1SYSTEMD-ANALYZE(1) systemd-analyze SYSTEMD-ANALYZE(1)
2
6 systemd-analyze - Analyze and debug system manager
7
9 systemd-analyze [OPTIONS...] [time]
10 systemd-analyze [OPTIONS...] blame
12 systemd-analyze [OPTIONS...] critical-chain [UNIT...]
14 systemd-analyze [OPTIONS...] dump
16 systemd-analyze [OPTIONS...] plot [>file.svg]
18 systemd-analyze [OPTIONS...] dot [PATTERN...] [>file.dot]
20 systemd-analyze [OPTIONS...] unit-paths
22 systemd-analyze [OPTIONS...] exit-status [STATUS...]
24 systemd-analyze [OPTIONS...] condition CONDITION...
26 systemd-analyze [OPTIONS...] syscall-filter [SET...]
28 systemd-analyze [OPTIONS...] calendar SPEC...
30 systemd-analyze [OPTIONS...] timestamp TIMESTAMP...
32 systemd-analyze [OPTIONS...] timespan SPAN...
34 systemd-analyze [OPTIONS...] cat-config NAME|PATH...
36 systemd-analyze [OPTIONS...] verify [FILE...]
38 systemd-analyze [OPTIONS...] security UNIT...
40
42 systemd-analyze may be used to determine system boot-up performance
43 statistics and retrieve other state and tracing information from the
44 system and service manager, and to verify the correctness of unit
45 files. It is also used to access special functions useful for advanced
46 system manager debugging.
47 If no command is passed, systemd-analyze time is implied.
49 systemd-analyze time
51 This command prints the time spent in the kernel before userspace has
52 been reached, the time spent in the initial RAM disk (initrd) before
53 normal system userspace has been reached, and the time normal system
54 userspace took to initialize. Note that these measurements simply
55 measure the time passed up to the point where all system services have
56 been spawned, but not necessarily until they fully finished
57 initialization or the disk is idle.
58 Example 1. Show how long the boot took
60 # in a container
62 $ systemd-analyze time
63 Startup finished in 296ms (userspace)
64 multi-user.target reached after 275ms in userspace
65 # on a real machine
67 $ systemd-analyze time
68 Startup finished in 2.584s (kernel) + 19.176s (initrd) + 47.847s (userspace) = 1min 9.608s
69 multi-user.target reached after 47.820s in userspace
70 systemd-analyze blame
72 This command prints a list of all running units, ordered by the time
73 they took to initialize. This information may be used to optimize
74 boot-up times. Note that the output might be misleading as the
75 initialization of one service might be slow simply because it waits for
76 the initialization of another service to complete. Also note:
77 systemd-analyze blame doesn't display results for services with
78 Type=simple, because systemd considers such services to be started
79 immediately, hence no measurement of the initialization delays can be
80 done. Also note that this command only shows the time units took for
81 starting up, it does not show how long unit jobs spent in the execution
82 queue. In particular it shows the time units spent in "activating"
83 state, which is not defined for units such as device units that
84 transition directly from "inactive" to "active". This command hence
85 gives an impression of the performance of program code, but cannot
86 accurately reflect latency introduced by waiting for hardware and
87 similar events.
88 Example 2. Show which units took the most time during boot
90 $ systemd-analyze blame
92 32.875s pmlogger.service
93 20.905s systemd-networkd-wait-online.service
94 13.299s dev-vda1.device
95 ...
96 23ms sysroot.mount
97 11ms initrd-udevadm-cleanup-db.service
98 3ms sys-kernel-config.mount
99 systemd-analyze critical-chain [UNIT...]
102 This command prints a tree of the time-critical chain of units (for
103 each of the specified UNITs or for the default target otherwise). The
104 time after the unit is active or started is printed after the "@"
105 character. The time the unit takes to start is printed after the "+"
106 character. Note that the output might be misleading as the
107 initialization of services might depend on socket activation and
108 because of the parallel execution of units. Also, similar to the blame
109 command, this only takes into account the time units spent in
110 "activating" state, and hence does not cover units that never went
111 through an "activating" state (such as device units that transition
112 directly from "inactive" to "active"). Moreover it does not show
113 information on jobs (and in particular not jobs that timed out).
