from The Wizard of Oz, L. Frank Baum
You need help from the monkeys.
Last update: 14-oct-09 22:23
The ntpd program is an operating system daemon that synchronises the system clock with remote NTP time servers or local reference clocks. It is a complete implementation of the Network Time Protocol (NTP) version 4, but also retains compatibility with version 3, as defined by RFC-1305, and version 1 and 2, as defined by RFC-1059 and RFC-1119, respectively. The program can operate in any of several modes, as described on the Association Management page, and with both symmetric key and public key cryptography, as described on the Authentication Options page.
The ntpd program ordinarily requires a configuration file as desccribe on the Configuration Commands and Options collection above. However a client can discover remote servers and configure them automatically. This makes it possible to deploy a fleet of workstations without specifying configuration details specific to the local environment. Further details are on the Automatic Server Discovery page.
Once the NTP software distribution has been compiled and installed and the configuration file constructed, the next step is to verify correct operation and fix any bugs that may result. Usually, the command line that starts the daemon is included in the system startup file, so it is executed only at system boot time; however, the daemon can be stopped and restarted from root at any time. Once started, the daemon will begin sending and receiving messages, as specified in the configuration file.
The ntpd program operates by exchanging messages with one or more servers at designated intervals ranging from about one minute to about 17 minutes. When started, the program requires several exchanges while the algorithms accumulate and groom the data before setting the clock. The initial delay to set the clock can be reduced using options on the Server Options page.
Most compters today incorporate a time-of-year (TOY) chip to maintain the time during periods when the power is off. When the machine is booted, the chip is used to initialize the operating system time. In case there is no TOY chip or the TOY time is more than 1000 s from the server time, ntpd assumes something must be terribly wrong and exits with a panic message to the system operator. With the -g option the clock will be initially set to the server time regardless of the chip time. However, once the clock has been set, an error greater than 1000 s will cause ntpd to exit anyway.
Under ordinary conditions, ntpd slews the clock so that the time is effectively continuous and never runs backwards. If due to extreme network congestion an error spike exceeds the step threshold, by default 128 ms, the spike is discarded. However, if the error persists for more than the stepout threshold, by default 900 s, the system clock is stepped to the correct value. In practice the need for a step has is extremely rare and almost always the result of a hardware failure. With the -x option the step threshold is increased to 600 s. Other options are available using the tinker command on the Miscellaneous Options page.
The issues should be carefully considered before using these options. The maximum slew rate possible is limited to 500 parts-per-million (PPM) by the Unix kernel. As a result, the clock can take 2000 s for each second the clock is outside the acceptable range. During this interval the clock will not be consistent with any other network clock and the system cannot be used for distributed applications that require correctly synchronized network time.
The frequency file, usually called ntp.drift, contains the latest estimate of clock frequency. If this file does not exist when ntpd is started, it enters a special mode designed to measure the particular frequency directly. The measurement takes 15 minutes, after which the frequency is set and ntpd resumes normal mode where the time and frequency are continuously adjusted. The frequency file is updated at intervals of an hour or more depending on the measured clock stability.
The ntpd program normally operates continuously while adjusting the time and frequency, but in some cases it may not be practical to run it continuously. With the -q option ntpd operates as in continous mode, but exits just after setting the clock for the first time. Most applications will probably want to specify the iburst option with the server command. With this option a volley of messages is exchanged to groom the data and set the clock in about 10 s. If nothing is heard after a few minutes, the daemon times out and exits.
NTP uses an intricate heuristic algorithm to automatically control the poll interval for maximum accuracy consistent with minimum network overhead. The algorithm measures the incidental offset and jitter to determine the best poll interval. When ntpd starts, the interval is the default minimum 64 s. Under normal conditions when the clock discipline has stabilized, the interval increases in steps to the default maximum 1024 s. In addition, should a server become unreachable after some time, the interval increases in steps to the maximum in order to reduce network overhead.
The default poll interval range is suitable for most conditions, but can be changed using options on the Server Options and Miscellaneous Options pages. However, when using maximum intervals much larger than the default, the residual clock frequency error must be small enough for the discipline loop to capture and correct. The capture range is 500 PPM with a 64-s interval decreasing by a factor of two for each interval doubling. At a 36-hr interval, for example, the capture range is only 0.24 PPM.
In scenarios where a considerable amount of data are to be downloaded or uploaded over telephone modems, timekeeping quality can be seriously degraded. This occurs because the differential delays on the two directions of transmission can be quite large. In many cases the apparent time errors are so large as to exceed the step threshold and a step correction can occur during and after the data transfer.
