The chrony suite

This manual describes how to use

the programs chronyd and chronyc

Richard P. Curnow


Introduction

Overview

Chrony is a software package for maintaining the accuracy of computer system clocks. It consists of a pair of programs :

Acknowledgements

The chrony suite makes use of the algorithm known as RSA Data Security, Inc. MD5 Message-Digest Algorithm for authenticating messages between different machines on the network.

In writing the chronyd program, extensive use has been made of RFC1305, written by David Mills. I have occasionally referred to the xntp suite's source code to check details of the protocol that the RFC did not make absolutely clear. The core algorithms in chronyd are all completely distinct from xntp, however.

Availability

Getting the software

Links on the chrony home page describe how to obtain the software.

Platforms

Although most of the program is portable between Unix-like systems, there are parts that have to be tailored to each specific vendor's system. These are the parts that interface with the operating system's facilities for adjusting the system clock; different operating systems may provide different function calls to achieve this, and even where the same function is used it may have different quirks in its behaviour.

So far, the software is able to run in the following environments:

Closely related systems may work too, but they have not been tested.

Porting the software to other system (particularly to those supporting an adjtime system call) should not be difficult, however it requires access to such systems to test out the driver.

Relationship to other software packages

xntpd

The `reference' implementation of the Network Time Protocol is the program xntpd, available via The NTP home page.

xntpd is designed to support all the operating modes defined by RFC1305, and has driver support for a large number of reference clocks (such as GPS receivers) that can be connected directly to a computer, thereby providing a so-called 'stratum 1' server.

Things chronyd can do that xntpd can't:

Things xntpd can do that chronyd can't:

timed

timed is a program that is part of the BSD networking suite. It uses broadcast packets to find all machines running the daemon within a subnet. The machines elect a master which periodically measures the system clock offsets of the other computers using ICMP timestamps. Corrections are sent to each member as a result of this process.

Problems that may arise with timed are :

timed does have the benefit over chronyd that for isolated networks of computers, they will track the `majority vote' time. For such isolated networks, chronyd requires one computer to be the `master' with the others slaved to it. If the master has a particular defective clock, the whole set of computers will tend to slip relative to real time (but they will stay accurate relative to one another).

Distribution rights and (lack of) warranty

{@tensf

This is the licence for the programs "chronyd" and "chronyc". Their source code files which reference this licence, object files / programs (binary form) built from such source code, and the supporting documentation are collectively referred to in this licence as "the Software".

Certain source code files required for the construction of the Software may be included under different licensing arrangements; you should refer to such source code files directly for further information. Any such files are outside the scope of this licence.

Person(s) and/or organisation(s) holding the copyright to any part of the source code to which this licence applies are referred to in this licence as "the Copyright Holder(s)".

Copying and use of this licence itself, with or without modification, are permitted for any purpose subject only to condition 6 below.

Copying, use and redistribution of the Software in source and/or binary forms, with or without modification, and use of any part of the Software in the creation of another work, are permitted without fee subject to the following conditions 1 through 8 and the disclaimer below :

  1. Redistributions of the Software as source code, in whole or in part, with or without modification, must retain all existing copyright notices and references to this licence, and must be accompanied by a copy of this licence. You may rename the file containing this licence and change all references to it accordingly if you need to. Such redistributions must be under terms that require this list of conditions and the disclaimer below to be carried on any subsequent redistribution.
  2. Redistributions in binary form (of the original version of the Software, a modified version of the Software, or any other work which incorporates any part of the Software), must carry in accompanying documentation and/or other materials a notice to the effect of either (A) or (B) below.
    1. This original version of this software was written by <name(s) of Copyright Holder(s)> and is used subject to the disclaimer in the file <name of file containing the disclaimer>.
    2. The parts of this software which <describe their role in your software> were originally written by <name(s) of Copyright Holder(s)> and are used subject to the disclaimer in the file <name of file containing the disclaimer>.
    (You must replace each <...> with the appropriate information.) A copy of the disclaimer below must also be included. Such redistributions must be under terms that require the relevant notice (A) or (B) and the disclaimer below to be carried on any subsequent redistribution. As an exception, if all recipients of your distribution receive additional rights under condition 5, and those rights supersede the disclaimer in its entirety, then you need not include the disclaimer, any reference to it, or the requirement for further redistributions to carry it.
  3. Unless modifications to the original Software are retained for the exclusive personal use of the modifier, each source code file that has been modified must identify the original source version and carry a notice describing the nature and purpose of all modifications and the person responsible for them. Where modified source code is widely redistributed (e.g. outside the modifier's own organisation), it should preferably be supplied as the original version plus `patches' which can be applied to the original version to obtain the modified version.
  4. If a modified version of the Software is distributed in any form, it must not be misrepresented (whether by the method of distribution or by any related advertisement or documentation) as being the original version. Both the documentation accompanying the software and any advertisement related to the distribution must expressly state that the distributed software is not the original work, and must identify both the original work (by name and version) and the nature and purpose of all modifications made.
  5. You may offer to any party (and charge a fee for) additional warranty, liability rights and/or support for the Software (or a work derived from it). Any such offer made must be entirely on your own behalf, and you must indemnify the Copyright Holder(s) and all other parties through whose redistributions you have received the Software for any liability arising out of such an offer.
  6. This list of conditions and the disclaimer below, when applied either to the Software as a whole or to any part of the Software, may only be modified by or with the express permission of the Copyright Holder(s).
  7. The Copyright Holder(s) retain all rights to the Software, including without limitation the rights to use the Software for any purpose, to create modified and/or enhanced versions of the Software, and to distibute such versions under terms different to those in this licence.
  8. Except as required by this licence, the name(s) of the Copyright Holder(s) may not be used to endorse or promote products derived from the Software without their express permission obtained in advance.

== Start of disclaimer ==

BECAUSE THE SOFTWARE IS PROVIDED FREE OF CHARGE, THERE IS NO WARRANTY FOR IT, TO THE MAXIMUM EXTENT THAT APPLICABLE LAW ALLOWS. UNLESS OTHERWISE STATED IN WRITING OR REQUIRED BY APPLICABLE LAW, THE COPYRIGHT HOLDER(S) AND/OR OTHER PARTIES PROVIDE THE SOFTWARE `AS IS' AND DISCLAIM ALL WARRANTIES, EITHER EXPRESSED OR IMPLIED, INCLUDING WITHOUT LIMITATION WARRANTIES THAT THE SOFTWARE IS FREE OF DEFECTS, MERCHANTABLE OR FIT FOR A PARTICULAR PURPOSE. THE USER BEARS THE ENTIRE RISK AS TO THE QUALITY, ACCURACY AND PERFORMANCE OF THE SOFTWARE AND THE COST OF ANY NECESSARY SERVICING, REPAIR OR CORRECTION IN THE EVENT THAT THE SOFTWARE SHOULD PROVE DEFECTIVE.

IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE SOFTWARE, BE LIABLE TO YOU OR TO ANY OTHER PARTY FOR DAMAGES OF ANY CHARACTER, INCLUDING WITHOUT LIMITATION ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES WHICH ARISE FROM THE USE OF OR INABILITY TO USE THE SOFTWARE (INCLUDING WITHOUT LIMITATION LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE SOFTWARE TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF SUCH COPYRIGHT HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

== End of disclaimer ==

}

Installation

The software is distributed as source code which has to be compiled unless you have installed the Debian package, in which case it is already compiled and installed with a default configuration. The source code is supplied in the form of a gzipped tar file, which unpacks to a subdirectory identifying the name and version of the program.

After unpacking the source code, change directory into it, and type

./configure

This is a shell script that automatically determines the system type. There is a single optional parameter, --prefix which indicates the directory tree where the software should be installed. For example,

./configure --prefix=/opt/free

will install the chronyd daemon into /opt/free/sbin and the chronyc control program into /opt/free/bin. The default value for the prefix is /usr/local.

If the software cannot (yet) be built on your system, an error message will be shown. Otherwise, the files `options.h' and `Makefile' will be generated.

Now type

make

to build the programs.

Once the programs have been successfully compiled, they need to be installed in their target locations. This step normally needs to be performed by the superuser, and requires the following command to be entered.

make install

Now that the software is successfully installed, the next step is to set up a configuration file. The contents of this depend on the network environment in which the computer operates. The Debian package installs a simple configuration file suitable for a dial-up pc. You should edit it to suit your situation. Typical scenarios are described in the following section of the document.

Typical operating scenarios

Computers connected to the internet

In this section we discuss how to configure chrony for computers that have permanent connections to the internet (or to any network containing true NTP servers which ultimately derive their time from a reference clock).

