[atlas] DNS probes: spurious SERVFAIL

On Sun, 30 May 2021 at 14:49, Lukas Tribus <lukas at ltri.eu> wrote:
>
> Hello,
>
> On Sun, 30 May 2021 at 14:39, Stephane Bortzmeyer <bortzmeyer at nic.fr> wrote:
> > > To reproduce, just try to resolve zoom.us or google.us against 8.8.8.8
> > > or 1.1.1.1 on your probe.
> >
> > It works and I don't see one SERVFAIL:
> >
> > % blaeu-resolve --requested 100 --area North-Central --nameserver 8.8.8.8 --type A google.us
> > ...
> > Test #30484765 done at 2021-05-30T12:34:54Z
>
> That's a major difference to what I'm seeing, and I don't get why the
> difference is so huge.
>
> Just now:1000 probes, area North-Central, A record google.us against 8.8.8.8:
>
> Above 90% of ther answers are SERVFAIL (the rest probably is not in
> yet, or timeout).
> https://atlas.ripe.net/measurements/30484987/#probes
>
> Only 12 (twelve) probes returned NOERROR.

Here's the cause:

By default, the Atlas UI for measurements has the "RECURSION DESIRED"
flag disabled. It is buried under advanced options, as it would be
some optional flag.

blaeu-resolve set this by default.



Thanks,

Lukas