114 Example 3. systemd-analyze critical-chain
116 $ systemd-analyze critical-chain
118 multi-user.target @47.820s
119 └─pmie.service @35.968s +548ms
120 └─pmcd.service @33.715s +2.247s
121 └─network-online.target @33.712s
122 └─systemd-networkd-wait-online.service @12.804s +20.905s
123 └─systemd-networkd.service @11.109s +1.690s
124 └─systemd-udevd.service @9.201s +1.904s
125 └─systemd-tmpfiles-setup-dev.service @7.306s +1.776s
126 └─kmod-static-nodes.service @6.976s +177ms
127 └─systemd-journald.socket
128 └─system.slice
129 └─-.slice
130 systemd-analyze dump
132 This command outputs a (usually very long) human-readable serialization
133 of the complete server state. Its format is subject to change without
134 notice and should not be parsed by applications.
135 Example 4. Show the internal state of user manager
137 $ systemd-analyze --user dump
139 Timestamp userspace: Thu 2019-03-14 23:28:07 CET
140 Timestamp finish: Thu 2019-03-14 23:28:07 CET
141 Timestamp generators-start: Thu 2019-03-14 23:28:07 CET
142 Timestamp generators-finish: Thu 2019-03-14 23:28:07 CET
143 Timestamp units-load-start: Thu 2019-03-14 23:28:07 CET
144 Timestamp units-load-finish: Thu 2019-03-14 23:28:07 CET
145 -> Unit proc-timer_list.mount:
146 Description: /proc/timer_list
147 ...
148 -> Unit default.target:
149 Description: Main user target
150 ...
151 systemd-analyze plot
153 This command prints an SVG graphic detailing which system services have
154 been started at what time, highlighting the time they spent on
155 initialization.
156 Example 5. Plot a bootchart
158 $ systemd-analyze plot >bootup.svg
160 $ eog bootup.svg&
161 systemd-analyze dot [pattern...]
163 This command generates textual dependency graph description in dot
164 format for further processing with the GraphViz dot(1) tool. Use a
165 command line like systemd-analyze dot | dot -Tsvg >systemd.svg to
166 generate a graphical dependency tree. Unless --order or --require is
167 passed, the generated graph will show both ordering and requirement
168 dependencies. Optional pattern globbing style specifications (e.g.
169 *.target) may be given at the end. A unit dependency is included in the
170 graph if any of these patterns match either the origin or destination
171 node.
172 Example 6. Plot all dependencies of any unit whose name starts with
174 "avahi-daemon"
175 $ systemd-analyze dot 'avahi-daemon.*' | dot -Tsvg >avahi.svg
177 $ eog avahi.svg
178 Example 7. Plot the dependencies between all known target units
180 $ systemd-analyze dot --to-pattern='*.target' --from-pattern='*.target' \
182 | dot -Tsvg >targets.svg
183 $ eog targets.svg
184 systemd-analyze unit-paths
186 This command outputs a list of all directories from which unit files,
187 .d overrides, and .wants, .requires symlinks may be loaded. Combine
188 with --user to retrieve the list for the user manager instance, and
189 --global for the global configuration of user manager instances.
190 Example 8. Show all paths for generated units
192 $ systemd-analyze unit-paths | grep '^/run'
194 /run/systemd/system.control
195 /run/systemd/transient
196 /run/systemd/generator.early
197 /run/systemd/system
198 /run/systemd/system.attached
199 /run/systemd/generator
200 /run/systemd/generator.late
201 Note that this verb prints the list that is compiled into
203 systemd-analyze itself, and does not communicate with the running
204 manager. Use
205 systemctl [--user] [--global] show -p UnitPath --value
207 to retrieve the actual list that the manager uses, with any empty
209 directories omitted.
210 systemd-analyze exit-status [STATUS...]
