{% extends "_layout.html" %} {% block title %}I2P Development Meeting 141{% endblock %} {% block content %}

I2P dev meeting, August 2, 2005

13:53 < jrandom2p> ok, as i'm here, is there anyone interested in having a brief meeting wrt the notes (or something else)?

13:54 < jrandom2p> anything in the notes people are concerned with, thoughts not related to 'em that people want to bring up, or other issues relevent and timely?

13:54 <@smeghead> sure

13:54 <+protokol> is icepick here?

13:55 <+protokol> i am wondering if i2p-mnet is testable yet and/or an ETA on it

13:55 < jrandom2p> idle for 9 hours atm..

13:56 < jrandom2p> from the channel logs, it didnt sound workable, but he did get the basic SAM integration going

13:56 < jrandom2p> i'm sure we'll hear more when there's more to hear

13:56 <+protokol> cooool

13:57 < jrandom2p> smeghead: has -1 fixed your port migration issue?

13:57 <@smeghead> i haven't noticed any funny business

13:58 <@smeghead> in 3 days or so

13:58 <@cervantes> glad to say I haven't had a loss of service for a day or two

13:58 <@smeghead> i think i can call it fixed

13:58 < jrandom2p> wr0d

13:58 < jrandom2p> (^2)

13:59 <@cervantes> and thetower is only reconnecting every 4 minutes now...so the network health in general must be improving

13:59 < jrandom2p> heh

13:59 <+thetower> A fresh install seemed to fix the problem, but it was really quite disturbing and I never could find a good reason for it.

14:00 < jrandom2p> hmm

14:00 < jrandom2p> was it irc only, or were you losing many peers?

14:00 <@cervantes> gremlins

14:01 <+thetower> Is it possible that changing the router.config file without restarting i2p would have caused the crashes?

14:01 < jrandom2p> hmm, no, i change router.config often

14:01 < jrandom2p> or, is there a particular change you're concerned with?

14:02 <@cervantes> I remember copying over my jbigi lib once while the router was still running.... THAT caused problems ;-)

14:02 <+thetower> I set up some script to alter the bandwidth limits based on current network usage and I was wondering if that was causing the problem.

14:02 < jrandom2p> heh yeah cervantes, that'll always kill the router

14:03 < jrandom2p> ah ok, no, that shouldnt be a problem... though... if it altered the limits to be too small for messages to get through...

14:04 <+thetower> Well, it had fairly reasonable lower limits so I guess that wasn't it.

14:04 < jrandom2p> ok cool, just checkin~ :)

14:05 < jrandom2p> i suppose we'll have 0.6.0.1 tomorrow then, as -1 seems to be a pretty good improvement

14:05 < jrandom2p> it'll be backwards compat, etc, yadda yadda.

14:06 < jrandom2p> anything else y'all know that needs to get pushed out there?

14:06 < jrandom2p> whats the status with i2phex?

14:06 <@smeghead> maybe push the cvs hosts.txt to dev.i2p.net... the current one is months old

14:06 < jrandom2p> i did the other night iirc

14:07 <@smeghead> sirup hasn't been around in a couple of weeks

14:07 < jrandom2p> ooh, hmmm..

14:07 <@smeghead> it's summer though

14:07 <@smeghead> maybe on holiday or something

14:08 <@cervantes> or he's been bum-raped by the riaa

14:08 < jrandom2p> ah yeah, its up there (it was just cached on squid.i2p)

14:08 <@smeghead> riaaped?

14:09 < jrandom2p> ($Id: meeting141.html,v 1.2 2005-08-04 16:21:39 duck Exp $)

14:09 < jrandom2p> *cough*

14:09 <+bar> there are some things that need to be added to bugzilla, like i2p 0.6 and java 1.5

14:09 <@smeghead> ok

14:09 < jrandom2p> ah right, yeah i still havent gotten my laptop online yet (grr)

14:10 < jrandom2p> ((the weekly status notes needed to be burnt to cd... a 1KB cd...))

14:10 < jrandom2p> woah heya mihi

14:10 <@duck> hi mihi!

14:10 < mihi> hi all :)

14:10 <@cervantes> could be dm :)

14:10 < jrandom2p> heh

14:10 <@smeghead> indeed

14:10 <@cervantes> 'lo mihi

14:10 < mihi> seemed to require a bit of tweaking in the config file till my router believed that *only* 8887/udp is open...

14:11 * jrandom2p mentioned i2ptunnel in the status notes and mihi appears ;)

14:11 < jrandom2p> ah, hmm, the i2np.udp.fixedPort=true thing?

14:11 < mihi> hmm? was it there?

14:11 * mihi read status notes only quickly

14:11 < mihi> hmm... is that better solution?

