It's interesting you mention NTT. My ISP uses NTT for transatlantic
connectivity too (not VM). Also UK based.

I have two time servers in the ntp pool. One is in a datacenter in France.
It syncs to several S1/S2 servers and to my rPi based S1 via GPS/PPS, which
is the other server.

The two servers are:

France: http://www.pool.ntp.org/scores/149.202.2.105 (+/- 3ms).
UK Stratum 1: http://www.pool.ntp.org/scores/82.70.138.66 (+10/-70ms)

The UK one has NEVER had a good response from the monitor, in fact it's
only recently it's become this good, usually having anywhere between
+100/-1000ms on the graph). You might think maybe it's my home connection,
or the local LAN or the raspberry pi or my GPS setup.. Well, I thought that
too. So, I setup my server in France (which has a solid +/-3ms almost all
the time on the monitoring station) to monitor my home ntp.

The results are here:

+---------------------+---------------------+-----------+-----------+-----------+-----------+-----------+-----------+
| minDate             | maxDate             | minOffset | maxOffset |
avgOffset | minJitter | maxJitter | avjJitter |
+---------------------+---------------------+-----------+-----------+-----------+-----------+-----------+-----------+
| 2017-01-20 10:40:30 | 2017-02-21 01:40:01 |   -99.906 |     4.186 |
0.0892094 |     0.044 |    99.916 | 1.0873850 |
+---------------------+---------------------+-----------+-----------+-----------+-----------+-----------+-----------+

This min offset/max jitter are probably a one off. What this shows is that
the average ntp offset is 80microseconds. With an average jitter of 1ms.
So, from that I would say that both the ntp servers are keeping good time.

So, it's clearly something between this network and the monitor station.
What I see on traceroutes is my home connection goes via NTT whereas the
datacenter goes via a different route.

At any rate, last week it was really terrible for me too. But since
Saturday it's improved to the +10/-70ms I have now.


On 20 February 2017 at 21:54, Ron Hahn <[email protected]> wrote:

> Dan,
>
> I am seeing approx. 170ms times to ntplax7.ntppool.net (monitoring server)
> on a traceroute.
>
> At hop 7 it enters ntt.net (NTT America) in NYC where almost 100ms is
> added
> to the transit time.  London time is 24ms, NYC time is 120ms so it appears
> there is high latency on the transatlantic link.
>
> My transit time to a Raleigh datacenter server is 128ms over the same
> transatlantic circuit.
>
> So it seems Virgin Media upstream provider NTT America is the possible
> cause of these problems.  I will open a ticket in the morning about this
>
> Thank you,
>
> Ron
>
>
>
> On Mon, Feb 20, 2017 at 9:23 PM, Dan Geist <[email protected]> wrote:
>
> > The latency tracing to your IP from my ISP is pretty dramatic. I'm at 17
> > ms from Atlanta, GA to Ashburn near D.C. From that point, there's a jump
> to
> > around 143 ms.
> >
> > I'd suspect an upstream network issue that started around the same time
> as
> > your stats went down the tubes. The last 4 hops between that point and
> the
> > server are only a couple of ms.
> >
> > Dan
> >
> > ----- Original Message -----
> > > From: "Ron Hahn" <[email protected]>
> > > To: "pool" <[email protected]>
> > > Sent: Monday, February 20, 2017 2:03:00 PM
> > > Subject: [Pool] Why is my server score suddenly so poor?
> >
> > > Colleagues,
> > >
> > > Looking for some help here as I have spent a lot of time looking at
> this
> > > without any improvements. :-(
> > >
> > > Since early Friday, 17th the pool monitoring server was reporting that
> my
> > > pool server:
> > >
> > > http://www.pool.ntp.org/scores/89.101.218.6
> > >
> > > in increasingly negative scores - anywhere between -20 and -60 range.
> > All
> > > over up and down. :-(
> > >
> > > So thinking that something has gone horrible wrong on my side I check:
> > >
> > > - fibre interface (and reboot)
> > > - switch (also reboot, no changes in some months)
> > > - firewall config (logs and rules)
> > > - LeoNTP appliance (reboot and also firmware upgrade as well)
> > > - swap of stratum 1 source
> > > - WAN utilisation (20-30%)
> > >
> > > I am finding nothing to explain why such low scores all of a sudden.
> > >
> > > I compared the LeoNTP box to my other stratum 1 sources and they are
> > within
> > > <1ms of each other - so I think time is good and not false ticking.
> > >
> > > I also wrote a script on my Raleigh (US) server to run ntpdate against
> > this
> > > every 5 minutes overnight.  Only 4 timeouts in the entire night and
> some
> > > time this morning.
> > >
> > > Can anyone offer suggestions as to what to check next as this causing
> me
> > a
> > > big head ache now.
> > >
> > > Thank you for the help,
> > >
> > > Ron
> > > _______________________________________________
> > > pool mailing list
> > > [email protected]
> > > http://lists.ntp.org/listinfo/pool
> >
> > --
> > Dan Geist dan(@)polter.net
> >
> >
> _______________________________________________
> pool mailing list
> [email protected]
> http://lists.ntp.org/listinfo/pool
>
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool

Reply via email to