212 This command prints a list of exit statuses along with their "class",
213 i.e. the source of the definition (one of "glibc", "systemd", "LSB", or
214 "BSD"), see the Process Exit Codes section in systemd.exec(5). If no
215 additional arguments are specified, all known statuses are are shown.
216 Otherwise, only the definitions for the specified codes are shown.
217 Example 9. Show some example exit status names
219 $ systemd-analyze exit-status 0 1 {63..65}
221 NAME STATUS CLASS
222 SUCCESS 0 glibc
223 FAILURE 1 glibc
224 - 63 -
225 USAGE 64 BSD
226 DATAERR 65 BSD
227 systemd-analyze condition CONDITION...
229 This command will evaluate Condition*=... and Assert*=...
230 assignments, and print their values, and the resulting value of the
231 combined condition set. See systemd.unit(5) for a list of available
232 conditions and asserts.
233 Example 10. Evaluate conditions that check kernel versions
235 $ systemd-analyze condition 'ConditionKernelVersion = ! <4.0' \
237 'ConditionKernelVersion = >=5.1' \
238 'ConditionACPower=|false' \
239 'ConditionArchitecture=|!arm' \
240 'AssertPathExists=/etc/os-release'
241 test.service: AssertPathExists=/etc/os-release succeeded.
242 Asserts succeeded.
243 test.service: ConditionArchitecture=|!arm succeeded.
244 test.service: ConditionACPower=|false failed.
245 test.service: ConditionKernelVersion=>=5.1 succeeded.
246 test.service: ConditionKernelVersion=!<4.0 succeeded.
247 Conditions succeeded.
248 systemd-analyze syscall-filter [SET...]
250 This command will list system calls contained in the specified system
251 call set SET, or all known sets if no sets are specified. Argument SET
252 must include the "@" prefix.
253 systemd-analyze calendar EXPRESSION...
255 This command will parse and normalize repetitive calendar time events,
256 and will calculate when they elapse next. This takes the same input as
257 the OnCalendar= setting in systemd.timer(5), following the syntax
258 described in systemd.time(7). By default, only the next time the
259 calendar expression will elapse is shown; use --iterations= to show the
260 specified number of next times the expression elapses. Each time the
261 expression elapses forms a timestamp, see the timestamp verb below.
262 Example 11. Show leap days in the near future
264 $ systemd-analyze calendar --iterations=5 '*-2-29 0:0:0'
266 Original form: *-2-29 0:0:0
267 Normalized form: *-02-29 00:00:00
268 Next elapse: Sat 2020-02-29 00:00:00 UTC
269 From now: 11 months 15 days left
270 Iter. #2: Thu 2024-02-29 00:00:00 UTC
271 From now: 4 years 11 months left
272 Iter. #3: Tue 2028-02-29 00:00:00 UTC
273 From now: 8 years 11 months left
274 Iter. #4: Sun 2032-02-29 00:00:00 UTC
275 From now: 12 years 11 months left
276 Iter. #5: Fri 2036-02-29 00:00:00 UTC
277 From now: 16 years 11 months left
278 systemd-analyze timestamp TIMESTAMP...
280 This command parses a timestamp (i.e. a single point in time) and
281 outputs the normalized form and the difference between this timestamp
282 and now. The timestamp should adhere to the syntax documented in
283 systemd.time(7), section "PARSING TIMESTAMPS".
284 Example 12. Show parsing of timestamps
286 $ systemd-analyze timestamp yesterday now tomorrow
288 Original form: yesterday
289 Normalized form: Mon 2019-05-20 00:00:00 CEST
290 (in UTC): Sun 2019-05-19 22:00:00 UTC
291 UNIX seconds: @15583032000
292 From now: 1 day 9h ago
293 Original form: now
295 Normalized form: Tue 2019-05-21 09:48:39 CEST
296 (in UTC): Tue 2019-05-21 07:48:39 UTC
297 UNIX seconds: @1558424919.659757
298 From now: 43us ago
299 Original form: tomorrow
301 Normalized form: Wed 2019-05-22 00:00:00 CEST
302 (in UTC): Tue 2019-05-21 22:00:00 UTC
303 UNIX seconds: @15584760000
304 From now: 14h left
305 systemd-analyze timespan EXPRESSION...
