1CHRONYC(1) User manual CHRONYC(1)
2
6 chronyc - command-line interface for chrony daemon
7
9 chronyc [OPTION]... [COMMAND]...
10
12 chronyc is a command-line interface program which can be used to
13 monitor chronyd’s performance and to change various operating
14 parameters whilst it is running.
15 If no commands are specified on the command line, chronyc will expect
17 input from the user. The prompt chronyc> will be displayed when it is
18 being run from a terminal. If chronyc’s input or output are redirected
19 from or to a file, the prompt is not shown.
20 There are two ways chronyc can access chronyd. One is the Internet
22 Protocol (IPv4 or IPv6) and the other is a Unix domain socket, which is
23 accessible locally by the root or chrony user. By default, chronyc
24 first tries to connect to the Unix domain socket. The compiled-in
25 default path is /var/run/chrony/chronyd.sock. If that fails (e.g.
26 because chronyc is running under a non-root user), it will try to
27 connect to 127.0.0.1 and then ::1.
28 Only the following monitoring commands, which do not affect the
30 behaviour of chronyd, are allowed from the network: activity, manual
31 list, rtcdata, smoothing, sources, sourcestats, tracking, waitsync. The
32 set of hosts from which chronyd will accept these commands can be
33 configured with the cmdallow directive in the chronyd’s configuration
34 file or the cmdallow command in chronyc. By default, the commands are
35 accepted only from localhost (127.0.0.1 or ::1).
36 All other commands are allowed only through the Unix domain socket.
38 When sent over the network, chronyd will respond with a ‘Not
39 authorised’ error, even if it is from localhost. In chrony versions
40 before 2.2 they were allowed from the network if they were
41 authenticated with a password, but that is no longer supported.
42 Having full access to chronyd via chronyc is more or less equivalent to
44 being able to modify the chronyd’s configuration file and restart it.
45
47 -4
48 With this option hostnames will be resolved only to IPv4 addresses.
49 -6
51 With this option hostnames will be resolved only to IPv6 addresses.
52 -n
54 This option disables resolving of IP addresses to hostnames, e.g.
55 to avoid slow DNS lookups. Long addresses will not be truncated to
56 fit into the column.
57 -c
59 This option enables printing of reports in a comma-separated values
60 (CSV) format. IP addresses will not be resolved to hostnames, time
61 will be printed as number of seconds since the epoch and values in
62 seconds will not be converted to other units.
63 -d
65 This option enables printing of debugging messages if chronyc was
66 compiled with debugging support.
67 -m
69 Normally, all arguments on the command line are interpreted as one
70 command. With this option multiple commands can be specified. Each
71 argument will be interpreted as a whole command.
72 -h host
74 This option allows the user to specify which host (or
75 comma-separated list of addresses) running the chronyd program is
76 to be contacted. This allows for remote monitoring, without having
77 to connect over SSH to the other host first.
78 The default is to contact chronyd running on the same host where
80 chronyc is being run.
81 -p port
83 This option allows the user to specify the UDP port number which
84 the target chronyd is using for its monitoring connections. This
85 defaults to 323; there would rarely be a need to change this.
86 -f file
88 This option is ignored and is provided only for compatibility.
89 -a
91 This option is ignored and is provided only for compatibility.
92 -v
94 With this option chronyc displays its version number on the
95 terminal and exits.
96
98 This section describes each of the commands available within the
99 chronyc program.
100 System clock
102 tracking
103 The tracking command displays parameters about the system’s clock
104 performance. An example of the output is shown below.
105 Reference ID : CB00710F (foo.example.net)
107 Stratum : 3
108 Ref time (UTC) : Fri Jan 27 09:49:17 2017
109 System time : 0.000006523 seconds slow of NTP time
110 Last offset : -0.000006747 seconds
111 RMS offset : 0.000035822 seconds
112 Frequency : 3.225 ppm slow
113 Residual freq : -0.000 ppm
114 Skew : 0.129 ppm
115 Root delay : 0.013639022 seconds
116 Root dispersion : 0.001100737 seconds
117 Update interval : 64.2 seconds
118 Leap status : Normal
119 The fields are explained as follows:
121 Reference ID
123 This is the reference ID and name (or IP address) of the server
124 to which the computer is currently synchronised. For IPv4
125 addresses, the reference ID is equal to the address and for
126 IPv6 addresses it is the first 32 bits of the MD5 sum of the
127 address.
128 If the reference ID is 7F7F0101 and there is no name or IP
130 address, it means the computer is not synchronised to any
131 external source and that you have the local mode operating (via
132 the local command in chronyc, or the local directive in the
133 configuration file).
134 The reference ID is printed as a hexadecimal number. Note that
136 in older versions it used to be printed in quad-dotted notation
137 and could be confused with an IPv4 address.
138 Stratum
140 The stratum indicates how many hops away from a computer with
141 an attached reference clock we are. Such a computer is a
142 stratum-1 computer, so the computer in the example is two hops
143 away (i.e. foo.example.net is a stratum-2 and is synchronised
144 from a stratum-1).
145 Ref time
147 This is the time (UTC) at which the last measurement from the
148 reference source was processed.