To operate in this mode, you will need to know the names of the NTP server machines you wish to use. You may be able to find names of suitable servers by one of the following methods:

Assuming that you have found some servers, you need to set up a configuration file to run chrony. The (compiled-in) default location for this file is `/etc/chrony.conf'. In the Debian package the configuration files are in the directory `/etc/chrony'. Assuming that your ntp servers are called a.b.c and d.e.f, your `chrony.conf' file could contain as a minimum

server a.b.c
server d.e.f
server g.h.i

However, you will probably want to include some of the other directives described later. The following directives will be particularly useful : driftfile, commandkey, keyfile. The smallest useful configuration file would look something like

server a.b.c
server d.e.f
server g.h.i
keyfile /etc/chrony.keys
commandkey 1
driftfile /etc/chrony.drift

Infrequent connection to true NTP servers

In this section we discuss how to configure chrony for computers that have occasional connections to the internet.

Setting up the configuration file for infrequent connections

As in the previous section, you will need access to NTP servers on the internet. The same remarks apply for how to find them.

In this case, you will need some additional configuration to tell chronyd when the connection to the internet goes up and down. This saves the program from continuously trying to poll the servers when they are inaccessible.

Again, assuming that your ntp servers are called a.b.c and d.e.f, your `chrony.conf' file would need to contain something like

server a.b.c
server d.e.f
server g.h.i

However, the following issues need to be addressed:

  1. Your computer probably doesn't have DNS access whilst offline to turn the machine names into IP addresses.
  2. Your computer will keep trying to contact the servers to obtain timestamps, even whilst offline. If you operate a dial-on-demand system, things are even worse, because the link to the internet will keep getting established.

For this reason, it would be better to specify this part of your configuration file in the following way:

server 1.2.3.4 offline
server 5.6.7.8 offline
server 9.10.11.12 offline

Because numeric IP addresses have been used, the first problem is overcome. The offline keyword indicates that the servers start in an offline state, and that they should not be contacted until chronyd receives notification that the link to the internet is present.

In order to notify chronyd of the presence of the link, you will need to be able to log in to it with the program chronyc. To do this, chronyd needs to be configured with an administrator password. The Debian package puts a default key in `/etc/chrony/chrony.keys'. You should change it. To set up an administrator password, you can create a file `/etc/chrony.keys' containing a single line

1 xyzzy

and add the following line to `/etc/chrony.conf' (the order of the lines does not matter)

commandkey 1

The smallest useful configuration file would look something like

server 1.2.3.4 offline
server 5.6.7.8 offline
server 9.10.11.12 offline
keyfile /etc/chrony.keys
commandkey 1
driftfile /etc/chrony.drift

The next section describes how to tell chronyd when the internet link goes up and down.

How to tell chronyd when the internet link is available.

To use this option, you will need to configure a command key in chronyd's configuration file `/etc/chrony.conf', as described in the previous section.

To tell chronyd when to start and finish sampling the servers, the online and offline commands of chronyc need to be used. To give an example of their use, we assume that pppd is the program being used to connect to the internet, and that chronyc has been installed at its default location `/usr/local/bin/chronyc'. We also assume that the command key has been set up as described in the previous section.

In the file `/etc/ppp/ip-up' we add the command sequence

cat <<EOF | /usr/local/bin/chronyc
password xyzzy
online
EOF

and in the file `/etc/ppp/ip-down' we add the sequence

cat <<EOF | /usr/local/bin/chronyc
password xyzzy
offline
EOF

The Debian package puts scripts similar to those above in the directories `/etc/ppp/ip-up.d' and `/etc/ppp/ip-down.d'.

chronyd's polling of the servers will now only occur whilst the machine is actually connected to the Internet.

Isolated networks

In this section we discuss how to configure chrony for computers that never have network conectivity to any computer which ultimately derives its time from a reference clock.

In this situation, one computer is selected to be the master timeserver. The other computers are either direct clients of the master, or clients of clients.

The rate value in the master's drift file needs to be set to the average rate at which the master gains or loses time. chronyd includes support for this, in the form of the manual directive in the configuration file and the settime command in the chronyc program.

If the master is rebooted, chronyd can re-read the drift rate from the drift file. However, the master has no accurate estimate of the current time. To get around this, the system can be configured so that the master can initially set itself to a `majority-vote' of selected clients' times; this allows the clients to `flywheel' the master across its outage.

A typical configuration file for the master (called master) might be (assuming the clients are in the 192.168.165.x subnet and that the master's address is 192.168.169.170)

driftfile /etc/chrony.drift
commandkey 25
keyfile /etc/chrony.keys
initstepslew 10 client1 client3 client6
local stratum 8
manual
allow 192.168.165

For the clients that have to resynchronise the master when it restarts, the configuration file might be

server master
driftfile /etc/chrony.drift
logdir /var/log/chrony
log measurements statistics tracking
keyfile /etc/chrony.keys
commandkey 24
local stratum 10
initstepslew 20 master
allow 192.168.169.170

The rest of the clients would be the same, except that the local and allow directives are not required.

The home PC with a dial-up connection

Assumptions/how the software works

This section considers the home computer which has a dial-up connection. It assumes that Linux is run exclusively on the computer. Dual-boot systems may work; it depends what (if anything) the other system does to the system's real-time clock.

Much of the configuration for this case is discussed earlier (see section Infrequent connection to true NTP servers). This section addresses specifically the case of a computer which is turned off between 'sessions'.

In this case, chronyd relies on the computer's real-time clock (RTC) to maintain the time between the periods when it is powered up. The arrangement is shown in the figure below.

            trim if required                          PSTN
      +---------------------------+               +----------+
      |                           |               |          |
      v                           |               |          |
+---------+                    +-------+       +-----+     +---+
| System's|  measure error/    |chronyd|       |modem|     |ISP|
|real-time|------------------->|       |-------|     |     |   |
|  clock  |   drift rate       +-------+       +-----+     +---+
+---------+                       ^                          |
      |                           |                          |
      +---------------------------+                  --o-----o---
         set time at boot up                           |
                                                  +----------+
                                                  |NTP server|
                                                  +----------+

When the computer is connected to the Internet (via the modem), chronyd has access to external NTP servers which it makes measurements from. These measurements are saved, and straight-line fits are performed on them to provide an estimate of the computer's time error and rate of gaining/losing time.

When the computer is taken offline from the Internet, the best estimate of the gain/loss rate is used to free-run the computer until it next goes online.

Whilst the computer is running, chronyd makes measurements of the real-time clock (RTC) (via the `/dev/rtc' interface, which must be compiled into the kernel). An estimate is made of the RTC error at a particular RTC second, and the rate at which the RTC gains or loses time relative to true time.

For kernels in the 2.0 series prior to 2.0.32, the kernel was set up to trim the RTC every 11 minutes. This would be disasterous for chronyd -- there is no reliable way of synchronising with this trimming. For this reason, chronyd only supports the RTC in 2.0 kernels from v2.0.32 onwards. (I don't know anything about the kernel's RTC behaviour in other kernel series).

When the computer is powered down, the measurement histories for all the NTP servers are saved to files (if the dumponexit directive is specified in the configuration file), and the RTC tracking information is also saved to a file (if the rtcfile directive has been specified). These pieces of information are also saved if the dump and writertc commands respectively are issued through chronyc.

When the computer is rebooted, chronyd reads the current RTC time and the RTC information saved at the last shutdown. This information is used to set the system clock to the best estimate of what its time would have been now, had it been left running continuously. The measurement histories for the servers are then reloaded.

The next time the computer goes online, the previous sessions' measurements can contribute to the line-fitting process, which gives a much better estimate of the computer's gain/loss rate.

One problem with saving the measurements and RTC data when the machine is shut down is what happens if there is a power failure; the most recent data will not be saved. Although chronyd is robust enough to cope with this, some performance may be lost. (The main danger arises if the RTC has been changed during the session, with the trimrtc command in chronyc. Because of this, trimrtc will make sure that a meaningful RTC file is saved out after the change is completed).

The easiest protection against power failure is to put the dump and writertc commands in the same place as the offline command is issued to take chronyd offline; because chronyd free-runs between online sessions, no parameters will change significantly between going offline from the Internet and any power failure.

A final point regards home computers which are left running for extended periods and where it is desired to spin down the hard disc when it is not in use (e.g. when not accessed for 15 minutes). chronyd has been planned so it supports such operation; this is the reason why the RTC tracking parameters are not saved to disc after every update, but only when the user requests such a write, or during the shutdown sequence. The only other facility that will generate periodic writes to the disc is the log rtc facility in the configuration file; this option should not be used if you want your disc to spin down.

Typical configuration files.

To illustrate how a dial-up home computer might be configured, example configuration files are shown in this section.