307 This command parses a time span (i.e. a difference between two
308 timestamps) and outputs the normalized form and the equivalent value in
309 microseconds. The time span should adhere to the syntax documented in
310 systemd.time(7), section "PARSING TIME SPANS". Values without units are
311 parsed as seconds.
312 Example 13. Show parsing of timespans
314 $ systemd-analyze timespan 1s 300s '1year 0.000001s'
316 Original: 1s
317 μs: 1000000
318 Human: 1s
319 Original: 300s
321 μs: 300000000
322 Human: 5min
323 Original: 1year 0.000001s
325 μs: 31557600000001
326 Human: 1y 1us
327 systemd-analyze cat-config NAME|PATH...
329 This command is similar to systemctl cat, but operates on config files.
330 It will copy the contents of a config file and any drop-ins to standard
331 output, using the usual systemd set of directories and rules for
332 precedence. Each argument must be either an absolute path including the
333 prefix (such as /etc/systemd/logind.conf or
334 /usr/lib/systemd/logind.conf), or a name relative to the prefix (such
335 as systemd/logind.conf).
336 Example 14. Showing logind configuration
338 $ systemd-analyze cat-config systemd/logind.conf
340 # /etc/systemd/logind.conf
341 ...
342 [Login]
343 NAutoVTs=8
344 ...
345 # /usr/lib/systemd/logind.conf.d/20-test.conf
347 ... some override from another package
348 # /etc/systemd/logind.conf.d/50-override.conf
350 ... some administrator override
351 systemd-analyze verify FILE...
354 This command will load unit files and print warnings if any errors are
355 detected. Files specified on the command line will be loaded, but also
356 any other units referenced by them. The full unit search path is formed
357 by combining the directories for all command line arguments, and the
358 usual unit load paths. The variable $SYSTEMD_UNIT_PATH is supported,
359 and may be used to replace or augment the compiled in set of unit load
360 paths; see systemd.unit(5). All units files present in the directories
361 containing the command line arguments will be used in preference to the
362 other paths.
363 The following errors are currently detected:
365 · unknown sections and directives,
367 · missing dependencies which are required to start the given unit,
369 · man pages listed in Documentation= which are not found in the
371 system,
372 · commands listed in ExecStart= and similar which are not found in
374 the system or not executable.
375 Example 15. Misspelt directives
377 $ cat ./user.slice
379 [Unit]
380 WhatIsThis=11
381 Documentation=man:nosuchfile(1)
382 Requires=different.service
383 [Service]
385 Description=x
386 $ systemd-analyze verify ./user.slice
388 [./user.slice:9] Unknown lvalue 'WhatIsThis' in section 'Unit'
389 [./user.slice:13] Unknown section 'Service'. Ignoring.
390 Error: org.freedesktop.systemd1.LoadFailed:
391 Unit different.service failed to load:
392 No such file or directory.
393 Failed to create user.slice/start: Invalid argument
394 user.slice: man nosuchfile(1) command failed with code 16
395 Example 16. Missing service units
398 $ tail ./a.socket ./b.socket
400 ==> ./a.socket <==
401 [Socket]
402 ListenStream=100
403 ==> ./b.socket <==
405 [Socket]
406 ListenStream=100
407 Accept=yes
408 $ systemd-analyze verify ./a.socket ./b.socket
410 Service a.service not loaded, a.socket cannot be started.
411 Service b@0.service not loaded, b.socket cannot be started.
412 systemd-analyze security [UNIT...]
