Usually after a lot of traffic is sent to a network, inducing a high load, we can get into a state where libp2p receives messages (there is a swarm log entry for eg. a BlockResponse), but these are never delivered to the node application and we enter a vicious cycle of rerequesting blocks which are already in our input buffer.
Usually after a lot of traffic is sent to a network, inducing a high load, we can get into a state where libp2p receives messages (there is a swarm log entry for eg. a BlockResponse), but these are never delivered to the node application and we enter a vicious cycle of rerequesting blocks which are already in our input buffer.