17:04:57 #startmeeting Network team meeting, 20 March 2023 17:04:57 Meeting started Mon Mar 20 17:04:57 2023 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:04:57 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:04:58 hello hello 17:05:02 sorry for the delay 17:05:05 https://pad.riseup.net/p/tor-netteam-2023.1-keep is our pad 17:05:08 no problem! 17:05:13 Welcome :-) 17:05:23 thank you 17:05:29 how are folks doing with their boards? 17:05:38 https://gitlab.torproject.org/groups/tpo/core/-/boards 17:05:47 (It's a good sign about your usual punctuality when people start asking questions when you're 2 minutes late...) 17:05:48 IHNI, I am catching up on todos and reviews and ffi/rpc discussions 17:06:20 nickm: haha, yeah - i even got a text from micah when i was trying to get my mobile modem to boot 17:06:56 i have a todo item to make my kanban reflect reality... so um. :) 17:07:04 lol 17:07:41 okay, i don't spot anything entirely off 17:07:59 dgoulet: anything on C tor we need to think of? 17:08:18 not at the moment ... we are merging and Conflux-ing 17:08:25 no release anytime soon 17:08:35 \o/ 17:08:36 very nice 17:08:50 there is a small announcement from Diziet: [Diziet 2023-03-20] Ian will be AFK fromk ~1700 UTC tomorrow 21.3.23 17:08:55 thanks for heads up there 17:09:00 no discussions 17:09:07 mikeperry: you want to move to s61? 17:09:08 That's the Arti public bbb I think. 17:09:11 ok 17:09:20 So, err, we might want to reschedule it ? 17:09:22 small one for me: I'll be afk wednesday morning for a doctors appt 17:09:32 Diziet: ah, right, those two collides there - i am not sure if we want to reschedule when the meeting is public :-/ 17:09:39 Yes. 17:09:42 Diziet: maybe you and nick can do a checkin after if needed? 17:09:52 Sorry for the short notice. I only found out middle of last week when I was AFK 17:10:07 ahf: Yes, that would be a good idea. 17:10:38 maybe on thurs or next tues, diziet gabi and I can talk about next directions in arti onion services? 17:10:39 excellent, you and nick can talk after the meeting to see if there was anything there 17:10:46 nickm: sounds good 17:10:50 sgtm 17:10:52 nickm: Sure 17:10:57 mikeperry: go loco 17:11:10 so last week dgoulet and I spent quite a bit of time investigating a weird protocol interaction between ssh and conflux, where scp transfers could stall by not sending a success message.. on the way, we fixed a couple bugs, and that improved things, but we need to do more investigation this week. we've also got an assert to fix from a shadow run 17:11:45 cool 17:11:51 so most likely we will be deep in that. goal is to get full shadow runs completing asap 17:13:08 mikeperry: I successfully reserved a chameleon cloud runner for next week since that's now the planned move time for our "verylarge" runner 17:13:17 but other than the stalls and that assert, stability is looking good. we've got test coverage of the most sensitive bits, and plans to add some more.. integration tests, chutney, asan, etc all green 17:13:31 jnewsome: ah I did not see that update yet. thank you 17:14:41 dgoulet: I should be able to work on that assert fix after this meeting, but maybe should run it by you on bbb or sth 17:14:50 excellent 17:14:57 lets BBB after this 17:15:47 juga,GeKo: how are things in onbrisca and onbasca testing land? do we have the ability to look at the sets of bridges below the ratio to see if they are stable, for onbrisca? 17:16:08 GeKo: around?, want to answer?\ 17:16:49 so far looks like only 2% of the bridges measuared are below the ratio 17:17:05 meskio is going to up the threshold 17:17:12 that's good. checking to see if they change a lot or not might be a good plan 17:17:47 percentage is not changing much, do you mean to look at individual bridges 17:17:48 ? 17:18:12 so far we don't have some automatic way, would need to look at the logs, but that's doable 17:18:17 yeah, like to see if it's always the same ones in that 2%, or if they vary wildly (perhaps due to bugs, or slow exits, etc) 17:18:27 go it 17:18:31 got it 17:18:54 that would be a first step before raising the ratio, I think 17:19:11 making sure we can at least get rid of the really slow ones reliably would be a win by itself 17:19:24 ok, i'll tell meskio to wait, he was going to wait to this meeting anyway 17:19:33 ok great 17:19:44 and look at which ones are under ratio first 17:20:24 ok. I am still eyeing https://gitlab.torproject.org/tpo/network-health/analysis/-/issues/49, esp since the ddos let up last week.. but it is sputtering again now.. so we'll see 17:20:47 I think that is it for the meeting, unless there are other questions? 17:20:56 * juga is good 17:21:05 very nice - i don't have anything else here 17:22:01 sounds like everybody else are good too 17:22:03 thanks all o/ 17:22:04 #endmeeting