perfSONAR User Q&A and Other Discussion

[Larry Blunk <>]

On 11/05/2015 10:42 AM, Brian Candler wrote:
> On 05/11/2015 15:05, Brian Candler wrote:
>> This then begs the question: why the difference between the perfsonar
>> 3.4 and perfsonar 3.5 boxes?
> It seems that perfsonar 3.5 records traceroute from (remote host) to
> (local host), while perfsonar 3.4 records traceroute from (local host)
> to (remote host)
> 
> At least, that's what I see when I go to the psTracerouteViewer page on
> those two nodes.
> 
> I don't know if that's an intentional change in behaviour between 3.4
> and 3.5, or whether it's one of those things where perfsonar is
> non-deterministic as to what is "source" and what is "destination".
> 
> However, looking on the 3.5 box I can scan the measurement archive from
> before it upgraded to 3.4:
> 
> 
> 
> If I do this, I see there is traceroute information for (perfsonar node)
> -> (remote Mac) only up until Wed Sep 30 00:56:26 2015, which I think is
> when the upgrade took place.
> 
> There is IPv4 traceroute information for (remote Mac) -> (perfsonar
> node) starting Wed Sep 30 01:06:27 2015
> 
> And there is IPv6 traceroute information for (remote Mac) -> (perfsonar
> node) from when I fixed the traceroute6 parsing just now.
> 
> So it does look to me like perfsonar has changed which way it runs
> traceroutes.
> 
> Regards,
> 
> Brian.
> 


 I noticed that the ping tests also reversed direction when
going from 3.4 to 3.5.

      Speaking of which, it'd be nice if there was an
option to display the ping roundtrip times instead
of the owamp one-way latency times on the dashboard.   I don't find
the one-way latency times to be particularly useful due to
system clock skew and variation.  Tried various NTP sources and
tunings and it seems like it's impossible to get better
than 1 millisecond of accuracy.

 -Larry Blunk