IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#i2p-dev
/2024/02/24
orignal question about reseed
orignal it looks like that reseed script doesn't force to include few floodfills into su3 packet
orignal it it true?
zzz what script
orignal pythn script that gerenarates su3
orignal official one
orignal if if's true it might be a problem
orignal can you remind what we do if we don't have any floodfiils?
orignal it seems right time to take care about such corner case
zzz I don't know about any official python script
zzz but it's impossible to pick 75 at random and not have a ff in there
orignal but it happens now
orignal Vort noticed it
orignal anyway how should we handle such situation?
orignal if we have no floodfiils afters ressed
dr|z3d if no floodfills after reseed, reseed again over the http proxy.
dr|z3d or just reseed again. whatever works for you.
dr|z3d if you can get the python script modified so it always selects at least x number of floodfills when creating a bundle, all the better.
orignal no, you can send something to a non-ff
orignal and if should respond with DatabaseStore of closest FF
orignal that's why I'm asking zzz to remind
zzz right
zzz not store, but DSRM
zzz then ask him for those 3
zzz and *poof* you know 3 ffs
zzz but if you're running out of ffs, I doubt it's because of reseed. You're probably not doing enough exploration
orignal should an ordinary router respond to lookuo with DSM?
orignal Vort says he wanted to start a new router and was our of flooodfills right after ressed
zzz our non-ffs respond to RI lookups with a store or a DSRM, just like ffs
zzz thats really strange, you should investigate, check all the reseeds and count the ffs
orignal if the sceipt doesn't care it's possible
zzz I really don't know what this script is. Our official reseed software is in go, not python
zzz but if a typ. router has about 1/3 ff locally, that's (2/3) ** 75 chance of all non-ff
orignal eveybody uses pyseeder
orignal and it was official reccomendation on your site
orignal now it's 1/20
orignal when did you siwth to go?
zzz 5 years ago?
orignal I remeber this pyseeder from day one
orignal where is go reseed?
zzz ask eyedeekay
zzz you haven't given me a link to pyseeder or any page saying it's official, and it's never been supported by me
zzz where is pyseeder?
zzz look on our website
zzz we would never get to only 1/20 ff, so it was probably a i2pd reseed
zzz when you reseed how many reseeds to you get? 1? 2? more?
orignal it was an instruction how to build own reseed
zzz we do 2. you don't want to do just 1
orignal because reseeds like to ban you if you reseed too often
zzz because of this and other reasons. you don't want a single bad or evil reseed to leave you in a bad place
orignal fair enough will change it
zzz if you're doing it right you should only reseed once in your entire life
orignal yes, but you know people like to do wier things
orignal like run in dockers without storage, etc.
orignal of always delete netdb
zzz don't ask the same reseed twice of course, ask two different ones ))
orignal yes I know
orignal but also need to implement plan B
zzz (19/20) ** 75 ~= 2%
eyedeekay There are 2 go reseeds in use: "my" go reseed is i2pgit.org/idk/reseed-tools for the one I use, and github.com/diva-exchange/i2p-reseed for the one they use. Both are based on the old Martin61/Mdrollette code but reseed-tools takes all the stuff that we used to do with reverse proxies(rate-limiting, TLS/Let's Encrypt) and moves it into reseed-tools itself, whereas i2p-tools goes the other
eyedeekay direction and removes features in favor of implementing them with a reverse proxy. reseed-tools will also self-configure most of it's options(it will guess where the netDb is on most platforms), monitor the status of other reseeds, and can optionally be contacted over Tor or IPFS if for some reason the regular reseeds are unreachable.
eyedeekay reseed.i2pgit.org is what the current version of reseed-tools looks like in a browser
orignal I'm wondering how many resseds use go and and how many python
eyedeekay Mine, Diva's, and a couple of the German ones use Go, which I think makes it about 50/50
zzz eyedeekay, is pyseeder officially supported by us now, or was it ever?
orignal it was on your site
orignal in instructions
orignal R4SAS used it, acetone used it
eyedeekay Maybe many years ago but by the time I was revising the reseed instructions we were already favoring the Go server
zzz looks like drollete dates back to 2014
zzz orignal, the only pyseeder git I can find is in PurpleI2P github, and all commits are by i2pd people. If anybody is supporting it, it's i2pd
zzz if it was originally from us, I don't know, but your team either started or forked it in 2016 and we haven't touched it since
eyedeekay I don't see pyseeder in my dump of monotone either
zzz acetone, r4sas, you, supervillain, are all the committers
zzz the call is coming from inside the building ))
zzz git/pyseeder> cat LICENSE
zzz Copyright (c) 2016-2017 Darknet Villain
zzz QED
orignal zzz but what was used before go?
dr|z3d stormycloud, go, incognet, go.
zzz I don't remember, but go was 2014 and pyseeder was 2016
zzz you can do your own research ))
orignal it was not written by him from scratch
zzz well he slapped his copyright on it
eyedeekay If pyseeder is doing anything more sophisticated than picking RI's at random, if it can read what's in an RI at all, it should be straightforward to add a check to ensure that a bundle contains x number of floodfills