IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#i2p-dev
/2026/02/06
&zzz
+FreefallHeavens
+R4SAS
+RN_
+Romster
+T3s|4
+acetone
+eche|off
+mareki2p
+nilbog
+nyaa2pguy
+orignal
+postman
+qend-irc2p
Arch
Daddy_I2P
Danny
Irc2PGuest15434
Irc2PGuest16019
Irc2PGuest20377
Irc2PGuest31938
Irc2PGuest35412
Irc2PGuest48148
Irc2PGuest77921
MatrixBot
NiceBoat
Onn4l7h
Onn4|7h
Over
Sisyphus
T3s|4_
Teeed
aargh4
ahiru
anontor
b3t4f4c3__
cims
dr4wd3_
duanin2
ice_juice
mahlay
makoto
marek
n2
noidea
not_bob_afk2
o3d3
poriori
profetikla
r00tobo
rapidash
rednode
solidx66_
stormycloud[m]
sublimia
test7363673
uop23ip
urist_
vivid_reader56
wodencafe2
x74a6
zelgomer
orignal zzz, FYI
orignal seems botnet comes with wrong static key
not_bob Thank you.
nilbog I tried -6-rc on debian by building it, had to fix the debian build first. But maybe i didn't completely fixed it bc the web interface couldn't load and i got an error related to jetty afterwards and the router crashed repeatedly. Unfortunately lost the log (extra dumb..) so will try to replicate as soon as i can connect to git (seems down?)
zzz thanks for the report nilbog. we hadn't fully and successfully tested the deb builds until yesterday. Fixes are in -7-rc
zzz if you can't get it from gitea in-net, it's on github
nilbog ok thanks will do!
zzz please report results
zzz we're a little swamped trying to get ready for the previously-scheduled release while simultaneously working on the attack
zzz last call for translations, pulling from transifex in 3 hours
orignal or it's postponed again?
zzz still monday until eyedeekay says something different
zzz re: s mismatch, yeah, he's changing identities
orignal you mean attacker?
zzz yeah, re: your 333.i2p link
orignal that what a guy noticed
orignal so he changes router keys and doesn't change ntcp2 and ssu2 static keys?
orignal then how it even works?
orignal i2pd always closes such sessions
zzz yeah I don't know how he would change ident while running
zzz hackers gonna hack
orignal I can tell you what they do
zzz thinking about if we can ban for mismatched s, I think we can if the IP matches
orignal they delete router.info and router.keys
orignal and not ntcp2.keys and ssu2.keys
zzz we won't ever change while running
orignal as result he uses key from nntcp2/ssu2/.keys
orignal but sends new RouterInfo
zzz we compare everything at startup and redo everything on mismatch
orignal i2pd doesn't )))
zzz i think ))
orignal maybe we started doing it later
orignal Vort mentioned something like this
orignal but they use older version
orignal do you compare S with one from RI now?
zzz yes ofc. It's just whether we can also ban the hash from the RI. Have to be careful not to ban if he sends somebody else's RI
zzz so need some sort of assurance that it's his RI, either by IP or S match
zzz if neither IP nor S match, you can't ban (preliminary conclusion, still thinking about it)
orignal then let's do it
orignal what's you suggest ban time?
orignal 10 minutes of 24 hours?
zzz ok. what I have now may be sufficient, or might make a change
orignal what you have now for this secenario?
zzz up to you. I have to worry about ban table memory usage with these kinds of attacks
orignal I onle close the session
zzz sec...
orignal what? 4 or 16 bytes?
orignal you can ban whole network without noticable memry useage
zzz java object overhead... also depends on ban by IP or router hash too, still thinking...
zzz any java HashMap entry is about 200 bytes overhead
orignal by IP only
orignal that's what I'm going to do
zzz our inbound ban time is 39 minutes
orignal thanks. will do the same
zzz if you like, nothing special about 39 ))
nilbog Update on debian build (sorry if dupl, got disconnected), some fixes are still required to compile: in debian/rules add ",quilt" at end of line 54 'dh $@ --with systemd,bash-completion,quilt' as explained in an issue on gitea + add classpath of slf4j2-api and jetty-ee in javadoc target in build.xml
nilbog maybe it's not a complete fix as lots of bugs happen in the ui, router indicated as down sporadically, no graphs visible (they are saved though) and lot of pages like local router (in netDb) errors with : 500 javax.servlet.ServletException: java.lang.NoSuchFieldError: Class org.eclipse.jetty.http.UriCompliance$Violation does not have member field 'org.eclipse.jetty.http.UriCompliance$Violation
nilbog BAD_PERCENT_ENCODING' (running java 21)
zzz thanks nilbog I found the same things
nilbog ok cool
nilbog also much much harder to build tunnels it seems (compared to 2.10.0), even keeping irc alive is very complicated, couldn't do it so had to go back to 2.10.0 (even after 45+ min)
zzz I have a fix for the javadoc, trying to push but gitea won't talk to me
zzz the ,quilt thing is a known problem for years
zzz looking at the BAD_PERCENT_ENCODING thing now
StormyCloud Alright, we got 30 routers on -5 now
zzz the problem is a mismatch of the jetty versions
zzz super StormyCloud
zzz still need to get it up on LP, been working on trixie/resolute issues
zzz I'm ging to have to bundle some 12.0.17 jars to fix the PERCENT_ENCODING issue
zzz this could all have been avoided but debian didn't respond to my bug report a few months back
zzz I have the PERCENT_ENCODING thing fixed but having trouble pushing to gitea
zzz -9-rc with the trixie/resolute fixes pushed. trying to push the -8 and -9 tags now
lbt Hi! What's gonna happen to routers running on Debian Bookworm when the new version comes? Means no new Jetty, but it sounds like that's gonna be needed, yes?