For the `/etc/chrony.conf' file, the following can be used as an example. NOTE : The server directives are only applicable to customers of Demon Internet; users of other ISPs will need to use their own ISP's NTP servers or public NTP servers.

server 158.152.1.65 minpoll 5 maxpoll 10 maxdelay 0.4 offline
server 158.152.1.76 minpoll 5 maxpoll 10 maxdelay 0.4 offline
server 194.159.253.2 minpoll 5 maxpoll 10 maxdelay 0.4 offline
logdir /var/log/chrony
log statistics measurements tracking
driftfile /etc/chrony.drift
keyfile /etc/chrony.keys
commandkey 25
maxupdateskew 100.0
dumponexit
dumpdir /var/log/chrony
rtcfile /etc/chrony.rtc

I use pppd for connecting to my ISP. This runs two scripts `/etc/ppp/ip-up' and `/etc/ppp/ip-down' when the link goes online and offline respectively.

The relevant part of the `/etc/ppp/ip-up' file is (with a dummy password)

cat <<EOF | /usr/local/bin/chronyc
password xxxxxxxx
online
EOF

and the relevant part of the `/etc/ppp/ip-down' script is

cat <<EOF | /usr/local/bin/chronyc
password xxxxxxxx
offline
dump
writertc
EOF

(Because they have to contain the administrator password, it would be desirable to make the files readable only by root on a multiuser machine).

To start chronyd during the boot sequence, I have the following in `/etc/rc.d/rc.local' (this is a Slackware system)

if [ -f /usr/local/sbin/chronyd -a -f /etc/chrony.conf ]; then
  /usr/local/sbin/chronyd -r -s
  echo "Start chronyd"
fi

The Debian package puts a script which handles this and shutdown in `/etc/init.d/chrony'.

The placement of this command may be important on some systems. In particular, chronyd may need to be started several seconds (about 10 as a minimum) before any software that depends on the system clock not jumping or moving backwards, depending on the directives in chronyd's configuration file.

For the system shutdown, chronyd should receive a SIGTERM several seconds before the final SIGKILL; the SIGTERM causes the measurement histories and RTC information to be saved out. There should be no need to add anything to the shutdown sequence, unless (as my system had) there is no pause between the SIGTERM and SIGKILL being delivered to the remaining processes. So if you find something like

killall5 -15 
killall5 -9

in your /etc/rc.d/rc.0 script, you will need to insert a sleep, e.g.

killall5 -15 
sleep 5
killall5 -9

Otherwise, chronyd will not always save information on shutdown, which could be a problem if you don't use dump and writertc when you go offline.

Other important configuration options

The most common option to include in the configuration file is the driftfile option. One of the major tasks of chronyd is to work out how fast or how slow the system clock runs relative to real time - e.g. in terms of seconds gained or lost per day. Measurements over a long period are usually required to refine this estimate to an acceptable degree of accuracy. Therefore, it would be bad if chronyd had to work the value out each time it is restarted, because the system clock would not run so accurately whilst the determination is taking place.

To avoid this problem, chronyd allows the gain or loss rate to be stored in a file, which can be read back in when the program is restarted. This file is called the drift file, and might typically be stored in `/etc/chrony.drift'. By specifying an option like the following

driftfile /etc/chrony.drift

in the configuration file (`/etc/chrony.conf'), the drift file facility will be activated.

Usage reference

Starting chronyd

If chronyd has been installed to its default location `/usr/local/sbin/chronyd', starting it is simply a matter of entering the command

/usr/local/sbin/chronyd

The Debian package uses `/usr/sbin/chronyd'.

Information messages and warnings will be logged to syslog.

The command line options supported are as follows:

-d
When run in this mode, the program will not detach itself from the terminal, and all messages will be sent to the terminal instead of to syslog.
-f <conf-file>
This option can be used to specify an alternate location for the configuration file (default `/etc/chrony.conf').
-r
This option will reload sample histories for each of the servers being used. These histories are created by using the dump command in chronyc, or by setting the dumponexit directive in the configuration file. This option is useful if you want to stop and restart chronyd briefly for any reason, e.g. to install a new version. However, it only makes sense on systems where the kernel can maintain clock compensation whilst not under chronyd's control. The only version where this happens so far is Linux. On systems where this is not the case, e.g. Solaris and SunOS the option should not be used.
-s
This option will set the system clock from the computer's real-time clock. This is analogous to supplying the `-s' flag to the `/sbin/clock' program during the Linux boot sequence. Support for real-time clocks is limited at present - the criteria are described in the section on the rtcfile directive (see section rtcfile). If chronyd cannot support the real time clock on your computer, this option cannot be used and a warning message will be logged to the syslog. If used in conjunction with the `-r' flag, chronyd will attempt to preserve the old samples after setting the system clock from the real time clock. This can be used to allow chronyd to perform long term averaging of the gain or loss rate across system reboots, and is useful for dial-up systems that are shut down when not in use. For this to work well, it relies on chronyd having been able to determine accurate statistics for the difference between the real time clock and system clock last time the computer was on.
-v
This option displays chronyd's version number to the terminal and exits.

On systems that support an `/etc/rc.local' file for starting programs at boot time, chronyd can be started from there.

On systems with a System V style initialisation (e.g. Solaris), a suitable start/stop script might be as shown below. This might be placed in the file `/etc/rc2.d/S83chrony'.

#!/bin/sh
# This file should have uid root, gid sys and chmod 744
#

killproc() {            # kill the named process(es)
        pid=`/usr/bin/ps -e |
             /usr/bin/grep -w $1 |
             /usr/bin/sed -e 's/^  *//' -e 's/ .*//'`
        [ "$pid" != "" ] && kill $pid
}

case "$1" in

'start')
   if [ -f /opt/free/sbin/chronyd -a -f /etc/chrony.conf ]; then
     /opt/free/sbin/chronyd
   fi
   ;;
'stop')
   killproc chronyd
   ;;
*)
   echo "Usage: /etc/rc2.d/S83chrony { start | stop }"
   ;;
esac

(In both cases, you may want to bear in mind that chronyd can step the time when it starts. There may be other programs started at boot time that could be upset by this, so you may need to consider the ordering carefully. However, chronyd will need to start after daemons providing services that it may require, e.g. the domain name service.)

The chronyd configuration file

The configuration file is normally called `/etc/chrony.conf'; in fact, this is the compiled-in default. However, other locations can be specified with a command line option.

Each command in the configuration file is placed on a separate line. The following sections describe each of the commands in turn. The directives can occur in any order in the file.

allow

The allow command is used to designate a particular subnet from which NTP clients are allowed to access the computer as an NTP server.

The default is that no clients are allowed access, i.e. chronyd operates purely as an NTP client. If the allow directive is used, chronyd will be both a client of its servers, and a server to other clients.

Examples of use of the command are as follows:

allow foo.bar.com
allow 1.2
allow 3.4.5
allow

The first command allows the named host to be an NTP client of this computer. The second command allows any host with an IP address of the form 1.2.x.y (with x and y arbitrary) to be an NTP client of this computer. Likewise, the third command allows any host with an IP address of the form 3.4.5.x to have client NTP access. The fourth form allows access by any node on the entire Internet.

A second form of the directive, allow all, has a greater effect, depending on the ordering of directives in the configuration file. To illustrate the effect, consider the two examples

allow 1.2.3.4
deny 1.2.3
allow 1.2

and

allow 1.2.3.4
deny 1.2.3
allow all 1.2

In the first example, the effect is the same regardles of what order the three directives are given in. So the 1.2.x.y subnet is allowed access, except for the 1.2.3.x subnet, which is denied access, however the host 1.2.3.4 is allowed access.

In the second example, the allow all 1.2 directives overrides the effect of any previous directive relating to a subnet within the specified subnet. Within a configuration file this capability is probably rather moot; however, it is of greater use for reconfiguration at run-time via chronyc (see section allow all).

Note, if the initstepslew directive (see section initstepslew) is used in the configuration file, each of the computers listed in that directive must allow client access by this computer for it to work.

commandkey

The commandkey command is used to set the key number used for authenticating user commands via the chronyc program at run time. This allows certain actions of the chronyc program to be restricted to administrators.

An example of the commandkey command is

commandkey 20

In the key file (see the keyfile command) there should be a line of the form

20 foobar

When running the chronyc program to perform run-time configuration, the command

password foobar

must be entered before any commands affecting the operation of the daemon can be entered.

cmdport

The cmdport directive allows the port that is used for run-time command and monitoring (via the program chronyc) to be altered from its default (323/udp).

An example shows the syntax

cmdport 257

This would make chronyd use 257/udp as its command port. (chronyc would need to be run with the -p 257 switch to inter-operate correctly).

deny

This is similar to the allow directive (see section allow), except that it denies NTP client access to a particular subnet or host, rather than allowing it.

The syntax is identical.

There is also a deny all directive with similar behaviour to the allow all directive.

driftfile

One of the main activities of the chronyd program is to work out the rate at which the system clock gains or loses time relative to real time.

Whenever chronyd computes a new value of the gain/loss rate, it is desirable to record it somewhere. This allows chronyd to begin compensating the system clock at that rate whenever it is restarted, even before it has had a chance to obtain an equally good estimate of the rate during the new run. (This process may take many minutes, at least).

