[Nfd-dev] Extremely high RTT on raspberry pi with the clean version

[Davide Pesavento]

On Tue, Jan 2, 2018 at 1:22 PM, Anthony Dowling <dowlinah at clarkson.edu> wrote:
> Hello all,
>
> I am working with NFD on the raspberry pi 3 model B system. Lately, however,
> I have had a very strange issue. While using ndnping over 1 hop in an adhoc
> environment over ethernet multicast faces, the round trip time can be
> upwards of 10 minutes. I was working with a modified version when I
> encountered this issue, however, after reinstalling the clean version from
> github, the problem persisted. So, I used the trace level logging to see
> what could be going on, but I am unable to find the issue. Does anyone have
> an idea what could be causing this issue?
>
> Below is the trace from each side, organized into the order that I believe
> the events took place, however, I am unsure if this is actually the case, as
> there is a chance that the producer responded to the interest after the
> timeout took place on the consumer.
>

Hi Anthony,

You're showing logs from two different hosts, are the clocks synchronized?

Maybe it's just a case of high packet loss on the wireless medium, due
to interference, distance, or other factors. Can you check if/when the
packets are actually received on the physical interface using e.g.
tcpdump?

Another thing you could try is: build NFD in debug mode (./waf
configure --debug) and look for the log line that says "Detected <X>
dropped frame(s)".

Best regards,
Davide