[atlas] Bug when creating DNS6 with udp_reply_bytes via API

Oh yes, here are some measurements illustrating the problem:

https://atlas.ripe.net/api/v1/measurement/1009504/
https://atlas.ripe.net/api/v1/measurement/1009505/
https://atlas.ripe.net/api/v1/measurement/1009506/
https://atlas.ripe.net/api/v1/measurement/1009519/

Here is one given with the interface that has a proper value for
udp_payload_size:

https://atlas.ripe.net/api/v1/measurement/1009517/

And here two submitted via the API that succeed because no
udp_reply_bytes was given:

https://atlas.ripe.net/api/v1/measurement/1009516/
https://atlas.ripe.net/api/v1/measurement/1009518/

Op 24-05-13 16:42, Willem Toorop schreef:
> Hi RIPE Atlas developers,
> 
> I found a bug in the DNS API.
> 
> When creating a DNS measurements via the API that have udp_reply_bytes
> property set, they never leave the "Scheduled" state and eventually fail.
> 
> When I query the status of the measurements I notice that they have the
> udp_payload_size property set to true in stead of the value given with
> udp_reply_bytes. This is different from measurements created via the
> interface, where udp_payload_size reflects udp_reply_bytes.
> 
> Also, it appears I can only origin 80 probes (by probe id), whilst in
> the interface I can origin from 500 probes. Is that correct? Would it be
> possible to (temporarily) increase my limit to 1050 probes?
> 
> Best regards,
> 
> -- Willem
>