The driftfile command allows a file to be specified into which chronyd can store the rate information. Two parameters are recorded in the file. The first is the rate at which the system clock gains or loses time, expressed in parts per million, with gains positive. Therefore, a value of 100.0 indicates that when the system clock has advanced by a second, it has gained 100 microseconds on reality (so the true time has only advanced by 999900 microseconds). The second is an estimate of the error bound around the first value in which the true rate actually lies.

An example of the driftfile command is

driftfile /etc/chrony.drift

dumpdir

To compute the rate of gain or loss of time, chronyd has to store a measurement history for each of the time sources it uses.

Certain systems (so far only Linux) have operating system support for setting the rate of gain or loss to compensate for known errors. (On other systems, chronyd must simulate such a capability by periodically slewing the system clock forwards or backwards by a suitable amount to compensate for the error built up since the previous slew).

For such systems, it is possible to save the measurement history across restarts of chronyd (assuming no changes are made to the system clock behaviour whilst it is not running). If this capability is to be used (via the dumponexit command in the configuration file, or the dump command in chronyc), the dumpdir command should be used to define the directory where the measurement histories are saved.

An example of the command is

dumpdir /var/log/chrony

A source whose IP address is 1.2.3.4 would have its measurement history saved in the file `/var/log/chrony/1.2.3.4.dat'.

dumponexit

If this command is present, it indicates that chronyd should save the measurement history for each of its time sources recorded whenever the program exits. (See the dumpdir command above).

initstepslew

In normal operation, chronyd always slews the time when it needs to adjust the system clock. For example, to correct a system clock which is 1 second slow, chronyd slightly increases the amount by which the system clock is advanced on each clock interrupt, until the error is removed. (Actually, this is done by calling the adjtime() or similar system function which does it for us.) Note that at no time does time run backwards with this method.

On most Unix systems it is not desirable to step the system clock, because many programs rely on time advancing monotonically forwards.

When the chronyd daemon is initially started, it is possible that the system clock is considerably in error. Attempting to correct such an error by slewing may not be sensible, since it may take several hours to correct the error by this means.

The purpose of the initstepslew directive is to allow chronyd to make a rapid measurement of the system clock error at boot time, and to correct the system clock by stepping before normal operation begins. Since this would normally be performed only at an appropriate point in the system boot sequence, no other software should be adversely affected by the step.

If the correction required is less than a specified threshold, a slew is used instead. This makes it easier to restart chronyd whilst the system is in normal operation.

The initstepslew directive takes a threshold and a list of NTP servers as arguments. A maximum of 8 will be used. Each of the servers is rapidly polled several times, and a majority voting mechanism used to find the most likely range of system clock error that is present. A step (or slew) is applied to the system clock to correct this error. chronyd then enters its normal operating mode (where only slews are used).

An example of use of the command is

initstepslew 30 foo.bar.com baz.quz.com

where 2 NTP servers are used to make the measurement. The 30 indicates that if the system's error is found to be 30 seconds or less, a slew will be used to correct it; if the error is above 30 seconds, a step will be used.

The initstepslew directive can also be used in an isolated LAN environment, where the clocks are set manually. The most stable computer is chosen as the master, and the other computers are slaved to it. If each of the slaves is configured with the local option (see below), the master can be set up with an initstepslew directive which references some or all of the slaves. Then, if the master machine has to be rebooted, the slaves can be relied on to 'flywheel' the time for the master.

keyfile

This command is used to specify the location of the file containing ID/key pairs for the following 2 uses:

The format of the command is shown in the example below

keyfile /etc/chrony.keys

The argument is simply the name of the file containing the ID/key pairs. The format of the file is shown below

10 tulip
11 hyacinth
20 crocus
25 iris
 ...

Each line consists of an ID and a password. The ID can be any unsigned integer in the range 0 through 2**32-1. The password can be any string of characters not containing a space.

For NTP use, the MD5 authentication scheme is always used. This must be borne in mind if chronyd is to inter-operate in authenticated mode with xntpd running on other computers.

The ID for the chronyc authentication key is specified with the commandkey command (see earlier).

local

The local keyword is used to allow chronyd to appear synchronised to real time (from the viewpoint of clients polling it), even if it has no current synchronisation source.

This option is normally used on computers in an isolated network, where several computers are required to synchronise to one other, this being the "master" which is kept vaguely in line with real time by manual input.

An example of the command is

local stratum 10

The value 10 may be substituted with other values in the range 1 through 15. Stratum 1 indicates a computer that has a true real-time reference directly connected to it (e.g. GPS, atomic clock etc) &ndash; such computers are expected to be very close to real time. Stratum 2 computers are those which have a stratum 1 server; stratum 3 computers have a stratum 2 server and so on.

A large value of 10 indicates that the clock is so many hops away from a reference clock that its time is fairly unreliable. Put another way, if the computer ever has access to another computer which is ultimately synchronised to a reference clock, it will almost certainly be at a stratum less than 10. Therefore, the choice of a high value like 10 for the local command prevents the machine's own time from ever being confused with real time, were it ever to leak out to clients that have visibility of real servers.

log

The log command indicates that certain information is to be logged.

measurements
This option logs the raw NTP measurements and related information to a file called measurements.log.
statistics
This option logs information about the regression processing to a file called statistics.log.
tracking
This option logs changes to the estimate of the system's gain or loss rate, and any slews made, to a file called tracking.log.
rtc
This option logs information about the system's real-time clock.

The files are written to the directory specified by the logdir command.

An example of the command is

log measurements statistics tracking

Measurements log file format

An example line (which actually appears as a single line in the file) from the measurements log file is shown below.

22Jul98 05:40:50 158.152.1.76    N  8 1111 11 1111 10 10  1 \
   -4.966e-03  2.296e-01  1.577e-05  1.615e-01  7.446e-03

The columns are as follows (the quantities in square brackets are the values from the example line above) :

  1. Date [22Jul98]
  2. Hour:Minute:Second [05:40:50]. Note that the date/time pair is expressed in UTC, not the local time zone.
  3. IP address of server/peer from which measurement comes [158.152.1.76]
  4. Leap status (N means normal, - means that the last minute of today has 61 seconds, + means that the last minute of the day has 59 seconds, ? means the remote computer is not currently synchronised.) [N]
  5. Stratum of remote computer. [2]
  6. RFC1305 tests 1 through 4 (1=pass, 0=fail) [1111]
  7. Tests for maximum delay and maximum delay ratio, against user defined parameters (1=pass, 0=fail) [11]
  8. RFC1305 tests 5 through 8 (1=pass, 0=fail) [1111]
  9. Local poll [10]
  10. Remote poll [10]
  11. `Score' (an internal score within each polling level used to decide when to increase or decrease the polling level. This is adjusted based on changes to the variance of the measurements obtained from the source). [1]
  12. The estimated local clock error (`theta' in RFC1305). Positive indicates that the local clock is slow. [-4.966e-03].
  13. The peer delay (`delta' in RFC1305). [2.296e-01]
  14. The peer dispersion (`epsilon' in RFC1305). [1.577e-05]
  15. The root delay (`Delta' in RFC1305). [1.615e-01]
  16. The root dispersion (`E' in RFC1305). [7.446e-03]

A banner is periodically written to the log file to indicate the meanings of the columns.

Statistics log file format

An example line (which actually appears as a single line in the file) from the measurements log file is shown below.

22Jul98 05:40:50 158.152.1.76     6.261e-03 -3.247e-03 \
     2.220e-03  1.874e-06  1.080e-06 7.8e-02  16   0   8

The columns are as follows (the quantities in square brackets are the values from the example line above) :

  1. Date [22Jul98]
  2. Hour:Minute:Second [05:40:50]. Note that the date/time pair is expressed in UTC, not the local time zone.
  3. IP address of server/peer from which measurement comes [158.152.1.76]
  4. The estimated standard deviation of the measurements from the source (in seconds). [6.261e-03]
  5. The estimated offset of the source (in seconds, positive means the local clock is estimated to be fast, in this case). [-3.247e-03]
  6. The estimated standard deviation of the offset estimate (in seconds). [2.220e-03]
  7. The estimated rate at which the local clock is gaining or losing time relative to the source (in seconds per second, positive means the local clock is gaining). This is relative to the compensation currently being applied to the local clock, not to the local clock without any compensation. [1.874e-06]
  8. The estimated error in the rate value (in seconds per second). [1.080e-06].
  9. The ration of |old_rate - new_rate| / old_rate_error. Large values indicate the statistics are not modelling the source very well. [7.8e-02]
  10. The number of measurements currently being used for the regression algorithm. [16]
  11. The new starting index (the oldest sample has index 0; this is the method used to prune old samples when it no longer looks like the measurements fit a linear model). [0, i.e. no samples discarded this time]
  12. The number of runs. The number of runs of regression residuals with the same sign is computed. If this is too small it indicates that the measurements are no longer represented well by a linear model and that some older samples need to be discarded. The number of runs for the data that is being retained is tabulated. Values of approximately half the number of samples are expected. [8]

A banner is periodically written to the log file to indicate the meanings of the columns.

