issues with hip-106 deployment

[nosmaster89]

it has come to my attention that theres some issues still with what i think is the code for hip-106

in reference to the hotspot 112iFVTPaFnmah1idNqKDjT3Z7af3Vz6SfANbmQiwoqyutKeswAU

this hotspots witnesses are correctly being maked bad by hip 106 , but the oracles are validating all its beacons . on the 20th of march it validated 22 beacons in a 24 hour period

[andymck]

On initial look the beacons in question were indeed valid but they had zero witnesses. A valid beacon with zero witnesses does not count against reciprocity and thus its witnesses were declared invalid.

[nosmaster89]

thats correct. but no hotspot on the network should have more than 4 valid beacons per epoch , the oracle should be invalidating any subsequent ones in the same 6 hour period .

[andymck]

This is a side effect of a change made in #759 where beacons with zero witnesses and which pass regular verifications are rendered valid but do not result in the gateway's last_beacon timestamp being updated in order to prevent a valid beacon with zero witnesses impacting reciprocity.

The verification which determines if a gateways beacons are as per the schedule also utilises the 'last_beacon' timestamp and as this does not get updated in the above scenario we end up rendering out of schedule zero witnessed beacons as valid....they will however not be rewarded ( due to having zero witnesses )

Will consider further and look at our options

[nosmaster89]

thank you for the explination, i suspected as much that they are all only validated due to hip 106 and would not actually happen in a working setup

[andymck]

addressed via https://github.com/helium/oracles/pull/784

[riobah]

793 might be related