Too many connections to the farm on the node can lead to the BPS dropping to zero, and the node stops syncing

When my node connects to around 80 or more farms, the BPS becomes zero. If I turn off all the farms, it returns to normal. Why does this happen?

Hard to say, this is certainly not an expected usage of the node, but I also don’t see why would it cause issues.

If you can provide all logs (as text) from both node and all farmers, there should be something there indicating an issue.

I see block 729582 actually contains segment headers, which is something that I already noticed causes spikes on the network and will debug it. I don’t think it is necessarily related to the number of farms you have connected.

What parameters do I need to start node with to get more detailed logs?

Even default log level should be quite telling, but I think this particular situation is something I need to debug and fix, then you’ll just upgrade and the issue should go away.

OK, I will provide you with detailed logs the next time the node is disconnected.

This is the first node that was started entirely according to the official default parameters. The BPS of this process became 0 around 2023-12-21T08:49:15.119317Z.

2023-12-21T08:49:15.119317Z [Consensus] 💤 Idle (49 peers), best: #747868 (0x7b02…48b7), finalized #669423 (0x10a3…98d3), ⬇ 7.9kiB/s ⬆ 1.6kiB/s    
2023-12-21T08:49:20.119846Z [Consensus] ⚙️  Preparing  0.0 bps, target=#747874 (48 peers), best: #747868 (0x7b02…48b7), finalized #669423 (0x10a3…98d3), ⬇ 15.7kiB/s ⬆ 2.1kiB/s    

This is a node where I increased the number of connections.

Did farmers disconnected and restarted?
Today’s dec-22 release addresses this issue and I think you should not see 0 bps sync on segment archiving or hit RPC issues.

Please upgade and let me know how it works over next few days.

It has been running stably for 8 hours now.

1 Like

Give it another two days to make sure it went through a few archived segments without major disruption.