Tracking log file format

An example line (which actually appears as a single line in the file) from the measurements log file is shown below.

22Jul98 05:40:50 158.152.1.76     3    340.529      1.606  1.046e-03

The columns are as follows (the quantities in square brackets are the values from the example line above) :

  1. Date [22Jul98]
  2. Hour:Minute:Second [05:40:50]. Note that the date/time pair is expressed in UTC, not the local time zone.
  3. The IP address of the server/peer to which the local system is synchronised. [158.152.1.76]
  4. The stratum of the local system. [3]
  5. The local system frequency (in ppm, positive means the local system runs fast of UTC). [340.529]
  6. The error bounds on the frequency (in ppm) [1.606]
  7. The estimated local offset at the epoch (which is rapidly corrected by slewing the local clock. (In seconds, positive indicates the local system is fast of UTC). [1.046e-3]

A banner is periodically written to the log file to indicate the meanings of the columns.

Real-time clock log file format

An example line (which actually appears as a single line in the file) from the measurements log file is shown below.

22Jul98 05:40:50     -0.037360 1       -0.037434\
          -37.948  12   5  120

The columns are as follows (the quantities in square brackets are the values from the example line above) :

  1. Date [22Jul98]
  2. Hour:Minute:Second [05:40:50]. Note that the date/time pair is expressed in UTC, not the local time zone.
  3. The measured offset between the system's real time clock and the system (gettimeofday()) time. In seconds, positive indicates that the RTC is fast of the system time. [-0.037360].
  4. Flag indicating whether the regression has produced valid coefficients. (1 for yes, 0 for no). [1]
  5. Offset at the current time predicted by the regression process. A large difference between this value and the measured offset tends to indicate that the measurement is an outlier with a serious measurement error. [-0.037434].
  6. The rate at which the RTC is losing or gaining time relative to the system clock. In ppm, with positive indicating that the RTC is gaining time. [-37.948]
  7. The number of measurements used in the regression. [12]
  8. The number of runs of regression residuals of the same sign. Low values indicate that a straight line is no longer a good model of the measured data and that older measurements should be discarded. [5]
  9. The measurement interval used prior to the measurement being made (in seconds). [120]

A banner is periodically written to the log file to indicate the meanings of the columns.

logdir

This command allows the directory where log files are written to be specified.

An example of the use of this command is

logdir /var/log/chrony

manual

The manual directive enables support at run-time for the settime command in chronyc (see section settime). If no manual directive is included, any attempt to use the settime command in chronyc will be met with an error message.

Note that the settime command can be enabled at run-time using the manual command in chronyc (see section manual). (The idea of the two commands is that the manual command controls the manual clock driver's behaviour, whereas the settime command allows samples of manually entered time to be provided).

maxupdateskew

One of chronyd's tasks is to work out how fast or slow the computer's clock runs relative to its reference sources. In addition, it computes an estimate of the error bounds around the estimated value.

If the range of error is too large, it probably indicates that the measurements have not settled down yet, and that the estimated gain or loss rate is not very reliable.

The maxupdateskew parameter allows the threshold for determining whether an estimate may be so unreliable that it should not be used.

The syntax is

maxupdateskew <skew-in-ppm>

Typical values for <skew-in-ppm> might be 100 for a dial-up connection to servers over a phone line, and 5 or 10 for a computer on a LAN.

It should be noted that this is not the only means of protection against using unreliable estimates. At all times, chronyd keeps track of both the estimated gain or loss rate, and the error bound on the estimate. When a new estimate is generated following another measurement from one of the sources, a weighted combination algorithm is used to update the master estimate. So if chronyd has an existing highly-reliable master estimate and a new estimate is generated which has large error bounds, the existing master estimate will dominate in the new master estimate.

peer

The syntax of this directive is identical to that for the server directive (see section server), except that it is used to specify an NTP peer rather than an NTP server.

port

This option allows you to configure the port used for the NTP service on your machine.

The compiled in default is udp/123, the standard NTP port. It is unlikely that you would ever need to change this value. A possible exception would be if you wanted to operate strictly in client-only mode and never be available as a server to xntpd clients.

An example of the port command is

port 11123

This would change the NTP port served by chronyd on the computer to udp/11123.

rtcfile

The rtcfile directive defines the name of the file in which chronyd can save parameters associated with tracking the accuracy of the system's real-time clock (RTC).

The syntax is illustrated in the following example

rtcfile /etc/chrony.rtc

chronyd saves information in this file when it exits and when the writertc command is issued in chronyc. The information saved is the RTC's error at some epoch, that epoch (in seconds since January 1 1970), and the rate at which the RTC gains or loses time.

So far, the support for real-time clocks is limited - their code is even more system-specific than the rest of the software. You can only use the real time clock facilities (the rtcfile directive and the -s command line option to chronyd) if the following three conditions apply:

  1. You are running Linux v2.0.x with x>=32.
  2. You have compiled the kernel with extended real-time clock support (i.e. the `/dev/rtc' device is capable of doing useful things).
  3. You don't have other applications that need to make use of `/dev/rtc' at all.

rtconutc

chronyd assumes by default that the real time clock (RTC) keeps local time (including any daylight saving changes). This is convenient on PCs running Linux which are dual-booted with DOS or Windows.

An alternative is for the RTC to keep Universal Coordinated Time (UTC).

If the rtconutc directive appears, it means the RTC is required to keep UTC. The directive takes no arguments. It is equivalent to specifying the -u switch to the Linux `/sbin/clock' program.

server

The server directive allows NTP servers to be specified. The client/server relationship is strictly hierarchical : a client may synchronise its system time to that of the server, but the server's system time will never be influenced by that of a client.

The server directive is immediately followed by either the name of the server, or its IP address in dotted-quad notation. The server command also supports a number of subfields (which may be defined in any order):

port
This option allows the UDP port on which the server understands NTP requests to be specified. For normal servers this option should not be required (the default is 123, the standard NTP port).
minpoll
Although chronyd will trim the rate at which it samples the server during normal operation, the user may wish to constrain the minimum polling interval. This is always defined as a power of 2, so <tt/minpoll 5/ would mean that the polling interval cannot drop below 32 seconds. The default is 6 (64 seconds).
maxpoll
In a similar way, the user may wish to constrain the maximum polling interval. Again this is specified as a power of 2, so <tt/maxpoll 9/ indicates that the polling interval must stay at or below 512 seconds. The default is 10 (1024 seconds).
maxdelay
chronyd uses the network round-trip delay to the server to determine how accurate a particular measurement is likely to be. Long round-trip delays indicate that the request, or the response, or both were delayed. If only one of the messages was delayed the measurement error is likely to be substantial. For small variations in round trip delay, chronyd uses a weighting scheme when processing the measurements. However, beyond a certain level of delay the measurements are likely to be so corrupted as to be useless. (This is particularly so on dial-up or other slow links, where a long delay probably indicates a highly asymmetric delay caused by the response waiting behind a lot of packets related to a download of some sort). If the user knows that round trip delays above a certain level should cause the measurement to be ignored, this level can be defined with the maxdelay command. For example, <tt/maxdelay 0.3/ would indicate that measurements with a round-trip delay of 0.3 seconds or more should be ignored.
maxdelayratio
This option is similar to the maxdelay option above. chronyd keeps a record of the minimum round-trip delay amongst the previous measurements that it has buffered. If a measurement has a round trip delay that is greater than the maxdelayratio times the minimum delay, it will be rejected.
presend
If the timing measurements being made by chronyd are the only network data passing between two computers, you may find that some measurements are badly skewed due to either the client or the server having to do an ARP lookup on the other party prior to transmitting a packet. This is more of a problem with long sampling intervals, which may be similar in duration to the lifetime of entries in the ARP caches of the machines. In order to avoid this problem, the presend option may be used. It takes a single integer argument, which is the smallest polling interval for which a pair of packets will be exchanged between the client and the server prior to the actual measurement being initiated by the client. For example, with the following option included in a server directive :
presend 9
when the polling interval is 512 seconds or more, a UDP echo datagram will be sent to the server a short time (currently 4 seconds) before the NTP client mode datagram.
key
The NTP protocol supports the inclusion of checksums in the packets, to prevent computers having their system time upset by rogue packets being sent to them. The checksums are generated as a function of a password, using the MD5 algorithm. The association between key numbers and passwords is contained in the keys file, defined by the keyfile command. If the key option is present, chronyd will attempt to use authenticated packets when communicating with this server. The key number used will be the single argument to the key option. The server must have the same password for this key number configured, otherwise no relationship between the computers will be possible.
offline
If the server will not be reachable when chronyd is started, the offline option may be specified. chronyd will not try to poll the server until it is enabled to do so (by using the online option of chronyc).

