#ls2 (irc2p) | #ls2 I2P Protocol Development Discussion - Next Meeting Monday, October 9th 6:30 UTC

zlatinb zzz: I'm having serious problems fetching zzz.i2p despite having a successful dbStore of leaseset in the logs

zlatinb Leases: #1

zlatinb [Lease:

zlatinb End Date: Dec 23, 2021, 11:45 PM

zlatinb Gateway: [Hash: fSYNpvQVLb-BA23mRaYYIvhC-jDBUDkjnGp18x1Vlis=]

zlatinb which router isn't in the /netdb

zlatinb another attempt, again one lease. Are you running zzz.i2p with a single tunnel?

RN also having issues loading several sites including zzz, dr's comes up though, and not sure where to look for leasesets in logs

zlatinb RN: add the following to the logger.config file:

zlatinb logger.record.net.i2p.router.networkdb=DEBUG

zlatinb logger.record.net.i2p.router.networkdb.kademlia=DEBUG

zlatinb then in about 30 seconds the router will start logging the leaseset requests

zlatinb whenever a successful leaseset gets fetched, there will be a multi-line entry starting with "Handling dbStore of leaseset"

zlatinb it will contain the number of leases and the actual leases

zlatinb for reference, the zzz.i2p b32 is lhbd7ojcaiofbfku7ixh47qj537g572zmhdc4oilvugzxdpdghua.b32.i2p

zzz that's going to be an overwhelming amount of logging. Much easier to just look on the leasesets tab on /netdb in the console

orignal I don't see EfRZwfOuyrNQMnnTcSogZ8NKeAP-q~zaqD9lJAb-DVs= on my floodfills

zzz oh well

orignal can you descritbe what you see?

orignal duplicates? different adddresess?

zzz I'm still seeing an enormous number of i2pd routers with no network ID and/or localhost addresses, most are .51 or .52

zzz I think EfRZ had almost all addresses ::1

orignal no network id means bad i2pd fork

orignal because i2pd always adds netid

orignal local address basically means Tor

orignal some "very smart" guys do it

orignal intead using SOCKs proxy

orignal they forward traffic trougr iptables

zzz found it in the logs. 2 SSU (one with introducers, one without, both port 0) and 10 NTCP2 (caps 4, no host)

zzz all the NTCP2 ones were the same

orignal I don't think it's real i2pd

orignal also not sure what is better

orignal to publish or not publish local address

zzz I was being nice and just banning them for a little bit, but maybe I should go back and ban them until restart

zzz or maybe based on version

zzz what version did you fix it? a couple of years ago I think

orignal spring 2021

orignal for me it's not a problem

orignal I just ignore extra addresses

orignal and local

zzz sure but I can't ignore no network ID

orignal as I said

orignal netid was always there from day one

orignal i2pd never published without netid

zzz ok

orignal it can be zero

orignal also I think I check netid for RouterInfo to see if it matches ours

orignal and if not I drop it

orignal also I'm thnking to redo addresses

orignal always have 5 slots for each RI

orignal each can be filled of empty

zzz yeah I can't support ygg unless I redo addresses also

orignal it's not about ygg

orignal 4 addresses in your case

orignal also back to ygg

orignal you don't need to support ygg in transports but why you can't pick ygg-only router for a tunnel?

orignal e.g. tunnel A-B-C, B is ygg-only

orignal and A and C also support ygg but also have ipv4

zzz because my address code is such a huge mess

orignal how do you check compatibility?

orignal like if A and connect to B

orignal *can

zzz do I have a v4/v6 IP address, and is the IP in the peer's RI a valid public IP

orignal then it's easy

orignal when you check ipv6 check if both are in ygg range

orignal in general I think to make it more genreic

orignal e.g. address my belong to ranges we don't support yet

zzz the problem is that deep in the code, each transport can only publish one v4 and one v6 address. That's why it's not easy

zzz it would be a big refactor

orignal so you have 4 fixed slots?

orignal in your code

zzz sort of like that. Really 2 per transport

orignal in my opinion you should start refactoring it

orignal because it's possible to have multiple ipv6 addresses anyway

zzz sure. it's been a mess ever since we added ipv6

zlatinb If you hate IDEs, wait till you see this: https://github.com/github/copilot.vim