Skip to Content.
Sympa Menu

perfsonar-user - Re: [perfsonar-user] time too long

Subject: perfSONAR User Q&A and Other Discussion

List archive

Re: [perfsonar-user] time too long


Chronological Thread 
  • From:
  • To: Victor Olifer <>,"Karch, Roland" <>,陳怡安 <>
  • Cc: "" <>
  • Subject: Re: [perfsonar-user] time too long
  • Date: Fri, 30 May 2014 05:10:59 -1000

Hi Victor,

What is the frequency offset on the physical host?

As reported in loopstats, or 'ntptime' ?

-Alan

On May 30, 2014 3:51:11 AM HST, Victor Olifer <> wrote:
Hi Roland, Angela, all,

I’m experiencing something similar to what Angela described with one of my servers running owamp and my investigation of the NTP statistics files such as loopstats and peerstats shows that it is likely not a NTP problem.

In more detail:


* I’m running ps toolkit with owamp 3.4 on a virtual machine (VMWare) and this owamp instance produces plausible measurements with a good precision (tested against a CARNET host)
* I’m also running owamp 3.4 installed as a RPM on a physical host under CentOS 6.4; this owamp instances produces obviously wrong measurements, for example, against the ps toolkit host the results oscillate in +/- 800 ms range being in 1 hop from each other.
* NTP configs of both hosts are identical – they are synced with the same set of Janet production NTP servers; owampd.conf files are also identical.
* Analysing NTP statistics, loopstats in particular, shows the similar and good precision of both local clocks: the clock offset on both hosts doesn’t exceed +/- 0.0006 s on both hosts, the rpm owamp host mostly shows even better precision than ps toolkit which is probably understandable taking into account virtual environment of the latter.

So, I see no reason to blame the bad NTP sync and see no way forward to troubleshoot the RPM instance of owamp. Any ideas would be much appreciated.

All the best,
Victor Olifer
Technical Specialist
Janet

From: "Karch, Roland" <<mailto:>>
Date: Tuesday, 27 May 2014 09:05
To: 陳怡安 <<mailto:>>
Cc: "<mailto:>" <<mailto:>>
Subject: RE: [perfsonar-user] time too long

Hi Angela,

for one way delay metrics, it is important to have a well synchronized clock, especially when measuring between two end points very close to each other. The following might be resources worth to investigate:

RFC 2679<https://tools.ietf.org/html/rfc2679> chapter 3.5
OWAMP cookbook<http://e2epi.internet2.edu/npw/binder-docs/owamp-cookbook.pdf> search for “NTP”

Hope that helps!

With best wishes,
Roland
--
Roland Karch, DFN-Labor
Friedrich-Alexander-Universität Erlangen-Nürnberg
Regionales RechenZentrum Erlangen (RRZE)
Martensstraße 1, 91058 Erlangen, Germany
Tel. +49 9131 85-27806, -28800, Fax: +49 9131 302941
mailto:
http://www-win.rrze.uni-erlangen.de/



From: <mailto:> [mailto:] On Behalf Of ???
Sent: Sunday, May 25, 2014 3:12 AM
To: <mailto:>
Subject: [perfsonar-user] time too long

Dear all,

I install owampd on 10.21.10.120.
I uesd owping, from 10.21.11.171 to 10.21.10.120.
But it is too long.
I used ping, only need less than 1 ms.
Why the time of owping is too long?
Thanks in advance.

[root@angela ~]# owping 10.21.10.120
Approximately 13.1 seconds until results available

--- owping statistics from [ip247.puli11-21-10.ncnu.edu.tw<http://ip247.puli11-21-10.ncnu.edu.tw>]:8942 to [10.21.10.120]:49546 ---
SID: 0a150a78d72bbff184726d0437db9567
first: 2014-05-25T09:03:45.593
last: 2014-05-25T09:03:55.971
100 sent, 0 lost (0.000%), 0 duplicates
one-way delay min/median/max = 425/1.09e+03/1.11e+03 ms, (err=0.969 ms)
one-way jitter = 16.3 ms (P95-P50)
TTL not reported
no reordering


--- owping statistics from [10.21.10.120]:37393 to [ip247.puli11-21-10.ncnu.edu.tw<http://ip247.puli11-21-10.ncnu.edu.tw>]:8874 ---
SID: 0a150bf7d72bbff06ace57103357a55e
first: 2014-05-25T09:03:45.606
last: 2014-05-25T09:03:55.671
100 sent, 0 lost (0.000%), 0 duplicates
one-way delay min/median/max = -1.11e+03/-1.09e+03/-425 ms, (err=0.969 ms)
one-way jitter = 667 ms (P95-P50)
TTL not reported
no reordering

With best regards,
Angela
May 25



NCNU Cloudmail V 0.2 - 請尊重智慧財產權! 請勿開啟不明淶源郵件!

Janet(UK) is a trading name of Jisc Collections and Janet Limited, a
not-for-profit company which is registered in England under No. 2881024
and whose Registered Office is at Lumen House, Library Avenue,
Harwell Oxford, Didcot, Oxfordshire. OX11 0SG. VAT No. 614944238


--
Sent from my Android device with K-9 Mail. Please excuse my brevity.


Archive powered by MHonArc 2.6.16.

Top of Page