Running chronyc

Chronyc is the program that can be used to reconfigure options within the chronyd program whilst it is running. Chronyc can also be used to generate status reports about the operation of chronyd.

Basic use

The program chronyc is run by entering

chronyc

at the command line. The prompt chronyc is displayed whilst chronyc is expecting input from the user, when it is being run from a terminal. If chronyc's input or output are redirected from/to a file, the prompt is ow shown.

When you are finished entering commands, the commands exit or quit will terminate the program. (Entering Control-D will also terminate the program.)

Command line options

Chronyc supports the following command line options.

-v
Displays the version number of chronyc on the terminal, and exists.
-h <host>
This option allows the user to specify which host running the chronyd program is to be contacted. This allows for remote configuration, without having to telnet or rlogin to the other host first. The default is to contact chronyd running on the same host as that where chronyc is being run.
-p <port>
This option allows the user to specify the UDP port number which the target chronyd is using for its command & monitoring connections. This defaults to the compiled-in default; there would rarely be a need to change this.

Security with chronyc

Many of the commands available through chronyc have a fair amount of power to reconfigure the run-time behaviour of chronyd. Consequently, chronyc is quite dangerous for the integrity of the target system's clock performance. Having access to chronyd via chronyc is more or less equivalent to being able to modify chronyd's configuration file (typically `/etc/chrony.conf') and to restart chronyd.

Chronyc also provides a number of monitoring (as opposed to commanding) commands, which will not affect the behaviour of chronyd. However, you may still want to restrict access to these commands.

In view of this, access to some of the capabilities of chronyc will usually be tightly controlled. There are two mechanisms supported:

  1. The set of hosts from which chronyd will accept commands can be restricted. By default, commands will only be accepted from the same host that chronyd is running on.
  2. Any command that actually reconfigures some aspect of chronyd's behaviour requires the user of chronyc to know a password. This password is specified in chronyd's keys file (see section keyfile) and specified via the commandkey option in its configuration file (see section commandkey).

Only the following commands can be used without providing a password:

All other commands require a password to have been specified previously, because they affect chronyd's operation.

Command reference

This section describes each of the commands available within the chronyc program. Chronyc offers the user a simple command-line driven interface.

accheck

This command allows you to check whether client NTP access is allowed from a particular host.

Examples of use, showing a named host and a numeric IP address, are as follows:

accheck a.b.c
accheck 1.2.3.4

This command can be used to examine the effect of a series of allow, allow all, deny and deny all commands specified either via chronyc, or in chronyd's configuration file.

add peer

The add peer command allows a new NTP peer to be added whilst chronyd is running.

Following the words add peer, the syntax of the following parameters and options is identical to that for the peer directive in the configuration file (see section peer).

An example of using this command is shown below.

add peer foo.bar.com minpoll 6 maxpoll 10 authkey 25

add server

The add server command allows a new NTP server to be added whilst chronyd is running.

Following the words add server, the syntax of the following parameters and options is identical to that for the server directive in the configuration file (see section server).

An example of using this command is shown below.

add server foo.bar.com minpoll 6 maxpoll 10 authkey 25

allow

The effect of the allow command is identical to the allow directive in the configuration file (see section allow).

The syntax is illustrated in the following examples:

allow foo.bar.com
allow 1.2
allow 3.4.5

allow all

The effect of the allow command is identical to the allow all directive in the configuration file (see section allow).

burst

The burst command tells chronyd to make a set of measurements to each of its sources over a short duration (rather than the usual periodic measurements that it makes). After such a burst, chronyd will revert to the previous state for each source. This might be either online, if the source was being periodically measured in the normal way, or offline, if the source had been indicated as being offline. (Switching a source between the online and offline states is described in section online, section offline).

The syntax of the burst command is as follows

burst <n-good-measurements>/<max-measurements> [<mask>/<masked-address>]

The mask and masked-address arguments are optional, in which case chronyd will initiate a burst for all of its currently defined sources.

The arguments have the following meaning and format.

n-good-measurements
This defines the number of good measurements that chronyd will want to obtain from each source. A measurement is good if it passes certain tests, for example, the round trip time to the source must be acceptable. (This allows chronyd to reject measurements that are likely to be bogus.)
max-measurements
This defines the maximum number of measurements that chronyd will attempt to make, even if the required number of good measurements has not been obtained.
mask
This is a dotted quad argument (e.g. 255.255.255.0) with which the IP address of each of chronyd's sources is to be masked.
masked-address
This is a dotted quad argument (e.g. 1.2.3.0). If the masked IP address of a source matches this value then the burst command is applied to that source.

If no mask or masked address arguments are provided, the default is 0.0.0.0 and 0.0.0.0 respectively, which will match every source.

An example of the two-argument form of the command is

burst 2/10

This will cause chronyd to attempt to get two good measurements from each source, stopping after two have been obtained, but in no event will it try more than ten probes to the source.

An example of the four-argument form of the command is

burst 2/10 255.255.0.0/1.2.0.0

In this case, the two out of ten sampling will only be applied to sources whose IP addresses are of the form 1.2.x.y, where x and y are arbitrary.

cmdaccheck

This command is similar to the accheck command, except that it is used to check whether command access is permitted from a named host.

Examples of use are as follows:

cmdaccheck a.b.c
cmdaccheck 1.2.3.4

cmdallow

This is similar to the allow command, except that it is used to allow particular hosts or subnets to use the chronyc program to interact with chronyd on the current host.

cmdallow all

This is similar to the allow all command, except that it is used to allow particular hosts or subnets to use the chronyc program to interact with chronyd on the current host.

cmddeny

This is similar to the deny command, except that it is used to allow particular hosts or subnets to use the chronyc program to interact with chronyd on the current host.

cmddeny all

This is similar to the deny all command, except that it is used to allow particular hosts or subnets to use the chronyc program to interact with chronyd on the current host.

cyclelogs

The cyclelogs command causes all of chronyd's open log files to be closed and re-opened. This allows them to be renamed so that they can be periodically purged. An example of how to do this is shown below.

% mv /var/log/chrony/measurements.log /var/log/chrony/measurements1.log
% chronyc
chronyc> password aardvark
200 OK
chronyc> cyclelogs
200 OK
chronyc> exit
% ls -l /var/log/chrony
-rw-r--r--   1 root     root            0 Jun  8 18:17 measurements.log
-rw-r--r--   1 root     root        12345 Jun  8 18:17 measurements1.log

delete

The delete command allows an NTP server or peer to be removed from the current set of sources.

The syntax is illustrated in the examples below.

delete foo.bar.com
delete 1.2.3.4

There is one parameter, the name or IP address of the server or peer to be deleted.

deny

The effect of the allow command is identical to the deny directive in the configuration file (see section deny).

The syntax is illustrated in the following examples:

deny foo.bar.com
deny 1.2
deny 3.4.5

deny all

The effect of the allow command is identical to the deny all directive in the configuration file (see section deny).

dump

The dump command causes chronyd to write its current history of measurements for each of its sources to dump files, either for inspection or to support the -r option when chronyd is restarted.

The dump command is somewhat equivalent to the dumponexit directive in the chrony configuration file. See section dumponexit.

To use the dump, you probably want to configure the name of the directory into which the dump files will be written. This can only be done in the configuration file, see section dumpdir.

exit

The exit command exits from chronyc and returns the user to the shell (same as the quit command).

help

The help command displays a summary of the commands and their arguments.

local

The local command allows chronyd to be told that it is to appear as a reference source, even if it is not itself properly synchronised to an external source. (This can be used on isolated networks, to allow one computer to be a master time server with the other computers slaving to it.) The local command is somewhat equivalent to the local directive in the configuration file, see section local.

The syntax is as shown in the following examples.

local stratum 10
local off

The first example enables the local reference mode on the host, and sets the stratum at which it should claim to be synchronised.

The second example disables the local reference mode.

manual

The manual command enables and disables use of the settime command (see section settime), and is used to modify the behaviour of the manual clock driver.

Examples of the command are shown below.

manual on
manual off
manual reset

The first form of the command enables use of the settime command.

The second form of the command disables use of the settime command.

When a time is entered with the settime command, the current time and machine clock error are stored within chronyd. When another time is entered later, the rate of gain or loss of the machine clock is computed and used to adjust the running speed of the machine clock.

If you want to correct the machine time without this rate compensation being performed, use the third form of the manual command first. This will purge any previous measurements stored within chronyd.

maxdelay

This allows the maxdelay option for one of the sources to be modified, in the same way as specifying the maxdelay option for the server directive in the configuration file (see section server).

