Shortly; woke up in the morning, noticed that my node laptop had disconnected from the network (0 connections). Reconnected → all running got number of peers and conns for nodes. But now I noticed that there are no earnings since the break? All nodes still have connections. I’m running 24 nodes on the laptop, usually get ~1 reward / 2 hours. Now 15 hours, no rewards.
Earlier I did not restarted the nodes, just reconnected thenetwork. Now I did, let’s see if it helps.
It’s possible your node have been shunned due to being offline, and therefore earning again may be tricker… If you were notmally getting regualr earnings, then this is a pretty good indicator.
It might be worth resetting some, or all of them to see if that gets you going again
After restart, one reward. But anyway I will dump these nodes and restart new ones.
Is this not good for the network? The network ‘teaches’ the noderunners to behave badly? Badly behaved node keepers are rewarded and well behaved are punished…
In my opinion, old nodes having already loaded network data should be welcome to get back. This is even more important when the network is not almost empty like now.
Yes they are, they do come back in with existing data and can get up and running again — that’s part of the design — but something else may have been at play here, hence my suggestion to reset them.
In theory a break of more than an hour should see the nodes all shunned by the RT peers.
They comeback and can only connect to another set of RT peers, potential here for crippled set of nodes rarely being shown as a close node thus no earnings.
I would suggest that when instructions are written up that for an extended outage the nodes will need to be reset for the health of the nodes and network.
Mind you the actual design was and supposedly still is for a node restarting to gain a new Peer ID and this problem would never exist. But that is for a future update.
I see this a kind of a problem, but good to know it will be fixed.
But is there really need to mark a node shunned? Or consider “shunned” as temporarily away and just change connection interval to much longer than online nodes?
I am sure this will also be looked at, but consider that once a node is absent for a good period of time then other nodes have had to take up those chunks it was storing, the node is not performing anything useful for the network so its rightly shunned. Your node maybe fading in and out causing unnecessary churning each time and this is not good for the health of the network. So yes I would consider it worthwhile to shun.
That fading in and out is a valid reason to make it permanent since that behaviour could be happening for a month when someone runs them and only checks rarely.
Also permanent to resist malicious nodes using the same peerID to disrupt the network in some manner. In the scheme of things this will not be a real issue since there are 2^256 peer IDs out there and practically impossible for a new node to use a peerID that has been shunned
I ran up 60 new nodes and only 6 got any payments. In 12 hours there was a 1 in 10 chance if that rate is typical across the network.
It has to. The network could be in danger otherwise. It reduces the resilience of large network outages
Guessing malicious methods is a dangerous game. Better to just shun if bad behaviour is found and leave at that and remove the attack vector altogether
The network tried to assure it has n copies of every chunk. (n=3 currently?) What if it allows copies to be also more than that. So
1 a node disappears from the network
2 network copies all node’s chunks to other nodes
3 node comes back
4 nothing happens, there are just n+1 copies of the chunks
5 the same node goes out again
6 nothing happens, there are still n copies
I assume currently the network works hardly on steps 4 and 6? Might be wrong?
Btw, I think the network should not accept new nodes as quickly it currently does. It should take them into use lazily. A new node have a lot greater possibility to disappear soon compared to old long running nodes.