14:12 * mihi just reset the port to 8887 and restarted hard until it did not change the port...

14:12 < jrandom2p> whats the tweak you did to your router.config to make it believe only 8886?

14:12 < jrandom2p> er, 8887

14:12 < jrandom2p> hah

14:12 <@cervantes> can we perhaps rename I2PTunnel as you suggested to something like I2PProxy...?

14:12 < jrandom2p> ok, yeah, use i2np.udp.fixedPort=true

14:12 < jrandom2p> (deployed in 0.6-1 and to be released asap as 0.6.0.1)

14:12 <@cervantes> it can get very confusing talking about "the tunnel config page"

14:13 <+thetower> Oh I have a question, isn't i2p supposed to automatically detect which udp port to use? And if so, is it supposed to be hard coded in the default router.config?

14:13 < mihi> hmmkay...

14:14 < mihi> seems that i2p changed the port once again

14:14 < mihi> expect me to be away soon :)

14:14 < jrandom2p> thetower: yes, it should automatically detect, but there are some funky tap dances that we~re going through at the moment

14:14 <@cervantes> mihi: d'you have the latest cvs?

14:14 < jrandom2p> thats what the whole PeerTest thing is about (making it so that we always automatically configure it properly)

14:14 < mihi> nope.

14:14 <@cervantes> mihi: that would be why then :)

14:15 < mihi> only the version from i2pupdate.zip

14:15 <@cervantes> mihi: 0.6 has RandomPort (tm) functionality

14:15 < jrandom2p> heh

14:16 <@cervantes> :)

14:16 <+ant> * mihi 'd like FixedPorto functionality :)

14:16 <+ant> and disconnected...

14:16 <@cervantes> then you'd need 0.6-1 FixedPort Pro

14:16 < jrandom2p> heh

14:16 < jrandom2p> ok, anyone else have something to bring up for the meeting?

14:16 <@cervantes> or wait for 0.6.0.1

14:17 < jrandom2p> how has the latency/throughput been, barring the intermittent reachability?

14:17 <+ant> hmm. here is a cvs checkout from 2004-10-06. should try to update it :)

14:17 < jrandom2p> !thwap mihi

14:18 <@cervantes> I got i2pinstall.jar at 110k/sec from dev.i2p yesterday on a single stream

14:18 < jrandom2p> nice

14:19 <@cervantes> and 320k/sec using multiple

14:19 < jrandom2p> w0ah

14:19 < jrandom2p> 0hop, i assume

14:19 < jrandom2p> (dev.i2p is 0hop)

14:19 <@cervantes> yup

14:19 < jrandom2p> ((in case you couldn't tell ;)

14:19 <@cervantes> ;-)

14:19 <+thetower> download to: GTA San Andreas

14:19 <+thetower> download rate: 28.51 kB/s

14:20 <@cervantes> that was from multiple sources though...

14:20 < jrandom2p> ah cool thetower

14:20 <@cervantes> managed to push squid.i2p up to about 280

14:21 < lucky> jrandom2p :)

14:21 < lucky> would you push the new hosts.txt to the site

14:21 <@cervantes> lucky: tis done

14:21 < jrandom2p> yeah, once we can consistently pull that sort of rate cervantes, we'll need to add on some configurable delays to let people do 0hops safely

14:22 < jrandom2p> (so it delays AVG(tunnelTestTime/2) but doesnt waste bw or lose messages)

14:22 <@cervantes> to hide the fact that it's a 0 hop tunnel?

14:22 < lucky> i wonder if I2P will ever have speeds decent enough tha ti could let people log into my virtu-vax

14:23 < jrandom2p> yeah. otherwise, if you say "hey i~m getting 300KBps from your site", you can pretty safely guess that its 2 0hop tunnels

14:23 < jrandom2p> (otoh, 1 to 2 to 3 to 4hops don't have such a dramatic cut)

14:23 <@cervantes> so will i2p effectively have a bandwidth cap

14:23 < jrandom2p> ((as once you force true tunnel operation, each intermediate hop isn't much))

14:24 < jrandom2p> nah cervantes, large windows + delays

14:24 * cervantes cancels his plans for HDTV streaming anonymous pr0n

14:24 < jrandom2p> you can just have more messages in the air to get the same rate

14:25 <@cervantes> ah right

14:25 < jrandom2p> (but it'll take a few more rtts to get to the larger window, of course)

14:25 < jrandom2p> ok, anyone have anything else to bring up?

14:26 < mihi> bring up a *baf*er :)

14:26 <@cervantes> it's gone rusty with missuse

14:27 < jrandom2p> heh i suppose its time ;)

14:27 * jrandom2p winds up

14:27 * jrandom2p *baf*s the meeting closed

{% endblock %}