The following examples illustrate the syntax

maxdelay foo.bar.com 0.3
maxdelay 1.2.3.4 0.0015

The first example sets the maximum network delay allowed for a measurement to the host foo.bar.com to 0.3 seconds. The second example sets the maximum network delay for a measurement to the host with IP address 1.2.3.4 to 1.5 milliseconds.

(Any measurement whose network delay exceeds the specified value is discarded.)

maxdelayratio

This allows the maxdelayratio option for one of the sources to be modified, in the same way as specifying the maxdelayratio option for the server directive in the configuration file (see section server).

The following examples illustrate the syntax

maxdelayratio foo.bar.com 1.5
maxdelayratio 1.2.3.4 2.0

The first example sets the maximum network delay for a measurement to the host foo.bar.com to be 1.5 times the minimum delay found amongst the previous measurements that have been retained. The second example sets the maximum network delay for a measurement to the host with IP address 1.2.3.4 to be double the retained minimum.

As for maxdelay, any measurement whose network delay is too large will be discarded.

maxpoll

The maxpoll command is used to modify the minimum polling interval for one of the current set of sources. It is equivalent to the maxpoll option in the server directive in the configuration file (see section server).

The syntax is as follows

maxpoll <host> <new-maxpoll>

where the host can be specified as either a machine name or dotted-quad IP address. The new minimum poll is specified as a base-2 logarithm of the number of seconds between polls (e.g. specify 6 for 64 second sampling).

An example is

maxpoll foo.bar.com 10

which sets the maximum polling interval for the host foo.bar.com to 1024 seconds.

Note that the new maximum polling interval only takes effect after the next measurement has been made.

maxupdateskew

This command has the same effect as the maxupdateskew directive in the configuration file, see section maxupdateskew.

minpoll

The minpoll command is used to modify the minimum polling interval for one of the current set of sources. It is equivalent to the minpoll option in the server directive in the configuration file (see section server).

The syntax is as follows

minpoll <host> <new-minpoll>

where the host can be specified as either a machine name or dotted-quad IP address. The new minimum poll is specified as a base-2 logarithm of the number of seconds between polls (e.g. specify 6 for 64 second sampling).

An example is

minpoll foo.bar.com 5

which sets the minimum polling interval for the host foo.bar.com to 32 seconds.

Note that the new minimum polling interval only takes effect after the next measurement has been made.

offline

The offline command is used to warn chronyd that the network connection to a particular host or hosts is about to be lost. It should be used on computers with a dial-up or similar connection to their time sources, to warn chronyd that the connection is about to be broken.

An example of how to use offline in this case is shown in section How to tell chronyd when the internet link is available..

Another case where offline could be used is where a computer serves time to a local group of computers, and has a permanant connection to true time servers outside the organisation. However, the external connection is heavily loaded at certain times of the day and the measurements obtained are less reliable at those times. In this case, it is probably most useful to determine the gain/loss rate during the quiet periods and let the whole network coast through the loaded periods. The offline and online commands can be used to achieve this. The situation is shown in the figure below.

          +----------+
          |Ext source|
          +----------+
              |
              |
              |/| <-- Link with variable
                |     reliability
                |
      +-------------------+
      |Local master server|
      +-------------------+
                |
  +---+---+-----+-----+----+----+
  |   |   |     |     |    |    |
           Local clients

If the source to which chronyd is currently synchronised is indicated offline in this way, chronyd will continue to treat it as the synchronisation source. If the network connection were broken without the offline command being used, chronyd would assume that the source had failed and would attempt to pick another synchronisation source.

There are two forms of the offline command. The first form is a wildcard, meaning all sources. The second form allows a IP address mask and a masked address to be specified. These forms are illustrated below.

offline
offline 255.255.255.0/1.2.3.0

The second form means that the offline command is to be applied to any source whose IP address is in the 1.2.3 subnet. (The host's address is logically and-ed with the mask, and if the result matches the masked-address the host is processed).

The wildcard form of the address is actually equivalent to

offline 0.0.0.0/0.0.0.0

online

The online command is opposite in function to the offline command. It is used to advise chronyd that network connectivity to a particular source or sources has been restored.

The syntax is identical to that of the offline command, see section offline.

password

The password command is used to allow chronyc to send privileged commands to chronyd. The password can either be entered on the command line, or can be entered without echoing. The syntax for entering the password on the command line is as follows

password xyzzy

To enter the password without it being echoed, enter

password

The computer will respond with a `Password:' prompt, at which you should enter the password and press return. (Note that the no-echo mode is limited to 8 characters on SunOS 4.1 due to limitations in the system library. Other systems do not have this restriction.)

The password is any string of characters not containing whitespace. It has to match chronyd's currently defined command key (see section commandkey).

quit

The quit command exits from chronyc and returns the user to the shell (same as the exit command).

rtcdata

The rtcdata command displays the current real time clock RTC parameters.

An example output is shown below.

RTC ref time (GMT) : Sat May 30 07:25:56 1998
Number of samples  : 10
Number of runs     : 5
Sample span period :  549
RTC is fast by     :    -1.632736 seconds
RTC gains time at  :  -107.623 ppm

The fields have the following meaning

RTC ref time (GMT)
This is the RTC reading the last time its error was measured.
Number of samples
This is the number of previous measurements being used to determine the RTC gain/loss rate.
Number of runs
This is the number of runs of residuals of the same sign following the regression fit for (RTC error) versus (RTC time). A value which is small indicates that the measurements are not well approximated by a linear model, and that the algorithm will tend to delete the older measurements to improve the fit.
Sample span period
This is the period that the measurements span (from the oldest to the newest). Without a unit the value is in seconds; suffixes `m' for minutes, `h' for hours, `d' for days or `y' for years may be used.
RTC is fast by
This is the estimate of how many seconds fast the RTC when it thought the time was at the reference time (above). If this value is large, you may (or may not) want to use the trimrtc command to bring the RTC into line with the system clock. (Note, a large error will not affect chronyd's operation, unless it becomes so big as to start causing rounding errors.
RTC gains time at
This is the amount of time gained (positive) or lost (negative) by the real time clock for each second that it ticks. It is measured in parts per million. So if the value shown was +1, suppose the RTC was exactly right when it crosses a particular second boundary. Then it would be 1 microsecond fast when it crosses its next second boundary.

settime

The settime command allows the current time to be entered manually, if this option has been configured into chronyd. (It may be configured either with the manual directive in the configuration file (see section manual), or with the manual command of chronyc (see section manual).

It should be noted that the computer's sense of time will only be as accurate as the reference you use for providing this input (e.g. your watch), as well as how well you can time the press of the return key. When inputting time to an isolated network, I have a battery operated alarm clock that is synchronised to the Rugby MSF time signal in the UK.

Providing your computer's time zone is set up properly, you will be able to enter a local time (rather than UTC).

The response to a successful settime command indicates the amount that the computer's clock was wrong. It should be apparent from this if you have entered the time wrongly, e.g. with the wrong time zone.

The software always allows you to enter the time again, replacing the measurement you just entered, for example if you realise that it was wrong. You are given a 5 minute period for doing this.

The time is parsed by the public domain `getdate' algorithm. Consequently, you can only specify time to the nearest second.

Examples of inputs that are valid are shown below.

settime 16:30
settime 16:30:05
settime Nov 21, 1997 16:30:05

For a full description of getdate, get hold of the getdate documentation (bundled, for example, with the source for GNU tar).

sources

This command displays information about the current time sources that chronyd is accessing. It takes no arguments.

210 Number of sources = 3
MS Name/IP address      Str  Poll LastRx        Last sample
===================================================================
^+ a.b.c                  3    6   47m  -9491us[-6983us] +/-  159ms
^+ d.e.f                  3    6   47m    +32ms[  +35ms] +/-  274ms
^* g.h.i                  2    6   47m  +8839us[  +11ms] +/-  214ms

The columns are as follows:

M
This indicates the mode of the source. ^ means a server, = means a peer and # indicates a locally connected reference clock(1).
S
This column indicates the state of the sources. * indicates the source to which chronyd is current synchronised. + indicates other acceptable sources. ? indicates sources to which connectivity has been lost. x indicates a clock which chronyd thinks is is a falseticker (i.e. its time is inconsistent with a majority of other sources). ~ indicates a source whose time appears to have too much variability. The ~ condition is also shown at start-up, until at least 3 samples have been gathered from it.
IP address
This shows the name or the IP address of the source.
Str
This shows the stratum of the source, as reported in its most recently received sample. Stratum 1 indicates a computer with a locally attached reference clock. A computer that is synchronised to a stratum 1 computer is at stratum 2. A computer that is synchronised to a stratum 2 computer is at stratum 3, and so on.
Poll
This shows the rate at which the source is being polled, as a base-2 logarithm of the interval in seconds. Thus, a value of 6 would indicate that a measurement is being made every 64 seconds. chronyd automatically varies the polling rate in response to prevailing conditions.
LastRx
This column shows how long ago the last sample was received from the source. This is normally in seconds. The letters m, h, d or y indicate minutes, hours, days or years.
Last sample
This column shows the offset between the local clock and the source at the last measurement. The number in the square brackets shows the actual measured offset. This may be suffixed by us (indicating microseconds), ms (indicating milliseconds), or s (indicating seconds). The number to the left of the square brackets shows the original measurement, adjusted to allow for any slews applied to the local clock since. The number following the +/- indicator shows the margin of error in the measurement. Positive offsets indicate that the local clock is fast of the source.

sourcestats

The sourcestats command displays information about the drift rate and offset estimatation process for each of the sources currently being examined by chronyd.

