On Wed, 31 Jul 2019 18:48:18 +0300, Tauno Voipio wrote:

> On 31.7.19 15:04, Martin Gregorie wrote:
>> On Sun, 28 Jul 2019 17:18:29 +0000, A. Dumas wrote:
>>
>>> Martin Gregorie  wrote:
>>>> On Sun, 28 Jul 2019 05:19:16 +0000, A. Dumas wrote:
>>>>
>>>>> This is unusable information. Raspbian doesn't use ntp/ntpd since
>>>>> Debian switched to systemd, it's not even installed. Details here,
>>>>> for instance: https://wiki.archlinux.org/index.php/systemd-timesyncd
>>>>> It should work out of the box: if the local dhcp server provides
>>>>> (an) ntp server address(es), timesyncd will use it. Check which
>>>>> servers are used via "timedatectl show-timesync --all"
>>>>
>>>> Err, no.
>>>>
>>>> ntpd is included in Buster, but not as a separate package: according
>>>> to "apt search ntpd" its part of collectd-core/stable 5.8.1-1.3 armhf
>>>> The associated packages ntp-doc and ntpdate are also part of Buster.
>>>>
>>>> You're right that it wasn't part of wheezy, jessie or stretch but the
>>>> Buster insitu upgrade installed it when I did that the other week. It
>>>> was trivia to customise ntpd on my RPi by copying ntp.conf from this
>>>> laptop to my RPi.
>>>
>>> Your upgrade was weird before, too. A fresh install of Raspbian Buster
>>> *does not* have ntp(d). It is in the repo, obviously.
>>
>> As I said previously, I followed the directions for doing an in-situ
>> upgrade from Stretch to Buster exactly as specified on the raspbian.org
>> website and, as I also said, this pulled in ntpd without asking me if I
>> wanted it.
>>
>> So, obviously something is screwed up on the Raspbian website and/or
>> repositories if an in-situ upgrade gives a different result to a from-
>> scratch install of Buster. There are clues that something is wrong
>> since,
>> although it included ntpd in the in-situ upgrade, it did not provide an
>> ntpd.service definition so, although ntpd gets started at boot time,
>> there's no obvious way to start,stop or interrogate its status.
>
>
> It seems to me that your old installation (Stretch?) had a configured
> ntpd running. The in-situ upgrade attempts to do its best to preserve
> the functionality of the installation.
>
Nope - just checked. The only nonstandard packages under stretch were
installed from this script::

====================================================
#/bin/bash
#
# Additional packages
# ===================
#
apt-get install anacron
apt-get install cvs lynx rsync tightvncserver
apt-get install libncurses-dev ncurses-doc
apt-get install locate
apt-get install postfix
====================================================

which got created to add a few packages to my original wheezy install and
then added to later. Its under CVS (external repo on a different box
and the only changes I made to it were to replace the originally used
'at' and 'atd' packages with the 'anacron' package. That was the most
recent change and happened on 22Nov17

> Have you condsidered using ntpq?
>
No - I'm happy with ntpd - been using it for years onother Linux boxes.

> For start / stop / disable, have a look at /etc/init.d.
>
There's an 'ntp' sysvinit script there and a ntp.service file exists,
which I didn't notice earlier, probably because I was looking for
ntpd.service.

Anyway, mystery solved - thanks for making me look a little deeper.


--
Martin    | martin at
Gregorie  | gregorie dot org

--- SoupGate-Win32 v1.05