The huff-n'-puff filter is designed to correct the apparent time offset in these cases. It depends on knowledge of the propagation delay when no other traffic is present, such as during other than work hours. The filter remembers the minimum delay over the most recent interval measured usually in hours. Under conditions of severe delay, the filter corrects the apparent offset using the sign of the offset and the difference between the apparent delay and minimum delay. The name of the filter reflects the negative (huff) and positive (puff) correction, which depends on the sign of the offset. The filter is activated by the tinker huffpuff command, as described in the Miscellaneous Options page.
As provided by international agreement, an extra second is sometimes inserted in Coordinated Universal Time (UTC) at the end of a selected month, usually June or December. The National Institutes of Standards and Technology (NIST) provides an historic leapseconds file at time.nist.gov for retrieval via FTP. When this file, usually called ntp-leapseconds.list, is copied and installed in a directory. The leapfile configuration command specifies the path to this file. At startup, ntpd reads it and initializes three leapsecond values: the NTP seconds at the next leap event, the offset of UTC relative to International Atomic Time (TAI) after the leap and the NTP seconds when the leapseconds file expires and should be retrieved again.
If a host does not have the leapsecond values, they can be obtained over the net using the Autokey security protocol. Ordinarily, the leapseconds file is installed on the primary servers and the values flow from them via secondary servers to the clients. When multiple servers are involved, the values with the latest expiration time are used.
If the latest leap is in the past, nothing further is done other than to install the TAI offset. If the leap is in the future less than 28 days, the leap warning bits are set. If in the future less than 23 hours, the kernel is armed to insert one second at the end of the current day. If the kernel is enabled, the leap is done automatically at that time; otherwise, the clock is effectively stopped for one second at the leap. Additional details are in the The NTP Timescale and Leap Seconds white paper
If none of the above provisions are available, dsependent servers and clients tally the leap warning bits of surviving servers and reference clocks. When a majority of the survivors show warning, a leap is programmed at the end of the current month. During the month and day of insertion, they operate as above. In this way the leap is is propagated at all dependent servers and clients.
A new experimental feature called interleaved modes can be used in NTP symmetric or broadcast modes. It is designed to improve accuracy by avoiding kernel latency and queueing delay, as described on the NTP Interleaved Modes page. It is activated by the xleave option with the peer or broadcast configuration commands. The NTP protocol automatically reconfigures in normal or interleaved mode as required. Ordinary broadcast clients can use the same servers as interleaved clients at the same time. Further details are in the white paper NTP Interleaved On-Wire Protocol and the briefing Interleaved Synchronization Protocols for LANs and Space Data Links.
If ntpd, is configured with NetInfo support, it will attempt to read its configuration from the NetInfo service if the default ntp.conf file cannot be read and no file is specified by the -c option.
In contexts where a host name is expected, a -4 qualifier preceding the host name forces DNS resolution to the IPv4 namespace, while a -6 qualifier forces DNS resolution to the IPv6 namespace.
Various internal ntpd variables can be displayed and configuration options altered while the ntpd is running using the ntpq and ntpdc utility programs.
When ntpd starts it looks at the value of umask, and if zero ntpd will set the umask to 022.
Ordinarily, ntpd reads the ntp.conf configuration file at startup in order to determine the synchronization sources and operating modes. It is also possible to specify a working, although limited, configuration entirely on the command line, obviating the need for a configuration file. This may be particularly useful when the local host is to be configured as a broadcast client, with servers determined by listening to broadcasts at run time.
Usually, the configuration file is installed as/etc/ntp.conf, but could be installed elsewhere (see the -c conffile command line option). The file format is similar to other Unix configuration files - comments begin with a # character and extend to the end of the line; blank lines are ignored.
Configuration commands consist of an initial command keyword followed by a list of option keywords separated by whitespace. Commands may not be continued over multiple lines. Options may be host names, host addresses written in numeric, dotted-quad form, integers, floating point numbers (when specifying times in seconds) and text strings. Optional arguments are delimited by [ ] in the options pages, while alternatives are separated by |. The notation [ ... ] means an optional, indefinite repetition of the last item before the [ ... ].
| File | Default | Option | Command |
| configuration file | /etc/ntp.conf | -c | none |
| frequency file | none | -f | driftfile |
| leapseconds file | none | leapfile | |
| process ID file | none | -p | pidfile |
| log file | system log | -l | logfile |
| include file | none | none | includefile |
| statistics path | /var/NTP | -s | statsdir |
| keys path | /usr/local/etc | -k | keysdir |