Mark Henderson

#2 definitely sounds like it is a bug and should be fixed. #1 might be better expressed as an SCP? https://github.com/orbitdb/SCPs

We do something very similar already in `orbit-db` with caches, keystores and stores IIRC: https://github.com/orbitdb/orbit-db/blob/main/src/OrbitDB.js#L53-L56

@EmiM is it possible for you to contrive some example that I can reproduce?

Yeesh, 1.4 seconds just to read a dag node from IPFS :sweat:

Setting the timeout higher will mitigate the issue but it's still curious to see it take such a long time in a local dev environment

Hmm, could it simply be tor routing? i.e. can you do a simple websocket packet test

They would timeout after 1s which would cause the failure. Don't get me wrong - I think we should _definitely_ speed up replication even more but I would like to...

@phillmac refresh my memory here? :D

Ah yes... the event listener leak thing...

Not yet, that's next on the list because we need to refactor the replicator anyway