149 System time
151 In normal operation, chronyd by default never steps the system
152 clock, because any jump in the time can have adverse
153 consequences for certain application programs. Instead, any
154 error in the system clock is corrected by slightly speeding up
155 or slowing down the system clock until the error has been
156 removed, and then returning to the system clock’s normal speed.
157 A consequence of this is that there will be a period when the
158 system clock (as read by other programs) will be different from
159 chronyd’s estimate of the current true time (which it reports
160 to NTP clients when it is operating in server mode). The value
161 reported on this line is the difference due to this effect.
162 Last offset
164 This is the estimated local offset on the last clock update.
165 RMS offset
167 This is a long-term average of the offset value.
168 Frequency
170 The ‘frequency’ is the rate by which the system’s clock would
171 be wrong if chronyd was not correcting it. It is expressed in
172 ppm (parts per million). For example, a value of 1 ppm would
173 mean that when the system’s clock thinks it has advanced 1
174 second, it has actually advanced by 1.000001 seconds relative
175 to true time.
176 Residual freq
178 This shows the ‘residual frequency’ for the currently selected
179 reference source. This reflects any difference between what the
180 measurements from the reference source indicate the frequency
181 should be and the frequency currently being used.
182 The reason this is not always zero is that a smoothing
184 procedure is applied to the frequency. Each time a measurement
185 from the reference source is obtained and a new residual
186 frequency computed, the estimated accuracy of this residual is
187 compared with the estimated accuracy (see ‘skew’ next) of the
188 existing frequency value. A weighted average is computed for
189 the new frequency, with weights depending on these accuracies.
190 If the measurements from the reference source follow a
191 consistent trend, the residual will be driven to zero over
192 time.
193 Skew
195 This is the estimated error bound on the frequency.
196 Root delay
198 This is the total of the network path delays to the stratum-1
199 computer from which the computer is ultimately synchronised.
200 Root dispersion
202 This is the total dispersion accumulated through all the
203 computers back to the stratum-1 computer from which the
204 computer is ultimately synchronised. Dispersion is due to
205 system clock resolution, statistical measurement variations,
206 etc.
207 An absolute bound on the computer’s clock accuracy (assuming
209 the stratum-1 computer is correct) is given by:
210 clock_error <= |system_time_offset| + root_dispersion + (0.5 * root_delay)
212 Update interval
214 This is the interval between the last two clock updates.
215 Leap status
217 This is the leap status, which can be Normal, Insert second,
218 Delete second or Not synchronised.
219 makestep, makestep threshold limit
221 Normally chronyd will cause the system to gradually correct any
222 time offset, by slowing down or speeding up the clock as required.
223 In certain situations, the system clock might be so far adrift that
224 this slewing process would take a very long time to correct the
225 system clock.
226 The makestep command can be used in this situation. There are two
228 forms of the command. The first form has no parameters. It tells
229 chronyd to cancel any remaining correction that was being slewed
230 and jump the system clock by the equivalent amount, making it
231 correct immediately.
232 The second form configures the automatic stepping, similarly to the
234 makestep directive. It has two parameters, stepping threshold (in
235 seconds) and number of future clock updates for which the threshold
236 will be active. This can be used with the burst command to quickly
237 make a new measurement and correct the clock by stepping if needed,
238 without waiting for chronyd to complete the measurement and update
239 the clock.
240 makestep 0.1 1
242 burst 1/2
243 BE WARNED: Certain software will be seriously affected by such
245 jumps in the system time. (That is the reason why chronyd uses
246 slewing normally.)
247 maxupdateskew skew-in-ppm
249 This command has the same effect as the maxupdateskew directive in
250 the configuration file.
251 waitsync [max-tries [max-correction [max-skew [interval]]]]
253 The waitsync command waits for chronyd to synchronise.
254 Up to four optional arguments can be specified. The first is the
256 maximum number of tries before giving up and returning a non-zero
257 error code. When 0 is specified, or there are no arguments, the
258 number of tries will not be limited.
259 The second and third arguments are the maximum allowed remaining
261 correction of the system clock and the maximum allowed skew (in
262 ppm) as reported by the tracking command in the System time and
263 Skew fields. If not specified or zero, the value will not be
264 checked.
265 The fourth argument is the interval specified in seconds in which
267 the check is repeated. The interval is 10 seconds by default.
268 An example is:
270 waitsync 60 0.01
272 which will wait up to about 10 minutes (60 times 10 seconds) for
274 chronyd to synchronise to a source and the remaining correction to
275 be less than 10 milliseconds.
276 Time sources
278 sources [-v]
279 This command displays information about the current time sources
280 that chronyd is accessing.
281 The optional argument -v can be specified, meaning verbose. In this
283 case, extra caption lines are shown as a reminder of the meanings
284 of the columns.
285 210 Number of sources = 3
287 MS Name/IP address Stratum Poll Reach LastRx Last sample
288 ===============================================================================
289 #* GPS0 0 4 377 11 -479ns[ -621ns] +/- 134ns
290 ^? foo.example.net 2 6 377 23 -923us[ -924us] +/- 43ms
291 ^+ bar.example.net 1 6 377 21 -2629us[-2619us] +/- 86ms
292 The columns are as follows:
294 M
296 This indicates the mode of the source. ^ means a server, =
297 means a peer and # indicates a locally connected reference
298 clock.
