Comment 4 for bug 1593907

Revision history for this message
Christian Ehrhardt  (paelzer) wrote : Re: ntpdate startup routine prevents ntp service from launching up on Ubuntu 16.04 server on system boot; manually starting ntp service works: [FIX in DESCRIPTION], just need to apply it and release a new version

Later to be combined in SRU Template Reasoning - snippest updated from my old commit log to be readable here in the bug:

Cases of the codepath (Ubuntu Delta) that we intent to remove:

    Case 1 - bug that it meant to fix initially:
     - ntpd comes up and can not find peers to associate with
     - an interface comes up
     - stop ntpd
     - sync time once with ntpdate-debian
     - (re-)start ntpd (now able to find its peers, but it will throw away what
       ntpdate just has set, so not worth)

    Case 2 - more common case:
     - ntpd comes up normally
     - the admin brings interfaces up/down rather often
     - ntp is restarted very very often due to this for no reason
     => That behavior is close to a bug on its own, which is the reason for
    bugs like https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=823533

    Case 3 - still more common
     - ntpd is installed but disabled (for whatever reason)
     - all is fine on any ifup/down
     - one installs ntpdate, maybe even without realizing due to a depends
     - now on the next ifup ntpd (or openntpd) will be started

By dropping this section we fix it for Case 2 and Case 3.
Users of Case 1 still have no impact - their call to ntpdate will refuse to change things as there is a running ntp.
Like:
$ /usr/sbin/ntpdate-debian
  20 Jun 08:02:32 ntpdate[519]: the NTP socket is in use, exiting
And as outlined in Case 1 the running ntp would resync over it anyway.
If there was nothing running at all - well then the sync works and things are fine still.