Issue when running 10 nodes genload command

[coby0831]

Hi,
I read the https://chainspace.io/docs/?go#performance-testing and was trying to do the performance testing. I intended to run 6, 8 and 10 nodes so I changed the node_count in the genload_testnet script, however, I found out that when running 10 nodes, the avg_tps suddenly drop to 4 while the other two are over 20K+ tps (running for 1mins). Am I doing it in the correct way? or is there some limits for maximum nodes in this network?

[futurechimp]

That seems like a bug if it's happening, there shouldn't be anything I can think of which would cause it to behave that way.

[futurechimp]

To clarify, I mean that there should not be anything in the protocol itself which would make this behaviour happen.

[coby0831]

Hi, @futurechimp thanks for the prompt reply. Do you have any thoughts on how to solve this issue? can't wrap my head around right now
Would it be different if I run it on cloud, say AWS, instead of running it on local?

[futurechimp]

I think you'd have to give it a shot and see. I can't immediately think of a reason that this behaviour would happen.

I would definitely try it out under some different conditions (different hosting, etc) just to see what happens, but overall my guess would be that there is a bug here. We were very much in the process of debugging SBAC when we stopped work, so I'm not super-surprised if there are problems lurking in there.

[futurechimp]

Nice to see interest in it, by the way!

[coby0831]

I'd love to try but I don't know how to set it up on AWS to work because running genload script is on one computer. I have no idea how to make multiple instances to perform like that. Could you tell me how could I make this?