299 S
301 This column indicates the state of the source.
302 · * indicates the source to which chronyd is currently
304 synchronised.
305 · + indicates acceptable sources which are combined with the
307 selected source.
308 · - indicates acceptable sources which are excluded by the
310 combining algorithm.
311 · ? indicates sources to which connectivity has been lost or
313 whose packets do not pass all tests. It is also shown at
314 start-up, until at least 3 samples have been gathered from
315 it.
316 · x indicates a clock which chronyd thinks is a falseticker
318 (i.e. its time is inconsistent with a majority of other
319 sources).
320 · ~ indicates a source whose time appears to have too much
322 variability.
323 Name/IP address
325 This shows the name or the IP address of the source, or
326 reference ID for reference clocks.
327 Stratum
329 This shows the stratum of the source, as reported in its most
330 recently received sample. Stratum 1 indicates a computer with a
331 locally attached reference clock. A computer that is
332 synchronised to a stratum 1 computer is at stratum 2. A
333 computer that is synchronised to a stratum 2 computer is at
334 stratum 3, and so on.
335 Poll
337 This shows the rate at which the source is being polled, as a
338 base-2 logarithm of the interval in seconds. Thus, a value of 6
339 would indicate that a measurement is being made every 64
340 seconds. chronyd automatically varies the polling rate in
341 response to prevailing conditions.
342 Reach
344 This shows the source’s reachability register printed as an
345 octal number. The register has 8 bits and is updated on every
346 received or missed packet from the source. A value of 377
347 indicates that a valid reply was received for all from the last
348 eight transmissions.
349 LastRx
351 This column shows how long ago the last sample was received
352 from the source. This is normally in seconds. The letters m, h,
353 d or y indicate minutes, hours, days, or years.
354 Last sample
356 This column shows the offset between the local clock and the
357 source at the last measurement. The number in the square
358 brackets shows the actual measured offset. This can be suffixed
359 by ns (indicating nanoseconds), us (indicating microseconds),
360 ms (indicating milliseconds), or s (indicating seconds). The
361 number to the left of the square brackets shows the original
362 measurement, adjusted to allow for any slews applied to the
363 local clock since. The number following the +/- indicator shows
364 the margin of error in the measurement. Positive offsets
365 indicate that the local clock is ahead of the source.
366 sourcestats [-v]
368 The sourcestats command displays information about the drift rate
369 and offset estimation process for each of the sources currently
370 being examined by chronyd.
371 The optional argument -v can be specified, meaning verbose. In this
373 case, extra caption lines are shown as a reminder of the meanings
374 of the columns.
375 An example report is:
377 210 Number of sources = 1
379 Name/IP Address NP NR Span Frequency Freq Skew Offset Std Dev
380 ===============================================================================
381 foo.example.net 11 5 46m -0.001 0.045 1us 25us
382 The columns are as follows:
384 Name/IP Address
386 This is the name or IP address of the NTP server (or peer) or
387 reference ID of the reference clock to which the rest of the
388 line relates.
389 NP
391 This is the number of sample points currently being retained
392 for the server. The drift rate and current offset are estimated
393 by performing a linear regression through these points.
394 NR
396 This is the number of runs of residuals having the same sign
397 following the last regression. If this number starts to become
398 too small relative to the number of samples, it indicates that
399 a straight line is no longer a good fit to the data. If the
400 number of runs is too low, chronyd discards older samples and
401 re-runs the regression until the number of runs becomes
402 acceptable.
403 Span
405 This is the interval between the oldest and newest samples. If
406 no unit is shown the value is in seconds. In the example, the
407 interval is 46 minutes.
408 Frequency
410 This is the estimated residual frequency for the server, in
411 parts per million. In this case, the computer’s clock is
412 estimated to be running 1 part in 10^9 slow relative to the
413 server.
414 Freq Skew
416 This is the estimated error bounds on Freq (again in parts per
417 million).
418 Offset
420 This is the estimated offset of the source.
421 Std Dev
423 This is the estimated sample standard deviation.
424 reselect
426 To avoid excessive switching between sources, chronyd can stay
427 synchronised to a source even when it is not currently the best one
428 among the available sources.
429 The reselect command can be used to force chronyd to reselect the
431 best synchronisation source.
432 reselectdist distance
434 The reselectdist command sets the reselection distance. It is
435 equivalent to the reselectdist directive in the configuration file.
436 NTP sources
438 activity
439 This command reports the number of servers and peers that are
440 online and offline. If the auto_offline option is used in
441 specifying some of the servers or peers, the activity command can
442 be useful for detecting when all of them have entered the offline
443 state after the network link has been disconnected.
444 The report shows the number of servers and peers in 5 states:
446 online
448 the server or peer is currently online (i.e. assumed by chronyd
449 to be reachable)
450 offline
452 the server or peer is currently offline (i.e. assumed by
453 chronyd to be unreachable, and no measurements from it will be
454 attempted.)
455 burst_online
457 a burst command has been initiated for the server or peer and
458 is being performed; after the burst is complete, the server or
459 peer will be returned to the online state.