415 This command analyzes the security and sandboxing settings of one or
416 more specified service units. If at least one unit name is specified
417 the security settings of the specified service units are inspected and
418 a detailed analysis is shown. If no unit name is specified, all
419 currently loaded, long-running service units are inspected and a terse
420 table with results shown. The command checks for various
421 security-related service settings, assigning each a numeric "exposure
422 level" value, depending on how important a setting is. It then
423 calculates an overall exposure level for the whole unit, which is an
424 estimation in the range 0.0...10.0 indicating how exposed a service is
425 security-wise. High exposure levels indicate very little applied
426 sandboxing. Low exposure levels indicate tight sandboxing and strongest
427 security restrictions. Note that this only analyzes the per-service
428 security features systemd itself implements. This means that any
429 additional security mechanisms applied by the service code itself are
430 not accounted for. The exposure level determined this way should not be
431 misunderstood: a high exposure level neither means that there is no
432 effective sandboxing applied by the service code itself, nor that the
433 service is actually vulnerable to remote or local attacks. High
434 exposure levels do indicate however that most likely the service might
435 benefit from additional settings applied to them.
436 Please note that many of the security and sandboxing settings
438 individually can be circumvented — unless combined with others. For
439 example, if a service retains the privilege to establish or undo mount
440 points many of the sandboxing options can be undone by the service code
441 itself. Due to that is essential that each service uses the most
442 comprehensive and strict sandboxing and security settings possible. The
443 tool will take into account some of these combinations and
444 relationships between the settings, but not all. Also note that the
445 security and sandboxing settings analyzed here only apply to the
446 operations executed by the service code itself. If a service has access
447 to an IPC system (such as D-Bus) it might request operations from other
448 services that are not subject to the same restrictions. Any
449 comprehensive security and sandboxing analysis is hence incomplete if
450 the IPC access policy is not validated too.
451 Example 17. Analyze systemd-logind.service
453 $ systemd-analyze security --no-pager systemd-logind.service
455 NAME DESCRIPTION EXPOSURE
456 ✗ PrivateNetwork= Service has access to the host's network 0.5
457 ✗ User=/DynamicUser= Service runs as root user 0.4
458 ✗ DeviceAllow= Service has no device ACL 0.2
459 ✓ IPAddressDeny= Service blocks all IP address ranges
460 ...
461 → Overall exposure level for systemd-logind.service: 4.1 OK 🙂
462
464 The following options are understood:
465 --system
467 Operates on the system systemd instance. This is the implied
468 default.
469 --user
471 Operates on the user systemd instance.
472 --global
474 Operates on the system-wide configuration for user systemd
475 instance.
476 --order, --require
478 When used in conjunction with the dot command (see above), selects
479 which dependencies are shown in the dependency graph. If --order is
480 passed, only dependencies of type After= or Before= are shown. If
481 --require is passed, only dependencies of type Requires=,
482 Requisite=, Wants= and Conflicts= are shown. If neither is passed,
483 this shows dependencies of all these types.
484 --from-pattern=, --to-pattern=
486 When used in conjunction with the dot command (see above), this
487 selects which relationships are shown in the dependency graph. Both
488 options require a glob(7) pattern as an argument, which will be
489 matched against the left-hand and the right-hand, respectively,
490 nodes of a relationship.
491 Each of these can be used more than once, in which case the unit
493 name must match one of the values. When tests for both sides of the
494 relation are present, a relation must pass both tests to be shown.
495 When patterns are also specified as positional arguments, they must
496 match at least one side of the relation. In other words, patterns
497 specified with those two options will trim the list of edges
498 matched by the positional arguments, if any are given, and fully
499 determine the list of edges shown otherwise.
500 --fuzz=timespan
502 When used in conjunction with the critical-chain command (see
503 above), also show units, which finished timespan earlier, than the
504 latest unit in the same level. The unit of timespan is seconds
505 unless specified with a different unit, e.g. "50ms".
506 --man=no
508 Do not invoke man(1) to verify the existence of man pages listed in
509 Documentation=.
510 --generators
512 Invoke unit generators, see systemd.generator(7). Some generators
513 require root privileges. Under a normal user, running with
514 generators enabled will generally result in some warnings.
515 --root=PATH
517 With cat-files, show config files underneath the specified root
518 path PATH.
519 --iterations=NUMBER
521 When used with the calendar command, show the specified number of
522 iterations the specified calendar expression will elapse next.
523 Defaults to 1.
524 --base-time=TIMESTAMP
526 When used with the calendar command, show next iterations relative
527 to the specified point in time. If not specified defaults to the
528 current time.
