IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#saltr
/2024/05/03
~dr|z3d
@RN
@RN_
@StormyCloud
@eyedeekay
@orignal
@postman
@zzz
%Liorar
+FreefallHeavens
+Xeha
+bak83_
+cumlord
+hk
+poriori
+profetikla
+uop23ip
Arch
DeltaOreo
FreeRider
Irc2PGuest10850
Irc2PGuest19353
Irc2PGuest23854
Irc2PGuest46029
Irc2PGuest48064
Meow
Nausicaa
Onn4l7h
Onn4|7h
Over
T3s|4__
acetone_
anon4
anu
boonst
enoxa
mareki2pb
mittwerk
plap
shiver_
simprelay
solidx66
u5657_1
weko_
zzz not out of ideas yet
zzz there's a lot going on if you know where to look
orignal I see few exploratory lookups per second as 2RRY
orignal is it expected?
zzz I don't know what a typical expl. lookup rate is
orignal the question is if you are capable to handle them
orignal 2RRY returns 16 hashes now
dr|z3d are you asking about lookups, orignal?
orignal if too many exploratory lookup and they are heavy
dr|z3d I think zzz mentioned before that canon's lookup throttle is pretty low.
dr|z3d I don't remember the numbers exactly, something like 9 is 2m?
dr|z3d *9/2m
orignal I receive few exploratory lookups per second at 2RRY
dr|z3d with my new limits, I'm not seeing much in the way of bans right now.
dr|z3d max 30/30s before ban, or 60/90s before ignore.
orignal I'm asking about not from the same source
orignal but overall
orignal and for exploratory only
orignal it seems it's a potential attack vector
orignal if an adersary start sending tonns of exploratory lookups to floodfill from different sources
zzz sure I can handle 16
dr|z3d so you're suggesting some sort of global rate limit for expl lookups, orignal?
zzz wo was having persistent trouble with the form nonces? I know why
orignal no I mean how many lookups you can handle
dr|z3d zzz: go on, enlighten us..
dr|z3d that was the channel mascot blob, mesh.
zzz tabs for consoles for multiple routers up in same browser, same IP (or ssh forwarding), solution is different IPs like 127.0.0.2, ::1, localhost, etc
zzz browsers don't respect different ports when sending cookies
dr|z3d yeah, I mentioned that to him before, but it depends on which browser.
dr|z3d if you're on firefox, things might be ok. chrome and friends, dunno.
zzz simultaneous http and https to one router also will do it
dr|z3d firefox now does tab-isolation, so you should be able to run multiple routers on the same address.
zzz no, firefox not ok
dr|z3d ah, then you probably don't have tab isolation active.
dr|z3d I've given up trying to track when mozilla decide to default enable various features.
dr|z3d librewolf does appear to default to that behavior.
dr|z3d I've got at least 3 consoles via ssh running on different ports on localhost, no issue with nonces.
dr|z3d localhost as in //localhost:port
dr|z3d I definitely used to have issues, which is why I was emphatic when telling mesh to not run more than one console in a browser session, with only 1 tab for the console.
dr|z3d and definitely don't run the same console in multiple browsers, that's asking for trouble.
zzz different IPs for different routers always works
dr|z3d indeed it does.
dr|z3d if you want to get clever, add multiple hostnames to your hosts file all pointing at 127.0.0.1 with names for each router.
orignal guys what timeout should I set between two requests of the same destination?
orignal like 3 minutes?
orignal e.g. if I requessted a dest and it was not found I wait for 3 minutes
zzz i.e. the negative lookup cache time?
orignal doesn't matter negative or poisitive basically
orignal lookup might return a malformed router
zzz well, it's important not to have two lookups for the same thing in progress at the same time, "combine" them together
orignal e.g. lookup itself was positive but the router is not in netdb
orignal that's I'm doing now
zzz our negative cache time is two minutes
zzz but it's bucketed, so average is one minute
orignal if I see a lookup going on I only add callback to it if presented
orignal 2 minutes. thanks.
orignal then 1 minutes after lookup complete
zzz I think we require 2 or 3 consecutive fails before we put it in the negative cache though
zzz play around with it
orignal I try 7 attaempts
zzz dunno, maybe a transient issue
orignal until it failes
zzz can't remember
orignal if it fails no reson to try again
zzz we do 5. 7 sounds like a lot, but you can look at your data and see if 7 is worth it
zzz hard to get to 7 with a reasonable timeout
orignal I don't remeber where 7 attamepts came from
orignal it was long time ago
orignal and I suspect it was your suggestion
orignal my timeout is 5 second per request
orignal and 30 second for exploratory
zzz netDb.successRetries: How many additional queries for an iterative search
zzz Lifetime average value: 1.299 (53,616 events)
zzz so we're querying 2.3 peers on average (for a successful lookup)
orignal Vort measured it
orignal usually stops at 3
orignal so you suggest to descrease number of attempts to 5?
zzz I suggest you research it, look at your data. Maybe 7 is too many, maybe not
orignal question
orignal what we do if we receive an exploratory request with real ident?
orignal leaseset lookups are also up to 7