460 burst_offline
462 a burst command has been initiated for the server or peer and
463 is being performed; after the burst is complete, the server or
464 peer will be returned to the offline state.
465 unresolved
467 the name of the server or peer was not resolved to an address
468 yet; this source is not visible in the sources and sourcestats
469 reports.
470 ntpdata [address]
472 The ntpdata command displays the last valid measurement and other
473 NTP-specific information about the specified NTP source, or all NTP
474 sources if no address was specified. An example of the output is
475 shown below.
476 Remote address : 203.0.113.15 (CB00710F)
478 Remote port : 123
479 Local address : 203.0.113.74 (CB00714A)
480 Leap status : Normal
481 Version : 4
482 Mode : Server
483 Stratum : 1
484 Poll interval : 10 (1024 seconds)
485 Precision : -24 (0.000000060 seconds)
486 Root delay : 0.000015 seconds
487 Root dispersion : 0.000015 seconds
488 Reference ID : 47505300 (GPS)
489 Reference time : Fri Nov 25 15:22:12 2016
490 Offset : -0.000060878 seconds
491 Peer delay : 0.000175634 seconds
492 Peer dispersion : 0.000000681 seconds
493 Response time : 0.000053050 seconds
494 Jitter asymmetry: +0.00
495 NTP tests : 111 111 1111
496 Interleaved : No
497 Authenticated : No
498 TX timestamping : Kernel
499 RX timestamping : Kernel
500 Total TX : 24
501 Total RX : 24
502 Total valid RX : 24
503 The fields are explained as follows:
505 Remote address
507 The IP address of the NTP server or peer, and the corresponding
508 reference ID.
509 Remote port
511 The UDP port number to which the request was sent. The standard
512 NTP port is 123.
513 Local address
515 The local IP address which received the response, and the
516 corresponding reference ID.
517 Leap status, Version, Mode, Stratum, Poll interval, Precision, Root
519 delay, Root dispersion, Reference ID, Reference time
520 The NTP values from the last valid response.
521 Offset, Peer delay, Peer dispersion
523 The measured values.
524 Response time
526 The time the server or peer spent in processing of the request
527 and waiting before sending the response.
528 Jitter asymmetry
530 The estimated asymmetry of network jitter on the path to the
531 source. The asymmetry can be between -0.5 and 0.5. A negative
532 value means the delay of packets sent to the source is more
533 variable than the delay of packets sent from the source back.
534 NTP tests
536 Results of RFC 5905 tests 1 through 3, 5 through 7, and tests
537 for maximum delay, delay ratio, delay dev ratio, and
538 synchronisation loop.
539 Interleaved
541 This shows if the response was in the interleaved mode.
542 Authenticated
544 This shows if the response was authenticated.
545 TX timestamping
547 The source of the local transmit timestamp. Valid values are
548 Daemon, Kernel, and Hardware.
549 RX timestamping
551 The source of the local receive timestamp.
552 Total TX
554 The number of packets sent to the source.
555 Total RX
557 The number of all packets received from the source.
558 Total valid RX
560 The number of valid packets received from the source.
561 add peer address [option]...
563 The add peer command allows a new NTP peer to be added whilst
564 chronyd is running.
565 Following the words add peer, the syntax of the following
567 parameters and options is similar to that for the peer directive in
568 the configuration file. The following peer options can be set in
569 the command: port, minpoll, maxpoll, presend, maxdelayratio,
570 maxdelay, key.
571 An example of using this command is shown below.
573 add peer foo.example.net minpoll 6 maxpoll 10 key 25
575 add server address [option]...
577 The add server command allows a new NTP server to be added whilst
578 chronyd is running.
579 Following the words add server, the syntax of the following
581 parameters and options is similar to that for the server directive
582 in the configuration file. The following server options can be set
583 in the command: port, minpoll, maxpoll, presend, maxdelayratio,
584 maxdelay, key.
585 An example of using this command is shown below:
587 add server foo.example.net minpoll 6 maxpoll 10 key 25
589 delete address
591 The delete command allows an NTP server or peer to be removed from
592 the current set of sources.
593 burst good/max [mask/masked-address], burst good/max
595 [masked-address/masked-bits], burst good/max [address]
596 The burst command tells chronyd to make a set of measurements to
597 each of its NTP sources over a short duration (rather than the
598 usual periodic measurements that it makes). After such a burst,
599 chronyd will revert to the previous state for each source. This
600 might be either online, if the source was being periodically
601 measured in the normal way, or offline, if the source had been
602 indicated as being offline. (A source can be switched between the
603 online and offline states with the online and offline commands.)
604 The mask and masked-address arguments are optional, in which case
606 chronyd will initiate a burst for all of its currently defined
607 sources.
608 The arguments have the following meaning and format:
610 good
612 This defines the number of good measurements that chronyd will
613 want to obtain from each source. A measurement is good if it
614 passes certain tests, for example, the round trip time to the
615 source must be acceptable. (This allows chronyd to reject
616 measurements that are likely to be bogus.)
617 max
619 This defines the maximum number of measurements that chronyd
620 will attempt to make, even if the required number of good
621 measurements has not been obtained.
