Nicolas Gailly

Full line was ``` ERROR: 2021/09/22 19:24:36 [core] grpc: server failed to encode response: rpc error: code = Internal desc = grpc: error while marshaling: proto: Marshal called with nil...

I think this PR solved this issue https://github.com/drand/drand/pull/752 - @willscott do you agree to close ?

We should probably extends this to "versions compatibility" like having a test that checks previous minor patches can still talk with the latest patch or stg.

We should test that in our testnet but with the kyber [fix](https://github.com/drand/kyber/pull/25), it is not a problem to have non consecutive indices. It doesn't matter for the cryptographic part (roughly,...

Can't we create a duplication, read only, static copy of the real dashboard, refreshed every minute or stg ? Maybe not, I don't know but that would be cheap then,...

As discussed, I dont think this works out because the "absent nodes" that wish to catch up need to make sure they receive the same shares than the others, and...

Would it be okay if I print the origin (address in the group) of the partial signature received, and we monitor for the _absence_ of such messages for each group...

The reason why I'm asking if we can only deal with positive "received partials" is that having a system that logs out things _we should have received_ is not 2...

I think what will means is that even locally on our node adding some logging on the different phases (computation, broadcast, reconstruction) would help understand why it is that much...

I think this issue has more to do with this https://github.com/drand/drand/blob/master/cmd/drand-cli/daemon.go#L21 Currently the logic to know if a daemon has already ran a DKG is very brittle and not unified:...