529 -H, --host=
531 Execute the operation remotely. Specify a hostname, or a username
532 and hostname separated by "@", to connect to. The hostname may
533 optionally be suffixed by a port ssh is listening on, separated by
534 ":", and then a container name, separated by "/", which connects
535 directly to a specific container on the specified host. This will
536 use SSH to talk to the remote machine manager instance. Container
537 names may be enumerated with machinectl -H HOST. Put IPv6 addresses
538 in brackets.
539 -M, --machine=
541 Execute operation on a local container. Specify a container name to
542 connect to.
543 -h, --help
545 Print a short help text and exit.
546 --version
548 Print a short version string and exit.
549 --no-pager
551 Do not pipe output into a pager.
552
554 On success, 0 is returned, a non-zero failure code otherwise.
555
557 $SYSTEMD_PAGER
558 Pager to use when --no-pager is not given; overrides $PAGER. If
559 neither $SYSTEMD_PAGER nor $PAGER are set, a set of well-known
560 pager implementations are tried in turn, including less(1) and
561 more(1), until one is found. If no pager implementation is
562 discovered no pager is invoked. Setting this environment variable
563 to an empty string or the value "cat" is equivalent to passing
564 --no-pager.
565 $SYSTEMD_LESS
567 Override the options passed to less (by default "FRSXMK").
568 Users might want to change two options in particular:
570 K
572 This option instructs the pager to exit immediately when Ctrl+C
573 is pressed. To allow less to handle Ctrl+C itself to switch
574 back to the pager command prompt, unset this option.
575 If the value of $SYSTEMD_LESS does not include "K", and the
577 pager that is invoked is less, Ctrl+C will be ignored by the
578 executable, and needs to be handled by the pager.
579 X
581 This option instructs the pager to not send termcap
582 initialization and deinitialization strings to the terminal. It
583 is set by default to allow command output to remain visible in
584 the terminal even after the pager exits. Nevertheless, this
585 prevents some pager functionality from working, in particular
586 paged output cannot be scrolled with the mouse.
587 See less(1) for more discussion.
589 $SYSTEMD_LESSCHARSET
591 Override the charset passed to less (by default "utf-8", if the
592 invoking terminal is determined to be UTF-8 compatible).
593 $SYSTEMD_PAGERSECURE
595 Takes a boolean argument. When true, the "secure" mode of the pager
596 is enabled; if false, disabled. If $SYSTEMD_PAGERSECURE is not set
597 at all, secure mode is enabled if the effective UID is not the same
598 as the owner of the login session, see geteuid(2) and
599 sd_pid_get_owner_uid(3). In secure mode, LESSSECURE=1 will be set
600 when invoking the pager, and the pager shall disable commands that
601 open or create new files or start new subprocesses. When
602 $SYSTEMD_PAGERSECURE is not set at all, pagers which are not known
603 to implement secure mode will not be used. (Currently only less(1)
604 implements secure mode.)
605 Note: when commands are invoked with elevated privileges, for
607 example under sudo(8) or pkexec(1), care must be taken to ensure
608 that unintended interactive features are not enabled. "Secure" mode
609 for the pager may be enabled automatically as describe above.
610 Setting SYSTEMD_PAGERSECURE=0 or not removing it from the inherited
611 environment allows the user to invoke arbitrary commands. Note that
612 if the $SYSTEMD_PAGER or $PAGER variables are to be honoured,
613 $SYSTEMD_PAGERSECURE must be set too. It might be reasonable to
614 completly disable the pager using --no-pager instead.
615 $SYSTEMD_COLORS
617 The value must be a boolean. Controls whether colorized output
618 should be generated. This can be specified to override the decision
619 that systemd makes based on $TERM and what the console is connected
620 to.
621 $SYSTEMD_URLIFY
623 The value must be a boolean. Controls whether clickable links
624 should be generated in the output for terminal emulators supporting
625 this. This can be specified to override the decision that systemd
626 makes based on $TERM and other conditions.
627
629 systemd(1), systemctl(1)
630systemd 246 SYSTEMD-ANALYZE(1)