@eyedeekay
+R4SAS
+RN
+RN_
+T3s|4
+Xeha
+orignal
FreeRider
Irc2PGuest46029
Onn4l7h
Onn4|7h
T3s|4_
aargh3
acetone_
anon4
cancername
eyedeekay_bnc
not_bob_afk
profetikla
shiver_
u5657_1
weko_
x74a6
orignal
implemted for MNcW
zzz
is it big enough for 2 fragments?
orignal
not yet
orignal
15000
orignal
1500
orignal
but I will add more
zzz
before compression though?
zzz
ok
orignal
yes size of router.info
orignal
two fragments now
orignal
1712 bytes RI
orignal
⇒ -Akx: [2a02:180:2:92:92:92:11:15] [2064:3096]
orignal
just connected to you from MNcW
orignal
please check for errors with fragments
R4SAS
restarted YXEA and gpUB with latest commits
zzz
MNcW no errors, looks good
zzz
R4SAS, orignal, still no MTU advertised on gpUB HE address, lots of errors last 24 hours
zzz
what's the real MTU?
zzz
errors last 21 hours:
zzz
11 sQ8dVd
zzz
10 gpUBQf
zzz
9 ~GIB3b
zzz
3 p~8-FO
zzz
2 2RRYXk
zzz
1 YXEAXl
zzz
1 xZ9nsA
zzz
1 tXc6dH
zzz
1 oBZUnh
zzz
1 bAU~6X
zzz
1 3PYq0v
R4SAS
zzz: how I can find mtu in RI?
zzz
it's in each SSU/SSU2 address
zzz
if not the default
zzz
what's the real interface MTU?
R4SAS
look... I have HE tunnel on one machine, there server-HE tunnel has MTU 1480
zzz
ok
R4SAS
gpUB sits on another server, to which I routing subnet, and that tunnel also has 1480
R4SAS
on both sides
zzz
ok
R4SAS
so everywhere 1480
zzz
then your address should have mtu=1480, it's a bug
R4SAS
hmmmm
R4SAS
all i2pd's with SSU2 on first service died
R4SAS
hmm... looking for dumps
R4SAS
first server*
zzz
lol use in-i2p hostname, eat your own dogfood :)
R4SAS
hah
R4SAS
s/https:\/\/paste.i2pd.xyz/http:\/\/privatebin.i2p/e
R4SAS
that's just backtrace
R4SAS
orignal: bit different, but same place: privatebin.i2p/?969d1bd233ceabda#GpGxSXAhJNSm8mdFQwF2azeJ48uPngKi1xKba7sYRaSa
orignal
I have fixed
orignal
zzz, question for you
orignal
how can you send a temination reason like "invalid S" if you don't know an address ans i?
orignal
R4SAS you need to delete router.info
orignal
to start advertisign mtu
R4SAS
orignal: already tried, nothing appears
orignal
2RRY still runs older version
orignal
it should apper for ipv6 only
R4SAS
look yourself: s4.i2pd.xyz/files/router.info
zzz
orignal, you can't, many of those reasons are impossible
zzz
I just copied them all from NTCP2
orignal
lol
orignal
that's what my mistake
orignal
R4SAS so it recongnizes ipv6 address but not mtu
orignal
please look at that code yourself in RouterContext.cpp
orignal
and try again
R4SAS
I restarted with deletion of router.info just 15 minutes ago
orignal
as I said print more logs into UpdateAddress
R4SAS
hm, ok
orignal
and check
orignal
why it don't set mtu
orignal
zzz but the source of the problem that somebody tries to connect to me with wrong S
orignal
e.g. I can't find it thier RI
zzz
and you're crashing?
orignal
yes, because I was trying to send termination with that reason ))
zzz
lol
orignal
the question is why they do ut
zzz
might be old java... possibly at startup... I fixed some problems with that a month or two ago
zzz
you have an IP for them I can look up?
orignal
no
orignal
I saw it in the core dump only
zzz
I haven't seen that error from anybody in a long time, at least a month
orignal
look
orignal
I have made this change a wekk ago
orignal
but it started this night
R4SAS
orignal: maybe because you sed MTU only for SSU, not SSU2 in UpdateAddress?
R4SAS
set*
R4SAS
if (address->ssu) address->ssu->mtu = mtu;
orignal
yes, but SSU2 also has ssu portion
R4SAS
hm
orignal
add logs and see where it fails
R4SAS
adding now
orignal
zzz also it seems it happens with v4 only
zzz
orignal, I'll keep an eye out for any missing/wrong S, will let you know if I see any
orignal
will do
R4SAS
orignal: Router: UpdateAddress: processing 127.0.0.1
R4SAS
nothing more
R4SAS
that's inside if (address->host != host && address->IsCompatible (host) && !i2p::util::net::IsYggdrasilAddress (address->host))
orignal
then where ipv6 comes from
orignal
from address6 ?
