@eyedeekay
&eche|on
&zzz
+R4SAS
+RN
+RN_
+StormyCloud
+T3s|4
+acetone
+eche|off
+orignal
+postman
+qend-irc2p
+snex
+weko
+wodencafe
Arch
BubbRubb
Danny
DeltaOreo
FreefallHeavens
HowardPlayzOfAdmin
Irc2PGuest18900
Irc2PGuest38175
Irc2PGuest62240
Irc2PGuest64908
Irc2PGuest87067
Irc2PGuest88693
Sleepy
SlippyJoe_
T3s|4_
Teeed
aargh2
ac9f
b3t4f4c3__
bak83
cumlord
death
dr4wd3
dr|z3d
duanin2
eyedeekay_bnc
idk_afk
leopold
makoto
nilbog
not_bob_afk
poriori_
profetikla
r00tobo_BNC
rapidash
shiver_sc
solidx66_
thetia
u5657
uop23ip
w8rabbit
x74a6
eyedeekay
Put anubis ahead of the git server, seems to have made a huge difference, still have some oddness server-side but it should be substantially more tolerable to use
eyedeekay
No anubis for I2P users btw, only on the clearnet side for now
dr|z3d
there we go :)
dr|z3d
it's clearnet that's killing your gitea instance. or not (anymore) :)
dr|z3d
night and day difference.
dr|z3d
it won't 503 for me.
snex
Use go-away instead of anubis
dr|z3d
this PR still missing commits, I think this was the one zzz mentioned before? git.idk.i2p/I2P_Developers/i2p.i2p/pulls/519
eyedeekay
Yes it is, something to do with health checks or missing refs I think, I have some more ideas to try
dr|z3d
might just be easier to nuke it and get zzz to make another PR.
eyedeekay
Probably less of a headache
eyedeekay
I don't *think* it will keep happening to new ones if the server doesn't have to play whack-a-mole with git processes anyway
dr|z3d
I think you've probably addressed 95% of the issue.
dr|z3d
I don't think I'd relish looking at the clearnet access logs, but they probably indicate wherein the problem.
eyedeekay
Yeah they're a lot of "why the hell is this AWS IP constantly visiting repositories that don't exist since the migration"
eyedeekay
I didn't want to go as far as Anubis(also I would have named it `Zuul`) but it's clearly making a difference
eyedeekay
Also speaking of headache's I've got a real doozy coming on so I'll be back around 9EST tomorrow
dr|z3d
hope it subsides :| ibuprofren always works for me.
snex
how do i host two different service types on the same b32? ive seen sites that do this but when i try to tell the service tunnels to reuse a privkey then they both try to compete over it and cant both run at the same time