An example report is

210 Number of sources = 1
Name/IP          NP   NR  Span       Freq      Skew      S.D./us
================================================================
abc.def.ghi      11    5   46m      -0.001       0.045        25

The columns are as follows

Name/IP
This is the name or dotted-quad IP address of the NTP server (or peer) to which the rest of the line relates.
NP
This is the number of sample points currently being retained for the server. The drift rate and current offset are estimated by performing a linear regression through these points.
NR
This is the number of runs of residuals having the same sign following the last regression. If this number starts to become too small relative to the number of samples, it indicates that a straight line is no longer a good fit to the data. If the number of runs is too low, chronyd discards older samples and re-runs the regression until the number of runs becomes acceptable.
Span
This is the interval between the oldest and newest samples. If no unit is shown the value is in seconds. In the example, the interval is 46 minutes.
Freq
This is the estimated residual frequency for the server, in parts per million. In this case, the computer's clock is estimated to be running 1 part in 10**9 slow relative to the server.
Skew
This is the estimated error bounds on Freq (again in parts per million).
Var/us
This is the estimated sample variance in microseconds.

tracking

The tracking command displays parameters about the system's clock performance. An example of the output is shown below.

Reference ID    : 1.2.3.4 (a.b.c)
Stratum         : 3
Ref time (UTC)  : Sun May 17 06:13:11 1998
System time     : 0.000000 seconds fast of NTP time
Frequency       : 331.898 ppm fast
Residual freq   : 0.004 ppm
Skew            : 0.154 ppm
Root delay      : 0.373169 seconds
Root dispersion : 0.024780 seconds

The fields are explained as follows.

Reference ID
This is the IP address, and name if available, of the server to which the computer is currently synchronised. If this is 127.127.1.1 it means the computer is not synchronised to any external source and that you have the `local' mode operating (via the local command in chronyc (see section local), or the local directive in the `/etc/chrony.conf' file (see section local)).
Stratum
The stratum indicates how many hops away from a computer with an attached reference clock we are. Such a computer is a stratum-1 computer, so the computer in the example is two hops away (i.e. a.b.c is a stratum-2 and is synchronised from a stratum-1).
Ref time
This is the time (GMT) at which the last measurement from the reference source was processed.
System time
In normal operation, chronyd never steps the system clock, because any jump in the timescale can have adverse consequences for certain application programs. Instead, any error in the system clock is corrected by slightly speeding up or slowing down the system clock until the error has been removed, and then returning to the system clock's normal speed. A consequence of this is that there will be a period when the system clock (as read by other programs using the gettimeofday() system call, or by the date command in the shell) will be different from chronyd's estimate of the current true time (which it reports to NTP clients when it is operating in server mode). The value reported on this line is the difference due to this effect. On systems such as Solaris and SunOS, chronyd has no means to adjust the fundamental rate of the system clock, so keeps the system time correct by periodically making offsets to it as though an error had been measured. The build up of these offsets will be observed in this report. On systems such as Linux where chronyd can adjust the fundamental rate of the system clock, this value will show zero unless a very recent measurement has shown the system to be error.
Frequency
The `frequency' is the rate by which the system's clock would be would be wrong if chronyd was not correcting it. It is expressed in ppm (parts per million). For example, a value of 1ppm would mean that when the system's clock thinks it has advanced 1 second, it has actually advanced by 1.000001 seconds relative to true time. As you can see in the example, the clock in the computer I developed chrony on is not a very good one - it gains about 30 seconds per day! This was the reason I started to write chrony in the first place.
Residual freq
This shows the `residual frequency' for the currently selected reference source. This reflects any difference between what the measurements from the reference source indicate the frequency should be and the frequency currently being used. The reason this is not always zero is that a smoothing procedure is applied to the frequency. Each time a measurement from the reference source is obtained and a new residual frequency computed, the estimated accuracy of this residual is compared with the estimated accuracy (see `skew' next) of the existing frequency value. A weighted average is computed for the new frequency, with weights depending on these accuracies. If the measurements from the reference source follow a consistent trend, the residual will be driven to zero over time.
Skew
This is the estimated error bound on the the frequency.
Root delay
This is the total of the network path delays to the stratum-1 computer from which the computer is ultimately synchronised. In certain extreme situations, this value can be negative. (This can arise in a symmetric peer arrangement where the computers' frequencies are not tracking each other and the network delay is very short relative to the turn-around time at each computer.)
Root dispersion
This is the total dispersion accumulated through all the computers back to the stratum-1 computer from which the computer is ultimately synchronised. Dispersion is due to system clock resolution, statistical measurement variations etc. An absolute bound on the computer's clock accuracy (assuming the stratum-1 computer is correct) is given by
clock_error <= root_dispersion + (0.5 * |root_delay|)

trimrtc

The trimrtc command is used to correct the system's real time clock (RTC) to the main system clock. It has no effect if the error between the two clocks is currently estimated at less than a second (the resolution of the RTC is only 1 second).

The command takes no arguments. It performs the following steps (if the RTC is more than 1 second away from the system clock):

  1. Remember the currently estimated gain/loss rate of the RTC and flush the previous measurements.
  2. Step the real time clock to bring it within a second of the system clock.
  3. Make several measurements to accurately determine the new offset between the RTC and the system clock (i.e. the remaining fraction of a second error)
  4. Save the RTC parameters to the RTC file (specified with the rtcfile directive in the configuration file (see section rtcfile).

The last step is done as a precaution against the computer suffering a power failure before either the daemon exits or the writertc command is issued.

chronyd will still work perfectly well both whilst operating and across machine reboots even if the trimrtc command is never used (and the RTC is allowed to drift away from true time). The trimrtc command is provided as a method by which it can be corrected, in a manner compatible with chronyd using it to maintain accurate time across machine reboots.

writertc

The writertc command writes the currently estimated error and gain/loss rate parameters for the RTC to the RTC file (specified with the rtcfile directive (see section rtcfile)). This information is also written automatically when chronyd is killed (with SIGHUP, SIGINT, SIGQUIT or SIGTERM).

Porting guide

This appendix discusses issues that have arisen in writing the system-specific parts of the existing ports. This will provide useful information for those attempting to write ports to other systems.

System driver files

The system specific parts of the software are contained in files with names like sys_linux.c.

The following functions are required in a system driver file:

  1. A function to read the current frequency
  2. A function to set the current frequency
  3. A function to slew the system time by a specified delta
  4. A function to step the system time by a specified delta
  5. A function to work out the error at a particular time between the system's clock and chronyd's estimate of real time. (This is required because some systems have to track real time by making the system time follow it in a 'sawtooth' fashion).

The frequency is the rate at which the system gains or loses time, measured relative to the system when running uncompensated.

Quirks of particular systems

These sections describe quirks in each system type that needed to be investigated to port the software to each system type.

Linux

The following quirks have been found in developing the Linux port.

  1. In order to avoid floating point arithmetic, the kernel uses shifting and adding to approximate a scaling of 100/128. This approximation implies that the frequency set via the adjtimex() system call is not the frequency that is actually obtained. The method of approximation varies between kernel versions and must be determined by examining the kernel source. An inverse factor must be included in the driver to compensate.
  2. In some kernel versions, an adjtimex() system call with the flags bits all zeroed will return the amount of offset still to be corrected. In others (e.g. the 2.0 series beyond 2.0.32), the offset must be changed in order to get the old offset returned (similar to adjtime() on other systems).

Solaris 2.5

The following quirks have been found in developing the Solaris port.

  1. The adjtime() system call with a zero argument does not cancel an adjustment that is in progress - it just reports the remaining adjustment.
  2. The settimeofday() system call only observes the seconds part of the argument - any fractional seconds part is lost. second.
  3. The kernel variable dosynctodr has to be set to zero, otherwise the system clock is periodically reset to the real-time clock.

SunOS 4.1.4

The following quirks have been found in developing the SunOS port.

  1. The adjtime() system call truncates its argument to a multiple of the system's tickadj variable. (chronyd sets that to 100, giving a 1 part in 100 slewing capability for correcting offsets.)
  2. The kernel variable dosynctodr has to be set to zero, otherwise the system clock is periodically reset to the real-time clock.


This document was generated on 11 November 1998 using the texi2html translator version 1.51a.