622 mask
624 This is an IP address with which the IP address of each of
625 chronyd’s sources is to be masked.
626 masked-address
628 This is an IP address. If the masked IP address of a source
629 matches this value then the burst command is applied to that
630 source.
631 masked-bits
633 This can be used with masked-address for CIDR notation, which
634 is a shorter alternative to the form with mask.
635 address
637 This is an IP address or a hostname. The burst command is
638 applied only to that source.
639 If no mask or masked-address arguments are provided, every source
643 will be matched.
644 An example of the two-argument form of the command is:
646 burst 2/10
648 This will cause chronyd to attempt to get two good measurements
650 from each source, stopping after two have been obtained, but in no
651 event will it try more than ten probes to the source.
652 Examples of the four-argument form of the command are:
654 burst 2/10 255.255.0.0/1.2.0.0
656 burst 2/10 2001:db8:789a::/48
657 In the first case, the two out of ten sampling will only be applied
659 to sources whose IPv4 addresses are of the form 1.2.x.y, where x
660 and y are arbitrary. In the second case, the sampling will be
661 applied to sources whose IPv6 addresses have first 48 bits equal to
662 2001:db8:789a.
663 Example of the three-argument form of the command is:
665 burst 2/10 foo.example.net
667 maxdelay address delay
669 This allows the maxdelay option for one of the sources to be
670 modified, in the same way as specifying the maxdelay option for the
671 server directive in the configuration file.
672 maxdelaydevratio address ratio
674 This allows the maxdelaydevratio option for one of the sources to
675 be modified, in the same way as specifying the maxdelaydevratio
676 option for the server directive in the configuration file.
677 maxdelayratio address ratio
679 This allows the maxdelayratio option for one of the sources to be
680 modified, in the same way as specifying the maxdelayratio option
681 for the server directive in the configuration file.
682 maxpoll address maxpoll
684 The maxpoll command is used to modify the maximum polling interval
685 for one of the current set of sources. It is equivalent to the
686 maxpoll option in the server directive in the configuration file.
687 Note that the new maximum polling interval only takes effect after
689 the next measurement has been made.
690 minpoll address minpoll
692 The minpoll command is used to modify the minimum polling interval
693 for one of the current set of sources. It is equivalent to the
694 minpoll option in the server directive in the configuration file.
695 Note that the new minimum polling interval only takes effect after
697 the next measurement has been made.
698 minstratum address minstratum
700 The minstratum command is used to modify the minimum stratum for
701 one of the current set of sources. It is equivalent to the
702 minstratum option in the server directive in the configuration
703 file.
704 offline [address], offline [masked-address/masked-bits], offline
706 [mask/masked-address]
707 The offline command is used to warn chronyd that the network
708 connection to a particular host or hosts is about to be lost, e.g.
709 on computers with intermittent connection to their time sources.
710 Another case where offline could be used is where a computer serves
712 time to a local group of computers, and has a permanent connection
713 to true time servers outside the organisation. However, the
714 external connection is heavily loaded at certain times of the day
715 and the measurements obtained are less reliable at those times. In
716 this case, it is probably most useful to determine the gain or loss
717 rate during the quiet periods and let the whole network coast
718 through the loaded periods. The offline and online commands can be
719 used to achieve this.
720 There are four forms of the offline command. The first form is a
722 wildcard, meaning all sources. The second form allows an IP address
723 mask and a masked address to be specified. The third form uses CIDR
724 notation. The fourth form uses an IP address or a hostname. These
725 forms are illustrated below.
726 offline
728 offline 255.255.255.0/1.2.3.0
729 offline 2001:db8:789a::/48
730 offline foo.example.net
731 The second form means that the offline command is to be applied to
733 any source whose IPv4 address is in the 1.2.3 subnet. (The host’s
734 address is logically and-ed with the mask, and if the result
735 matches the masked-address the host is processed.) The third form
736 means that the command is to be applied to all sources whose IPv6
737 addresses have their first 48 bits equal to 2001:db8:789a. The
738 fourth form means that the command is to be applied only to that
739 one source.
740 The wildcard form of the address is equivalent to:
742 offline 0.0.0.0/0.0.0.0
744 offline ::/0
745 online [address], online [masked-address/masked-bits], online
747 [mask/masked-address]
748 The online command is opposite in function to the offline command.
749 It is used to advise chronyd that network connectivity to a
750 particular source or sources has been restored.
751 The syntax is identical to that of the offline command.
753 polltarget address polltarget
755 The polltarget command is used to modify the poll target for one of
756 the current set of sources. It is equivalent to the polltarget
757 option in the server directive in the configuration file.
758 refresh
760 The refresh command can be used to force chronyd to resolve the
761 names of configured sources to IP addresses again, e.g. after
762 suspending and resuming the machine in a different network.
763 Sources that stop responding will be replaced with newly resolved
765 addresses automatically after 8 polling intervals, but this command
766 can still be useful to replace them immediately and not wait until
767 they are marked as unreachable.
768 Manual time input
770 manual on, manual off, manual delete index, manual list, manual reset
771 The manual command enables and disables use of the settime command,
772 and is used to modify the behaviour of the manual clock driver.
