Didier Kryn <kryn@???> writes:

[...]

> The problem is the way the ntp service is started in Debian involves
> invoking ntpdate prior to ntpd.

While remember this, I've checked installations of Debian 7, 6 and 5 and
none of them did this (anymore). The ntpd manpage also says

    -q     Exit  the  ntpd  just  after the first time the clock is set.
     This behavior mimics that of the ntpdate program, which is to be
     retired.

> This can't work with the method of starting all services in the same
> time, in general, not specifically by systemd.

It absolutely can: This depends on how often the stub resolver in the C
library (or the systemd stub resolver) retries a query before giving
up. If this interval made up of this is longer than any actually
experienced 'network failure' (this is - of course - impossible as 'the
internet' is composed of many computers and at least some of them are
unusable at any given time), it will "work", otherwise, a user might
become aware of the network failure. But this can (cf remark about 'many
computers') really happen at any given time while a system is up and not
only during startup.

Fittingly, my 'internet connection' broke down while writing
this. Turning the router on before anything else and waiting until it
signals 'connected' is no cure for that.