Hacker News Reader

Comment by burfog | Hacker News Reader

You can still have problems.

It may be determined that the computer's clock got too far ahead. For example, it booted and you cared about time, but NTP hadn't yet made corrections. Suddenly the time runs backwards.

Leap seconds may get interesting too, especially if you have to predict ahead or if the OS isn't updated often enough. (there is a 6-month warning) If you want to call leap seconds an issue for humans, then you aren't using UTC at all. You're using TAI. Software interfaces often ignore the distinction between UTC and TAI, and even between UTC and UT1, preferring to pretend these issues don't exist. POSIX is in conflict with international timekeeping, effectively requiring that there are zero leap seconds.

View on HN

> Suddenly the time runs backwards

It is my understanding that the way NTP deamons work is that time is never adjusted backward. Instead, the ticks are "slowed down" on the local machine until it is in sync with the NTP time. However, if the difference is too great then I think NTP deamons might refuse to correct the time all together. So then, if my understanding is correct, your machine is "stuck in the future". But it will never make a jump backwards because of NTP.

However, I am not familiar with the intricate details of NTP so do take this with a grain of salt.

My understanding (i am also not an expert!) is that common NTP implementations will do the slowing down ("slewing") to correct small errors, but will just change the time ("stepping") to correct large errors. It will even do this if that means time going backwards.

Subject to configuration, of course. man ntpd [1] says:

> Sometimes, in particular when ntpd is first started, the error might exceed 128 ms. This may on occasion cause the clock to be set backwards if the local clock time is more than 128 s in the future relative to the server. In some applications, this behavior may be unacceptable. If the -x option is included on the command line, the clock will never be stepped and only slew corrections will be used.

[1] https://linux.die.net/man/8/ntpd

timesyncd also does this for "large offsets", but what "large" is is neither configurable nor documented, but in the source:

    /*
     * Maximum delta in seconds which the system clock is gradually adjusted
     * (slewed) to approach the network time. Deltas larger that this are set by
     * letting the system time jump. The kernel's limit for adjtime is 0.5s.
     */
    #define NTP_MAX_ADJUST                  0.4

> neither configurable nor documented, but in the source:

Build it yourself and pass -DNTP_MAX_ADJUST=XX

Leap seconds are only a problem for wall clocks (localization) and deciding whether to call something utc or tai.

The ntp case is contrived. Either you care and wait until ntp has connected to do your stuff. Or you care and don't let ntp rewind and instead smear. Or you don't care and deal with the consequences.

Leap seconds affect Unix time too, because leap seconds are excluded from "number of seconds since Unix epoch". You can't measure the length of time intervals spanning leap seconds with simple subtraction.

Are they truly? There must be a good reason but I can't fathom what it could be

Because otherwise you would get annoying off-by-27 errors whenever you did time() % 60.

https://en.wikipedia.org/wiki/Leap_second#Binary_representat...

That good reason is that there is no general "formula" for leap seconds, unlike for leap years, they have to be looked up. So you can't do "offline" date calculations if they included leap seconds.

I think that UNIX time stamps are generally a very good approximation, and if you are comparing long enough time intervals for the error to get over one second, and/or that error to matter, you are doing something wrong anyway.

For exact time interval measurements that you have to get exactly right, don't use UNIX time stamps.

The lookup database is incredibly small and updated very infrequently. Certainly smaller and less frequent than tzinfo which is also included in the kernel

AFAICT leap seconds are announced about 6 months in advance (January/June). How would you process times that are to occur after a leap second that may or may not be?

You could store something as '2022-02-02 02:22:22 UTC', but try storing it as seconds since epoch, if you don't know if there will be leap seconds. Not to mention software that may be years or decades old.

> Leap seconds affect Unix time too, because leap seconds are excluded from "number of seconds since Unix epoch".

They are excluded from "unix timestamps". They would still be part of the number of seconds since Unix epoch.

What does that mean and how would that work?

AFAIK, unix time skips a beat or repeats itself to remain in alignment with UTC, and just keeps chugging along.

"Number of seconds since the unix epoch" describes a physical measurement (that can be related directly to TAI) that would not be affected by leap seconds. You could store that number to allow you to recover TAI timestamps (I'm not aware of any OS that does this, but there's no reason it wouldn't be possible).

Unix time does indeed repeat itself so as to remain in alignment with UTC. But unix time is not the number of seconds since the unix epoch.

Thank you, I think I understand, but disagree on terminology.

Leap seconds are not excluded from number of seconds since Unix epoch, interpreted on a physical, TAI, UTC, typical local time scales.

However, they are indeed excluded from "number of seconds since Unix epoch", when interpreted in unix time. In unix time, those seconds simply don't exists (never happened) and the events of those seconds are smooshed sometime. Unix time representation forms a (quirky) timescale.

I'll quote a few excerpts from 'date' man page of linux, openbsd and for gettimeofday:

Convert seconds since the epoch (1970-01-01 UTC) to a date

Print out (in specified format) the date and time represented by seconds from the Epoch.

The time is expressed in seconds and microseconds since midnight (0 hour), January 1, 1970

Hmm. The first definition I saw of "unix timestamp" explicitly defined it as "86400 * number of days since 1970-01-01 + number of seconds since midnight", which I thought was clever and clear, but I can't find it now. Having looked up the POSIX standard at https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1... it seems you're right and I'm wrong; they define "Seconds Since the Epoch" as "A value that approximates the number of seconds that have elapsed since the Epoch...". In my view this is an extremely poor choice of terminology, because "seconds since the epoch" should have its plain English meaning of the number of SI seconds that have passed since that event; defining some "second" that is almost, but not quite, an SI second seems like a recipe for disaster.

Reply on news.ycombinator.com