[ Thread Index |
Date Index
| More chrony.tuxfamily.org/chrony-users Archives
]
Frank Huang
On Thu, Aug 22, 2019 at 1:41 AM Miroslav Lichvar <mlichvar@xxxxxxxxxx>
wrote:
> On Wed, Aug 21, 2019 at 12:11:15PM -0700, chuang213 wrote:
> > the local clock is the fallback in case none of the source is not
> available
> > so that the server could continue to provide time service to local
> network.
>
> For that the "local stratum" line is enough. Specifying 127.127.1.0 as
> a server will cause chronyd to poll itself over NTP and create a
> synchronization loop when selected (until stratum reaches 16). If it
> becomes unsynchronized for any reason, this source will be
> unsychronized too. You should remove that source. It's only causing
> problems.
>
> In ntpd it is different. Addresses in the 127.127.0.0/16 range specify
> refclocks. There are no NTP packets exchanged. 127.127.1.0 is a LOCAL
> driver, which simulates a reference clock based on the local clock.
> It's a hack that has problems and it was deprecated when ntpd got its
> own local reference mode (orphan mode).
>
> > > 2019-08-20 23:41:52 PPS 9.088e-06 -9.657e-06 4.871e-06
> > > -2.323e-07 6.526e-07 1.0e-01 7 0 4 0.00
> > > 2019-08-20 23:42:03 127.127.1.0 5.293e-06 1.146e-05 3.293e-06
> > > -1.253e-08 1.335e-07 3.7e-01 7 0 5 -0.50
>
> Here it looks like it considered itself to be significantly more
> stable than the PPS refclock. That's likely why it briefly reselected.
> We would need to see the debug output to confirm that.
>
> --
> Miroslav Lichvar
>
> --
> To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
> with "unsubscribe" in the subject.
> For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
> with "help" in the subject.
> Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.
>
>
--00000000000032758c0590b64bb4
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div dir=3D"ltr"><div>Thanks for the info. I'll remove the line 'se=
rver 127.127.1.0'.</div><div><br></div><div>From the first email of thi=
s thread, I think it would keep the PPS as the clock source for a long peri=
od, rather than switching to 127.127.1.0 briefly even PPS is reachable.</di=
v><div><br></div><div>Frank Huang<br></div></div><br><div class=3D"gmail_qu=
ote"><div dir=3D"ltr" class=3D"gmail_attr">On Thu, Aug 22, 2019 at 1:41 AM =
Miroslav Lichvar <<a href=3D"mailto:mlichvar@xxxxxxxxxx";>mlichvar@redhat=
..com</a>> wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"mar=
gin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1=
ex">On Wed, Aug 21, 2019 at 12:11:15PM -0700, chuang213 wrote:<br>
> the local clock is the fallback in case none of the source is not avai=
lable<br>
> so that the server could continue to provide time service to local net=
work.<br>
<br>
For that the "local stratum" line is enough. Specifying 127.127.1=
..0 as<br>
a server will cause chronyd to poll itself over NTP and create a<br>
synchronization loop when selected (until stratum reaches 16). If it<br>
becomes unsynchronized for any reason, this source will be<br>
unsychronized too. You should remove that source. It's only causing<br>
problems.<br>
<br>
In ntpd it is different. Addresses in the <a href=3D"http://127.127.0.0/16"=
rel=3D"noreferrer" target=3D"_blank">127.127.0.0/16</a> range specify<br>
refclocks. There are no NTP packets exchanged. 127.127.1.0 is a LOCAL<br>
driver, which simulates a reference clock based on the local clock.<br>
It's a hack that has problems and it was deprecated when ntpd got its<b=
r>
own local reference mode (orphan mode).<br>
<br>
> > 2019-08-20 23:41:52 PPS=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 9.088e-06 -9.657e-06=C2=A0 4.871e-06<br>
> > -2.323e-07=C2=A0 6.526e-07 1.0e-01=C2=A0 =C2=A07=C2=A0 =C2=A00=C2=
=A0 =C2=A04=C2=A0 0.00<br>
> > 2019-08-20 23:42:03 127.127.1.0=C2=A0 =C2=A0 =C2=A0 5.293e-06=C2=
=A0 1.146e-05=C2=A0 3.293e-06<br>
> > -1.253e-08=C2=A0 1.335e-07 3.7e-01=C2=A0 =C2=A07=C2=A0 =C2=A00=C2=
=A0 =C2=A05 -0.50<br>
<br>
Here it looks like it considered itself to be significantly more<br>
stable than the PPS refclock. That's likely why it briefly reselected.<=
br>
We would need to see the debug output to confirm that.<br>
<br>
-- <br>
Miroslav Lichvar<br>
<br>
-- <br>
To unsubscribe email <a href=3D"mailto:chrony-users-request@xxxxxxxxxxxxxxx=
y.org" target=3D"_blank">chrony-users-request@xxxxxxxxxxxxxxxxxxxx</a> <br>
with "unsubscribe" in the subject.<br>
For help email <a href=3D"mailto:chrony-users-request@xxxxxxxxxxxxxxxxxxxx"=
target=3D"_blank">chrony-users-request@xxxxxxxxxxxxxxxxxxxx</a> <br>
with "help" in the subject.<br>
Trouble?=C2=A0 Email <a href=3D"mailto:listmaster@xxxxxxxxxxxxxxxxxxxx"; tar=
get=3D"_blank">listmaster@xxxxxxxxxxxxxxxxxxxx</a>.<br>
<br>
</blockquote></div>
--00000000000032758c0590b64bb4--
--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.