R4SAS
yes
orignal
that's why
R4SAS
haha
orignal
host is not updated
orignal
comment out address6
R4SAS
now mtu is added
R4SAS
correct one, 1480
zzz
orignal, good catch, I'm checking for missing s but not checking for s mismatch!!! will add it
orignal
I do it differently
orignal
I find as address by S
orignal
assuming every address might have own S
R4SAS
orignal: will you add MTU detect for address6?
orignal
let me think how
orignal
will do
orignal
that's eary
orignal
if (!ssu->mtu)
R4SAS
asking because now that router publishing address for tunnel's endpoint address, not routed one
zzz
actually, you don't need an "i" to send that reason code in a termination; you would do it in the data phase
zzz
but probably not worth the trouble. I don't do it.
orignal
what message do I send?
orignal
temination is a block
orignal
not a message
orignal
simply speaking I don't have a message to send without address
zzz
right. but in theory you could just look for a v4 or v6 address, enter data phase, and send data pkt with termination block.
zzz
but it's not worth the trouble
orignal
but I still need i
zzz
no you don't, the data phase uses different keys
zzz
never mind, you're right
zzz
here's another problem from a few days ago:
zzz
07/14 16:12:43.385 INFO [ handler 1/1] outer.transport.udp.PeerState2: Got TERMINATION block, reason: 2 count: 241 on 205.178.177.93:26464 Cznve~
zzz
00000000 1b 88 9d b3 ec a2 e8 ac 00 00 00 ff 06 00 00 00 |................|
zzz
00000010 0c 00 05 00 00 00 f1 f1 06 00 09 00 00 00 00 00 |................|
zzz
00000020 00 00 f1 02 00 00 04 62 d0 24 5b fe 00 09 00 00 |.......b.$[.....|
zzz
00000030 00 00 00 00 00 00 00 69 35 a6 cd bc 4c 8d 11 a3 |.......i5...L...|
zzz
00000040 8d bd b4 cd 63 1b 6c |....c.l |
zzz
java.security.GeneralSecurityException: Data payload error
zzz
Caused by: java.io.IOException: Illegal block after termination: 0
zzz
4 blocks: ack, termination, datetime, padding
zzz
java router
zzz
I need to fix it
orignal
yes it must be before padding
dr|z3d
keeping zzz on his toes, orignal :)
zzz
orignal, more ack block problems:
orignal
what kind?
zzz
07-17 15:44:23.812 DEBUG [ handler 1/1] outer.transport.udp.PeerState2: Got new ACK block from k8vhnd ACK 2199 NACK 2198-1944 NACK 1943 ACK 1942-1688 ACK 1687-1672 NACK 1671 ACK 1670-1416
zzz
07-17 15:44:23.998 DEBUG [ handler 1/1] outer.transport.udp.PeerState2: Got new ACK block from k8vhnd ACK 2200 NACK 2199-1945 NACK 1944 ACK 1943-1689 ACK 1688-1672 NACK 1671 ACK 1670-1416
zzz
07-17 15:44:24.089 DEBUG [ handler 1/1] outer.transport.udp.PeerState2: Got new ACK block from k8vhnd ACK 2202 NACK 2201-1947 NACK 1946 ACK 1945-1691 ACK 1690-1672 NACK 1671 ACK 1670-1416
zzz
07-17 15:44:24.315 DEBUG [ handler 1/1] outer.transport.udp.PeerState2: Got new ACK block from k8vhnd ACK 2203 NACK 2202-1948 NACK 1947 ACK 1946-1692 ACK 1691-1672 NACK 1671 ACK 1670-1416
zzz
yin first line you ack 2199
zzz
after that you're nacking it
orignal
good ctach, will check
orignal
what is acnt? zero?
zzz
correct
orignal
must be 1 I guess
orignal
it this case everything will be correct
orignal
looks like I never send acnt of 1
orignal
*** afk ***
orignal
maybe it's older router?
orignal
because I don't see this problem in current code
zzz
ok
orignal
do you see mtu in this RI?
zzz
k8vhnd
orignal
will chek
orignal
on FF
zzz
1420
orignal
then I need to investigate more
orignal
becuase it's new
zzz
one sec and I'll pull the whole log
orignal
stange that acnt is zero
orignal
looks like Orion's again ))
orignal
Kirov
zzz
here's every non-dup ack I got for that session:
zzz
so you can look for where it went wrong
orignal
thanks
orignal
let me download
zzz
looks to me like the first bad one is ACK 990
zzz
that's when the acks of 989-734 turned into nacks
zzz
ant it was mostly bad from there to ACK 2215
orignal
do you see non-zero acnt?
zzz
it was all normal until 990
orignal
let me investiagte
zzz
from there to 1091 it was acnt = 0
zzz
then it kindof got normal again for a while
zzz
then messed up again starting at 1459
orignal
990-734 = 256 )))
orignal
and acnt is 255 max
orignal
and 256 makes acnt 0
orignal
will fix it
zzz
yeah you're not skipping the nacks field
orignal
it's clear now
zzz
if you want to ack 1000-0 you have to do thru=1000 acnt=255 and then ranges 0 255 0 255 0 255
orignal
yes I know
orignal
that's what I tried to do
orignal
maybe I flipped nack and ack fields
orignal
anyway let me fix
orignal
well I just don't consider this case yet
orignal
implementing
orignal
when acnt > 255
orignal
implemneted
R4SAS
looks like HE.net POP in France is died...
orignal
dead?
orignal
I use in New York only
R4SAS
I some strange route via NL
orignal
war never changes ))
R4SAS
that is on dedicated server in Online SAS
R4SAS
(Scaleway(
R4SAS
)*
orignal
Europe falls
R4SAS
tserv10.par1 - Paris, FR - Down
R4SAS
haha