Closed ghost closed 7 years ago
@karek314: This issue is caused by lisk-ui
. There is an issue already raised there which we will resolve: https://github.com/LiskHQ/lisk-ui/issues/86. Thanks for bringing it to our attention.
Thank you, may i ask you - according to logs. Why Lisk banned set of other nodes, and then tried to get height from them which obviously resulted in failure due ban? Why those peers has been banned at first place?
I'vent checked lisk source code, but it seems it's working on normal http requests, wouldn't it be better to implement network connectivity additionally over websockets, limiting connected peers to 15-25 per node, so network would be able to propagate data much, much faster and more reliably. Current propagation times are not good. I have also noticed that most nodes are usually "connected" to 100 or more peers, how it's being calculated ? If it's not websocket connection, then technically it can't be understand as connected, but maybe more like was connected in last N seconds?
@karek314 The peers were banned because they did not respond quickly enough to the request.
Regarding use of websockets vs http requests, please raise as a separate issue. We will deal with it there.
Version: 0.5.1 I've noticed that balance in /dashboard did not matched real balance, while sending - UI returned error with real balance, but still balance in /dashboard has not been updated. It happened during time when node forked and recovered itself from fork by deleting last block. Not sure if this issue belong here or lisk-ui.
It happened that i've catch logs for block rewinding and further behaviour. Im not sure why lisk banned so many peers after that ?