| Re: [chrony-dev] Frequency transfer in NTP |
[ Thread Index |
Date Index
| More chrony.tuxfamily.org/chrony-dev Archives
]
- To: chrony-dev@xxxxxxxxxxxxxxxxxxxx
- Subject: Re: [chrony-dev] Frequency transfer in NTP
- From: Miroslav Lichvar <mlichvar@xxxxxxxxxx>
- Date: Tue, 2 Feb 2021 13:46:51 +0100
- Authentication-results: relay.mimecast.com; auth=pass smtp.auth=CUSA124A263 smtp.mailfrom=mlichvar@xxxxxxxxxx
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1612270018; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:mime-version:mime-version:content-type:content-type: in-reply-to:in-reply-to:references:references; bh=g262DUl/ee+N9gEqQqW4sKK3RyRxXk/t53oexPBfUzY=; b=ft0d40ATWHB6Hcj525cxlLGIEQZSHse+oeN05TGVfiVMuGu0LzikPPzCMdUmoc5pPD6bXb dgYAd++8jyLHYXb1/Cdqt8UE2X2BDL3NrPtcyCAupaFQpOoiqIFKiZWCdLB0X1I967y1ih 4UzRiD3KRIteBbrXc27A/e9uQKOIISQ=
On Thu, Jan 28, 2021 at 11:31:47AM -0800, Bill Unruh wrote:
> Not sure how, since there is no frequency standard. Frequency is seconds per
> second. But for this not to be a tautalogy, what are the two seconds in that
> refering to? Anyway, yes it is certainly interesting. It is also slightly
> worrying that chrony has that oscillation as one goes up the stratum level.
> Can it go unstable? Certainly there is no evidence at stratum 4, but could
> stratum 10 say go unstable?
I ran a few tests to see how it looks at stratum 10.
The oscillation can get very bad, but I don't think it can break
completely, e.g. get to a state from which it cannot recover.
https://fedorapeople.org/~mlichvar/ntpresponse/chrony44.png
It seems the impact of the step can be reduced if the polling interval
is allowed to change (minpoll < maxpoll). As the step causes the
client to drop to a shorter interval, it has a chance to get over the
overshoot before it's polled by its own client. The responses in the
chain can be staggered and so they don't pile on top of each other as
much, but it oscillates for a longer time.
https://fedorapeople.org/~mlichvar/ntpresponse/chrony46.png
I have also done some testing over a longer interval, with no
artificial steps, to show the improvement with enabled frequency
transfer. This is a table with the RMS offset in microseconds measured
at all strata. At stratum 10, there is an improvement of a factor of 3.
stratum no-freq freq
2 11.7 12.0
3 18.1 15.4
4 24.7 17.2
5 30.1 20.7
6 40.1 22.5
7 50.4 25.9
8 64.2 26.9
9 77.1 29.0
10 91.7 31.4
--
Miroslav Lichvar
--
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.