Block spreading through WebSocket not effective
I noticed I oftenly had in Sakia a network split into 2 parts when a new block is issued:
However if I restart Sakia, I get this screen:
So there is something wrong with network refresh. I first wondered if the problem was coming from Sakia, so I've tested with an ad hoc mirror on TestNet.
And I get the same issue: the new block is not spread correctly through WebSocket all the time by all the nodes, even if the node did well receive and accepted the block.
So the problem must come from Duniter.