IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#i2p-dev
/2022/04/24
@eyedeekay
&kytv
&zzz
+RN
+RN_
+T3s|4
+dr|z3d
+hk
+not_bob
+orignal
+postman
+wodencafe
Arch
DeltaOreo
FreeRider
FreefallHeavens
Irc2PGuest28511
Irc2PGuest64530
Irc2PGuest75862
Irc2PGuest77854
Nausicaa
Onn4l7h
Onn4|7h
Over1
R4SAS
Sisyphus
Sleepy
Soni
T3s|4_
Teeed
aargh3
acetone_
anon4
b3t4f4c3
bak83
boonst
cancername
cumlord
dr4wd3
eyedeekay_bnc
hagen_
khb_
plap
poriori
profetikla
r3med1tz
rapidash
shiver_1
solidx66
u5657
uop23ip
w8rabbit
weko_
x74a6
dr|z3d zzz: minor to the point of "who cares?" but you're mixing cases for Alice and charlie all over. I'd suggest capitalizing names and sticking with it :)
zzz lol
dr|z3d glad it tickled you :)
zzz any more serious complaints?
dr|z3d no. you're good.
dr|z3d any ETA for when SSU2 peer testing can be switched on?
zzz I got it working an hour ago but it's going to be real tough to have it kick in on the live net
dr|z3d yeah, figured it would be a while yet before you're confident enough to go live with it.
zzz because Capital A Alice must randomly pick a Capital B Bob that is connected SSU2 and Bob must have a capital C Charlie also connected SSU2 but is not connected to Alice
dr|z3d so it's as much a question of numbers as anything else?
zzz yup
dr|z3d and in the absence of enough ssu2-enabled peers? things break, or they carry on regardless?
zzz in theory the test result is "unknown" and everything carries on
dr|z3d ok, what's the worst case scenario?
zzz hopefully, just errors in the logs
dr|z3d I mean, in the event of yet-to-be-discovereg bugs surfacing? router crash?
zzz havent seen anything worse
dr|z3d ok, so it seems to be that you could probably take it live round about now and just wait for more people to opt into ssu2?
dr|z3d if that's likely to help with testing, at least.
dr|z3d code-testing, I mean.
zzz I think I'd rather wait until i2pd is ready to test
zzz they're working on relay which I havent started yet
dr|z3d ok, sounds reasonable enough. can't be too long to wait, then, orignal seems to be motoring along.
zzz so we're planning on meeting in the middle in a week or two
dr|z3d is orignal still planning to ship ssu2 on in the next release?
zzz same as us. in there but off by default. they don't have any dev/test branches afaik.
dr|z3d ok. could be a good time to announce latest progress on the forum, get more people interested in enabling. we're still a little light on the uptake it seems.
dr|z3d eyedeekay1: receive !recieve :)
zzz eyedeekay1, zlatinb, misspelling or not, why a whole paragraph about bundles in the main feed? seems irrelevant and could be confusing to people
zzz also, I thought the issue was in java 15-18 but you're recommending "below Java 14"
dr|z3d bit of a strange intermittent issue you might have an idea about, zzz. remote console access occasionally locks up, "Input stream closed" is reported in the console logs.
dr|z3d router remains functional, just no console access.
dr|z3d apparently it's a client -> server issue, though I can't figure out what the issue is.
dr|z3d I have a feeling the same issue occasionally shows up for hosted sites.
zzz remote meaning SSH or tunnel?
dr|z3d ssh access otherwise a.ok
dr|z3d just to confirm, I closed the ssh session and opened a new one. issue persists.
zzz so once it's stuck it's stuck for good?
dr|z3d looks like it, it's been over 1/2 hour since I haven't been able to get in.
dr|z3d it may come back, not sure. if it's the same issue as with hosted sites, it resolves after a while. haven't seen it much, so can't say if it's a terminal issue with 100% confidence.
zzz two ideas: 1) deadlock; 2) open files ulimit
dr|z3d $ ulimit -n
dr|z3d 65535
dr|z3d I was looking at the console logs, let's have a look at the wrapper logs.
dr|z3d no deadlock. router's functioning normally other than no access to console. services still being served etc.
zzz thread dump to be sure
dr|z3d can I invoke a thread dump from the command line?
zzz i2prouter dump
dr|z3d thing is, in all other respects the router's behaving as expected, sites still being served, router logs indicating nothing amiss. so I don't think it's a thread lock issue, but I'll give that a shot.
dr|z3d ok, your hunch was right. there is a jetty deadlock.
dr|z3d you want to take a look, zzz?
zzz sure, unless you're going to fix it. it's not gonna fix itself ))
dr|z3d :) you want the whole dump?
zzz no, just the deadlock part at the bottom, as usual
zzz jetty version?
dr|z3d current, no diversion from upstream.
zzz you have a console password set?
dr|z3d on that box, yes.
zzz it looks like it's hung up in the authentication
zzz it's happened more than once?
dr|z3d yeah, not the first time, though I don't remember when it was it last happened. more frequently are web services becoming inaccessible, not sure if that's related.
zzz doesnt look like it would be related to password authentication
dr|z3d true.
zzz disabling password would probably be a workaround
dr|z3d yeah, wouldn't be happy with that, password is there for a reason :)
dr|z3d it's not frequent enough to be a hear tear-out issue, but it's annoying when it happens.
dr|z3d *hair
zzz it's pretty far down the stack from any of our code, so unless we've configured the authentication wrong, we'd have to look for a fix in 10.x and backport it
dr|z3d yeah, if the auth was wrongly configured, it would probably have turned up issues before now, no?
dr|z3d about as close as I can find, issue wise, when searching for 'jetty 9.3 authentication issue'
zzz please open an issue on our gitlab and include the stuff from the pastebin because this doesn't look fast or easy
dr|z3d sorry it's not more descriptive, not really sure how much more info I can convey.
zzz the problem is clear, and the problem is it isn't our problem, but we'd have to pay them to make it their problem
dr|z3d yeah, I hear you. maybe if it was an issue in whatever they're working on now (11?) they'd take an interest.
zzz that avenue doesn't work with them
dr|z3d oh, ok, sounds like you've had previous dealings with them and it wasn't all ponies and rainbows.
zzz just that you can't trick them into working for free
zzz on old versions
dr|z3d right, I was suggesting that _if_ the issue was with 11, and we were using 11, they'd be more inclined to want to fix it. as is, 9.3 is dead to them I suspect.
zzz sure, if you reproduce it on 11, they'll fix it in 11. you can't trick them into attempting to reproduce it on 11 or doing any research themselves
dr|z3d *nods and smiles*
dr|z3d on a parallel not, are we due an upgrade to 9.4 soon, given debian's now on 9.4?
dr|z3d *note
zzz 9.4 requires java 9, that's why we bundle 9.3, and we include patches in debian to work with 9.4
dr|z3d ah, gotcha
dr|z3d another jetty/i2ptunnel aside: jvmbrotli.com looks interesting