773 The on form of the command enables use of the settime command.
775 The off form of the command disables use of the settime command.
777 The list form of the command lists all the samples currently stored
779 in chronyd. The output is illustrated below.
780 210 n_samples = 1
782 # Date Time(UTC) Slewed Original Residual
783 ====================================================
784 0 27Jan99 22:09:20 0.00 0.97 0.00
785 The columns are as as follows:
787 1. The sample index (used for the manual delete command).
789 2. The date and time of the sample.
791 3. The system clock error when the timestamp was entered, adjusted
793 to allow for changes made to the system clock since.
794 4. The system clock error when the timestamp was entered, as it
796 originally was (without allowing for changes to the system
797 clock since).
798 5. The regression residual at this point, in seconds. This allows
800 ‘outliers’ to be easily spotted, so that they can be deleted
801 using the manual delete command.
802 The delete form of the command deletes a single sample. The
806 parameter is the index of the sample, as shown in the first column
807 of the output from manual list. Following deletion of the data
808 point, the current error and drift rate are re-estimated from the
809 remaining data points and the system clock trimmed if necessary.
810 This option is intended to allow ‘outliers’ to be discarded, i.e.
811 samples where the administrator realises they have entered a very
812 poor timestamp.
813 The reset form of the command deletes all samples at once. The
815 system clock is left running as it was before the command was
816 entered.
817 settime time
819 The settime command allows the current time to be entered manually,
820 if this option has been configured into chronyd. (It can be
821 configured either with the manual directive in the configuration
822 file, or with the manual command of chronyc.)
823 It should be noted that the computer’s sense of time will only be
825 as accurate as the reference you use for providing this input (e.g.
826 your watch), as well as how well you can time the press of the
827 return key.
828 Providing your computer’s time zone is set up properly, you will be
830 able to enter a local time (rather than UTC).
831 The response to a successful settime command indicates the amount
833 that the computer’s clock was wrong. It should be apparent from
834 this if you have entered the time wrongly, e.g. with the wrong time
835 zone.
836 The rate of drift of the system clock is estimated by a regression
838 process using the entered measurement and all previous measurements
839 entered during the present run of chronyd. However, the entered
840 measurement is used for adjusting the current clock offset (rather
841 than the estimated intercept from the regression, which is
842 ignored). Contrast what happens with the manual delete command,
843 where the intercept is used to set the current offset (since there
844 is no measurement that has just been entered in that case).
845 The time is parsed by the public domain getdate algorithm.
847 Consequently, you can only specify time to the nearest second.
848 Examples of inputs that are valid are shown below:
850 settime 16:30
852 settime 16:30:05
853 settime Nov 21, 2015 16:30:05
854 For a full description of getdate, see the getdate documentation
856 (bundled, for example, with the source for GNU tar).
857 NTP access
859 accheck address
860 This command allows you to check whether client NTP access is
861 allowed from a particular host.
862 Examples of use, showing a named host and a numeric IP address, are
864 as follows:
865 accheck foo.example.net
867 accheck 1.2.3.4
868 accheck 2001:db8::1
869 This command can be used to examine the effect of a series of
871 allow, allow all, deny, and deny all commands specified either via
872 chronyc, or in chronyd’s configuration file.
873 clients
875 This command shows a list of clients that have accessed the server,
876 through either the NTP or command ports. It does not include
877 accesses over the Unix domain command socket. There are no
878 arguments.
879 An example of the output is:
881 Hostname NTP Drop Int IntL Last Cmd Drop Int Last
883 ===============================================================================
884 localhost 2 0 2 - 133 15 0 -1 7
885 foo.example.net 12 0 6 - 23 0 0 - -
886 Each row shows the data for a single host. Only hosts that have
888 passed the host access checks (set with the allow, deny, cmdallow
889 and cmddeny commands or configuration file directives) are logged.
890 The intervals are displayed as a power of 2 in seconds.
891 The columns are as follows:
893 1. The hostname of the client.
895 2. The number of NTP packets received from the client.
897 3. The number of NTP packets dropped to limit the response rate.
899 4. The average interval between NTP packets.
901 5. The average interval between NTP packets after limiting the
903 response rate.
904 6. Time since the last NTP packet was received
906 7. The number of command packets received from the client.
908 8. The number of command packets dropped to limit the response
910 rate.
911 9. The average interval between command packets.
913 10. Time since the last command packet was received.
915 serverstats
917 The serverstats command displays how many valid NTP and command
918 requests chronyd as a server received from clients, how many of
919 them were dropped to limit the response rate as configured by the
920 ratelimit and cmdratelimit directives, and how many client log
921 records were dropped due to the memory limit configured by the
922 clientloglimit directive. An example of the output is shown below.
923 NTP packets received : 1598
925 NTP packets dropped : 8
926 Command packets received : 19
927 Command packets dropped : 0
928 Client log records dropped : 0
929 allow [all] [subnet]
931 The effect of the allow command is identical to the allow directive
932 in the configuration file.
933 The syntax is illustrated in the following examples:
935 allow foo.example.net
937 allow all 1.2
938 allow 3.4.5
939 allow 6.7.8/22
940 allow 6.7.8.9/22
941 allow 2001:db8:789a::/48
942 allow 0/0
943 allow ::/0
944 allow
945 allow all
946 deny [all] [subnet]
948 The effect of the allow command is identical to the deny directive
949 in the configuration file.
