-
Notifications
You must be signed in to change notification settings - Fork 2.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Increase global peer coverage #222
Comments
I'm experiencing the same issue. I haven't dug into the architecture of the project (apologies) but it seems that 9222 tcp/udp are the discovery ports for p2p. I've opened these up on the node but I'm still experiencing the same issue as @TangMonk - long periods of 0 peers. |
Thanks, will take a look. We set up some default public peers but perhaps they don't have enough capacity. As for the question in the title of this issue, yes we have the static peers that should be provided by the bootnodes in the default config here. This is a recent update, so make sure your config has this too: Line 3 in 18a9591
|
@TangMonk @mariokami Can you all also please make sure TCP / UDP for P2P is forwarded for 30303 and let us know the results? |
and no firewall enabled on my machine |
All above of your provided static peers node are from USA. Not convenient for people from other areas Could u please add some Asian static peers node, Hong Kong, Singapore, Japan is fine. I am living in China, Syncing node is so sloooow. |
Thanks for the feedback; leaving this issue open to track the need for increasing global peer coverage to facilitate easier discovery. |
@wbnns thanks for the response - 30303 is forwarded on this box, and I'll keep you posted with updates. I wasn't thinking about 30303 because it says in the docker-compose.yml file :
@TangMonk the bootnodes are just for discovery (ie: to provide addresses for other nodes via Kademlia DHT) so their geographic location shouldn't make too much of a difference. I'd be interested if there was something to suggest otherwise. |
Is it 30303 (conflicting with Ethereum) or is it 30301 ? The static bootnodes list has 30301 for all their ports... |
Are the boot nodes still overloaded? Same issue here with no connected nodes, no firewall. |
Took about 1 hour for my node to find one peer, after that the whole chain sync took 1 hour as my L1 beacon and exec are local to the infrastructure. To be fair it kind of feels like you dont need more than 1 peer to start syncing. Would it be too hard to provide a reliable list of 10 or so static nodes? |
The boot nodes don't (shouldn't) participate in peering iirc - they just provide pointers to additional peers via Kademlia DHT lookups. If anyone has a moment and is interested there is Nebula which can crawl DHT networks for analysis. It would be interesting to see what the properties of the Base network are like. |
We do provide a set of nodes available for peering, but apparently not enough. We will look into increasing capacity. |
been running for days without being able to find a single peer |
So.... any new peers? |
We cranked up our snapshot peering this week, let me know if you still have issues. |
I thought this was the basic goal of base eth. |
Yes, it's still an issue. I've been trying to resync after updating to .4 for about 4 days now. Its stuck at peercount=2 but makes no progress. |
We have added one more machine with additional 500 peer handling capacity, though I am not sure if that will help, as the issue may be discovery related. Still digging into this. One thing I found that helps with peer connectivity is to make sure inbound connections to 30303 are open, and you specify your external IP address appropriately (geth --nat=extip:[your external ip address]). Is that something your setup would allow? |
Wait, why 30303 ? That's Ethereum. |
right op-geth is just a fork of geth, so it defaults to 30303 for p2p. You can override it to whatever port you prefer with the --port flag however. |
now im in a situation where my chain finds 1-2 peers quick enough but it never syncs. just stays in looking for peers. why? |
I'm here, and it's not syncing. Doesn't matter if I start with an empty data dir or with a snapshot. Using the 0.8.4 release. |
Can you try opening inbound connections to port 30303 and specifying the HOST_IP appropriately? |
Both are done and it's not starting the sync |
Can you confirm you can "telnet <HOST_IP> 30303" (or whatever port you're using) to the node from an outside machine? I am having trouble replicating so wondering if your port is still blocked somehow. We're still working on allowing more peers to be found even if the port is blocked. Thanks for your patience. |
I can confirm, yes. It's a kube cluster but the NodePort 30303 is being
proxied through to the box, and the HOST_IP indeed can be telnet'd into
30303 from an external machine, and it arrives at the right pod.
Also I dont have issues finding peers. It's just not starting the sync.
|
I have 2 peers already within 2 minutes of restarting. Now running 0.8.3 I think. Same old. |
Would you mind sharing your enode id? (Feel free to e-mail...) I can see if I can peer my node with it and see if that gets it syncing. |
I also have 100+ peers connected now but it's not trying to catch up the chain. Can you advise of current configuration that might make this weird? |
Interesting so now peering is better but still no sync. assuming you already uncomment the flags here? Line 41 in 2dbf4a4
Is your config otherwise standard? |
Ah, no, I didn't realize there were new envs there, and we are supplying them via a cm so didn't pick up the file automatically. Will try with that and report back. Thanks. |
@roberto-bayardo yea, even with 0.9.0, the updated env, testing changing between local L1 rpc/beacon to a chainstack archive one, it still doesn't start syncing. data dir is wiped, it says starting snap sync mode and that the database is empty, but up to 10 nodes now and it doesn't start any sync -- what am I missing? :( |
I am having the same issue here. Long period of 'peercount=0' . Already changed the 'HOST_IP' for NAT (though I have a public ip) and added bootnodes. |
Can anyone share more bootnodes address here? |
Line 45 in 8e7460e
|
I had the same issue and solved it by deleting everything from root@full-2tb ~/base/blockchain # ls geth
chaindata LOCK nodekey nodes transactions.rl I believe the root issue is If you have trouble with downloading the state (after reaching some peers) please check local optimism node logs (
|
Did you find a solution? Got the exact same here. Peers but no blocks. |
It's hard to looking peers:
Very long time, still 0 peer
The text was updated successfully, but these errors were encountered: