Re: [Ntp] Finding leap-seconds.list

[Warner Losh <[e-mail]>]

On Fri, Nov 9, 2018 at 3:30 PM Jared Mauch <[e-mail]> wrote:

> There are the secondary issues of what NTP implementations use it.  I
> would like to know if it’s just FreeBSD users that fetch it and the
> frequency, etc.
>

FreeBSD users fetch it, by default, twice a year with a random smear around
a month before it expires. There's around ~100,000 FreeBSD systems
installed that fetch this, give or take. The project has no hard numbers,
and I've seen estimates as high as 1 million for the number of machines
running some flavor of FreeBSD.

I have no idea if other users fetch it as well...


> If it’s a relic of a single implementation (whatever form/type it is),
> perhaps it’s best for them to maintain the document and update it at their
> site.  It’s also perhaps too late, but if this file is so important, is it
> described in the BCP for operating NTP, or is this yet another esoteric NTP
> thing that’s not well documented so ends up being inconsistently used?
>

It's the best source of this file in the internet. That's why FreeBSD is
using it. I won't argue that it may not be the best documented thing in
NTP...


> I would be interested in seeing an implementation report for who uses it.
> (I would also perhaps argue that freebsd-update is not really a consumer of
> it).
>

freebsd-update isn't a consumer of it, nor has that been claimed. There's a
periodic script that runs a couple of times a year that fetches the file
independently of freebsd-update. It's hard to argue that FreeBSD users
doesn't use and/or fetch this file since it is enabled by default in a part
of the system that's not often changed from the defaults.

The FreeBSD project is looking for ways to transition away from doing this,
but at the moment it would create a big problem were the file to suddenly
disappear.

Warner