ITS ALIVE!

Reply to this note

Please Login to reply.

Discussion

I did something right here because the relay is down to 32ms response time from a varying 4000-12000 ms. Let’s see if it continues to behave like this.

make sure you have 1024 max_conn and `ulimit -a` max file handles is higher than 1024

What open files limit do you have? Seems like some do lower thousands and some do 50k, 100k+. I haven’t set max_conn yet though, will do that.

i noticed mine was set to 1024 and then the relay started throwing errors at around that many connections. Need to run `ulimit -n 99999` before running the relay if its too low.

amazing

Watching as it all unfolds and progresses is pretty amazing indeed 🐝⚡️

💯

Tbh I think most of this improvement is from #[3] ’s fantastic nostr-rs-relay improvements, as well as Anigma stopping their accidental DDOS attack with kind 70202 events.

DDoS attack was just a test to prepare you

lol, we’re not prepared for a real attack. We should definetly look at rate limiting or something similar. Would be nice to be able to whitelist certain event kinds too.

DDoSing on nostr is inevitable, but hopefully it will make us stronger, and deploying relays as easy as download+install. If they can be that easy, I think we will win. Also, if relays enforced perhaps lightning network micropayments, that would deter DDoS.

I think proof of work needs to be attached to each msg that is costly to the server

Yeah that will probably help too, makes it a bit more expensive at least