950 The syntax is illustrated in the following examples:
952 deny foo.example.net
954 deny all 1.2
955 deny 3.4.5
956 deny 6.7.8/22
957 deny 6.7.8.9/22
958 deny 2001:db8:789a::/48
959 deny 0/0
960 deny ::/0
961 deny
962 deny all
963 local [option]..., local off
965 The local command allows chronyd to be told that it is to appear as
966 a reference source, even if it is not itself properly synchronised
967 to an external source. (This can be used on isolated networks, to
968 allow one computer to be a master time server with the other
969 computers slaving to it.)
970 The first form enables the local reference mode on the host. The
972 syntax is identical to the local directive in the configuration
973 file.
974 The second form disables the local reference mode.
976 smoothing
978 The smoothing command displays the current state of the NTP server
979 time smoothing, which can be enabled with the smoothtime directive.
980 An example of the output is shown below.
981 Active : Yes
983 Offset : +1.000268817 seconds
984 Frequency : -0.142859 ppm
985 Wander : -0.010000 ppm per second
986 Last update : 17.8 seconds ago
987 Remaining time : 19988.4 seconds
988 The fields are explained as follows:
990 Active
992 This shows if the server time smoothing is currently active.
993 Possible values are Yes and No. If the leaponly option is
994 included in the smoothtime directive, (leap second only) will
995 be shown on the line.
996 Offset
998 This is the current offset applied to the time sent to NTP
999 clients. Positive value means the clients are getting time
1000 that’s ahead of true time.
1001 Frequency
1003 The current frequency offset of the served time. Negative value
1004 means the time observed by clients is running slower than true
1005 time.
1006 Wander
1008 The current frequency wander of the served time. Negative value
1009 means the time observed by clients is slowing down.
1010 Last update
1012 This field shows how long ago the time smoothing process was
1013 updated, e.g. chronyd accumulated a new measurement.
1014 Remaining time
1016 The time it would take for the smoothing process to get to zero
1017 offset and frequency if there were no more updates.
1018 smoothtime activate, smoothtime reset
1020 The smoothtime command can be used to activate or reset the server
1021 time smoothing process if it is configured with the smoothtime
1022 directive.
1023 Monitoring access
1025 cmdaccheck address
1026 This command is similar to the accheck command, except that it is
1027 used to check whether monitoring access is permitted from a named
1028 host.
1029 Examples of use are as follows:
1031 cmdaccheck foo.example.net
1033 cmdaccheck 1.2.3.4
1034 cmdaccheck 2001:db8::1
1035 cmdallow [all] [subnet]
1037 This is similar to the allow command, except that it is used to
1038 allow particular hosts or subnets to use chronyc to monitor with
1039 chronyd on the current host.
1040 cmddeny [all] [subnet]
1042 This is similar to the deny command, except that it is used to
1043 allow particular hosts or subnets to use chronyc to monitor chronyd
1044 on the current host.
1045 Real-time clock (RTC)
1047 rtcdata
1048 The rtcdata command displays the current RTC parameters.
1049 An example output is shown below.
1051 RTC ref time (GMT) : Sat May 30 07:25:56 2015
1053 Number of samples : 10
1054 Number of runs : 5
1055 Sample span period : 549
1056 RTC is fast by : -1.632736 seconds
1057 RTC gains time at : -107.623 ppm
1058 The fields have the following meaning:
1060 RTC ref time (GMT)
1062 This is the RTC reading the last time its error was measured.
1063 Number of samples
1065 This is the number of previous measurements being used to
1066 determine the RTC gain or loss rate.
1067 Number of runs
1069 This is the number of runs of residuals of the same sign
1070 following the regression fit for (RTC error) versus (RTC time).
1071 A value which is small indicates that the measurements are not
1072 well approximated by a linear model, and that the algorithm
1073 will tend to delete the older measurements to improve the fit.
1074 Sample span period
1076 This is the period that the measurements span (from the oldest
1077 to the newest). Without a unit the value is in seconds;
1078 suffixes m for minutes, h for hours, d for days or y for years
1079 can be used.
1080 RTC is fast by
1082 This is the estimate of how many seconds fast the RTC when it
1083 thought the time was at the reference time (above). If this
1084 value is large, you might (or might not) want to use the
1085 trimrtc command to bring the RTC into line with the system
1086 clock. (Note, a large error will not affect chronyd’s
1087 operation, unless it becomes so big as to start causing
1088 rounding errors.)
1089 RTC gains time at
1091 This is the amount of time gained (positive) or lost (negative)
1092 by the real time clock for each second that it ticks. It is
1093 measured in parts per million. So if the value shown was +1,
1094 suppose the RTC was exactly right when it crosses a particular
1095 second boundary. Then it would be 1 microsecond fast when it
1096 crosses its next second boundary.
1097 trimrtc
1099 The trimrtc command is used to correct the system’s real-time clock
1100 (RTC) to the main system clock. It has no effect if the error
1101 between the two clocks is currently estimated at less than a
1102 second.
