Why does disconnecting the node from the Internet and connecting back cause the node to all of a sudden receive a lot less traffic than doing that + restarting the node?
Receiving 10 Mbps+ during status quo, disconnect from Internet for 10 seconds -> once node recovers receiving small spikes of 4~ Mbps with pauses in-between. Wait 10 mins. No change.
Shut the node down, start it back up -> back to 10+ Mbps traffic within seconds.
Why the difference?
Are you stopping your internet or just pulling the plug out of the node? Im not sure if thats the same as stopping your node completely letting it finish all the queues and just stopping it abruptly. Other then that I dont know what the difference would cause. Possible when you just pull the plug it takes time to reconnect to everything vs restarting the node it connects right away.
I can’t say because I havent just disconnected my node from the internet, But I have restarted it many times and it seems to go back to getting data like normal.
Just disconnecting the Internet, the network stays connected.
The node isn’t stopped abruptly, it just loses connection to satellites abruptly. In any case it’s a bug IMO if satellites treat it differently because you didn’t restart the node, which also makes it waste rescources on calculating all that startup stuff again. I assume some startup stuff needs to be added that happens when a node loses conection and is reconnected. In both cases the node didn’t get to finish sending/receiving all the queued stuff but the results vary drastically.
Because during the downtime your node didn’t respond to an uptime check on some of the satellites. That satellite will then have marked your node as offline until the next uptime check. However if you restart the node, the node checks back in with all satellites itself and is automatically marked as online again.
I suspected something like that. But why does it still receive data then, just at much reduced rate?
The node should check for connection every minute or so and inform the satellite that it is online when it manages to connect, this won’t add any additional burden to the satellites.
Other satellites may not have noticed yet that your node was online. They do their own uptime checks.
I see, kind of too consistent for a random occurence, happens the same every time. I guess I should add an issue on Github to improve it if they want.