and group to exist.
212 systemd ships with four glibc NSS modules:
214 nss-myhostname resolves the local hostname to locally
215 configured IP addresses, as well as "localhost" to
218 nss-resolve enables DNS resolution via the systemd-resolved
219 DNS/LLMNR caching stub resolver "systemd-resolved".
221 nss-mymachines enables resolution of all local containers registered
222 with machined to their respective IP addresses. It also maps UID/GIDs
223 ranges used by containers to useful names.
225 nss-systemd enables resolution of all dynamically allocated service
226 users. (See the DynamicUser= setting in unit files.)
228 To make use of these NSS modules, please add them to the "hosts:",
229 "passwd:" and "group:" lines in /etc/nsswitch.conf. The "resolve"
230 module should replace the glibc "dns" module in this file (and don't
231 worry, it chain-loads the "dns" module if it can't talk to resolved).
233 The four modules should be used in the following order:
235 passwd: compat mymachines systemd
236 group: compat mymachines systemd
237 hosts: files mymachines resolve myhostname
240 When calling "systemctl enable/disable/is-enabled" on a unit which is a
241 SysV init.d script, it calls /usr/lib/systemd/systemd-sysv-install;
242 this needs to translate the action into the distribution specific
243 mechanism such as chkconfig or update-rc.d. Packagers need to provide
244 this script if you need this functionality (you don't if you disabled
247 Please see src/systemctl/systemd-sysv-install.SKELETON for how this
248 needs to look like, and provide an implementation at the marked places.
251 systemd will warn you during boot if /usr is on a different
252 file system than /. While in systemd itself very little will
253 break if /usr is on a separate partition, many of its
254 dependencies very likely will break sooner or later in one
255 form or another. For example, udev rules tend to refer to
256 binaries in /usr, binaries that link to libraries in /usr or
257 binaries that refer to data files in /usr. Since these
258 breakages are not always directly visible, systemd will warn
259 about this, since this kind of file system setup is not really
260 supported anymore by the basic set of Linux OS components.
262 systemd requires that the /run mount point exists. systemd also
263 requires that /var/run is a symlink to /run.
265 For more information on this issue consult
266 http://freedesktop.org/wiki/Software/systemd/separate-usr-is-broken
268 To run systemd under valgrind, compile with VALGRIND defined
269 (e.g. ./configure CPPFLAGS='... -DVALGRIND=1'). Otherwise,
270 false positives will be triggered by code which violates
271 some rules but is actually safe.
273 Currently, systemd-timesyncd defaults to use the Google NTP
274 servers if not specified otherwise at configure time. You
275 really should not ship an OS or device with this default
276 setting. See DISTRO_PORTING for details.
278 ENGINEERING AND CONSULTING SERVICES:
279 Kinvolk (https://kinvolk.io) offers professional engineering
280 and consulting services for systemd. Please contact Chris Kühl
281 <chris@kinvolk.io> for more information.