1103 The command takes no arguments. It performs the following steps (if
1105 the RTC is more than 1 second away from the system clock):
1106 1. Remember the currently estimated gain or loss rate of the RTC
1108 and flush the previous measurements.
1109 2. Step the real-time clock to bring it within a second of the
1111 system clock.
1112 3. Make several measurements to accurately determine the new
1114 offset between the RTC and the system clock (i.e. the remaining
1115 fraction of a second error).
1116 4. Save the RTC parameters to the RTC file (specified with the
1118 rtcfile directive in the configuration file).
1119 The last step is done as a precaution against the computer
1123 suffering a power failure before either the daemon exits or the
1124 writertc command is issued.
1125 chronyd will still work perfectly well both whilst operating and
1127 across machine reboots even if the trimrtc command is never used
1128 (and the RTC is allowed to drift away from true time). The trimrtc
1129 command is provided as a method by which it can be corrected, in a
1130 manner compatible with chronyd using it to maintain accurate time
1131 across machine reboots.
1132 The trimrtc command can be executed automatically by chronyd with
1134 the rtcautotrim directive in the configuration file.
1135 writertc
1137 The writertc command writes the currently estimated error and gain
1138 or loss rate parameters for the RTC to the RTC file (specified with
1139 the rtcfile directive). This information is also written
1140 automatically when chronyd is killed (by the SIGHUP, SIGINT,
1141 SIGQUIT or SIGTERM signals) or when the trimrtc command is issued.
1142 Other daemon commands
1144 cyclelogs
1145 The cyclelogs command causes all of chronyd’s open log files to be
1146 closed and re-opened. This allows them to be renamed so that they
1147 can be periodically purged. An example of how to do this is shown
1148 below.
1149 # mv /var/log/chrony/measurements.log /var/log/chrony/measurements1.log
1151 # chronyc cyclelogs
1152 # ls -l /var/log/chrony
1153 -rw-r--r-- 1 root root 0 Jun 8 18:17 measurements.log
1154 -rw-r--r-- 1 root root 12345 Jun 8 18:17 measurements1.log
1155 # rm -f measurements1.log
1156 dump
1158 The dump command causes chronyd to write its current history of
1159 measurements for each of its sources to dump files in the directory
1160 specified in the configuration file by the dumpdir directive. Note
1161 that chronyd does this automatically when it exits. This command is
1162 mainly useful for inspection of the history whilst chronyd is
1163 running.
1164 rekey
1166 The rekey command causes chronyd to re-read the key file specified
1167 in the configuration file by the keyfile directive.
1168 Client commands
1170 dns option
1171 The dns command configures how hostnames and IP addresses are
1172 resolved in chronyc. IP addresses can be resolved to hostnames when
1173 printing results of sources, sourcestats, tracking and clients
1174 commands. Hostnames are resolved in commands that take an address
1175 as argument.
1176 There are five options:
1178 dns -n
1180 Disables resolving IP addresses to hostnames. Raw IP addresses
1181 will be displayed.
1182 dns +n
1184 Enables resolving IP addresses to hostnames. This is the
1185 default unless chronyc was started with -n option.
1186 dns -4
1188 Resolves hostnames only to IPv4 addresses.
1189 dns -6
1191 Resolves hostnames only to IPv6 addresses.
1192 dns -46
1194 Resolves hostnames to both address families. This is the
1195 default behaviour unless chronyc was started with the -4 or -6
1196 option.
1197 timeout timeout
1199 The timeout command sets the initial timeout for chronyc requests
1200 in milliseconds. If no response is received from chronyd, the
1201 timeout is doubled and the request is resent. The maximum number of
1202 retries is configured with the retries command.
1203 By default, the timeout is 1000 milliseconds.
1205 retries retries
1207 The retries command sets the maximum number of retries for chronyc
1208 requests before giving up. The response timeout is controlled by
1209 the timeout command.
1210 The default is 2.
1212 keygen [id [type [bits]]]
1214 The keygen command generates a key that can be added to the key
1215 file (specified with the keyfile directive) to allow NTP
1216 authentication between server and client, or peers. The key is
1217 generated from the /dev/urandom device and it is printed to
1218 standard output.
1219 The command has three optional arguments. The first argument is the
1221 key number (by default 1), which will be specified with the key
1222 option of the server or peer directives in the configuration file.
1223 The second argument is the hash function (by default SHA1 or MD5 if
1224 SHA1 is not available) and the third argument is the number of bits
1225 the key should have, between 80 and 4096 bits (by default 160
1226 bits).
1227 An example is:
1229 keygen 73 SHA1 256
1231 which generates a 256-bit SHA1 key with number 73. The printed line
1233 should then be securely transferred and added to the key files on
1234 both server and client, or peers.
1235 exit, quit
1237 The exit and quit commands exit from chronyc and return the user to
1238 the shell.
1239 help
1241 The help command displays a summary of the commands and their
1242 arguments.
1243
1245 chrony.conf(5), chronyd(8)
1246
1248 For instructions on how to report bugs, please visit <https://
1249 chrony.tuxfamily.org/>.
1250
1252 chrony was written by Richard Curnow, Miroslav Lichvar, and others.
1253chrony 3.2 2017-09-15 CHRONYC(1)