17:00:20 #startmeeting Network team meeting, 8th november 2021 17:00:20 Meeting started Mon Nov 8 17:00:20 2021 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:20 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:23 hello hello 17:00:42 hi there! 17:00:49 * ahf waits for a matrix pop-up to go away soe he can click on the right tab 17:01:00 our pad is at https://pad.riseup.net/p/tor-netteam-2021.1-keep 17:01:14 thanks for all the help last week with turning the updates into the forum post! 17:01:51 how are folks' boards doing: https://gitlab.torproject.org/groups/tpo/core/-/boards ? 17:02:00 * ahf waits for a few more to join 17:02:08 o/ 17:02:24 o/ 17:02:26 dgoulet, eta: meeting ping :) 17:02:37 yes? 17:02:48 ah! wasn't sure you were here 17:02:49 o/ 17:02:57 sorry 17:02:58 o/ 17:03:06 how are folks' boards doing: https://gitlab.torproject.org/groups/tpo/core/-/boards ? 17:03:06 nickm, (cheers) 17:03:12 :) 17:03:55 * eta actually has things in hers now! 17:04:30 my board is legit 17:04:45 same here 17:04:50 Mine is fine; I spent a little time cleaning out my icebox. 17:05:30 mine is also OK.. now 17:05:36 have one annoying ticket in doing 17:05:38 excellent 17:06:08 we have our check-in on the release situation. we are still merging some smaller things to main right now, but also holding a few things back 17:06:17 we have a discussion item later by nickm about 0.4.7 17:06:43 we can sort of take that now as this is technically the last bullet on the must do part of the meeting before the s61 part kicks in 17:06:57 ok. 17:07:00 2021-11-08 [nickm] Is it time to revisit the 0.4.7 release plan? i.e., aim to stabilize what we have, then have S69 go in 048, which would release in late Q1 2022? 17:07:06 s69 = s61 ? 17:07:10 sorry, yes 17:07:28 so what is your thoughts here? i thought we had a plan here a week ago :-) 17:07:44 david and i talked with mike when he expects to have the next round of things that needs to go in for next alpha release ready 17:08:05 what will that include? And when does it get 047 stable out the door? :) 17:09:16 i don't think there is a fixed date yet. it sounded like the CC things were ready quite soon 17:10:07 okay. So, the situation I think we should avoid is delaying 047 to the point where we could have done two releases had we delayed CC. 17:10:10 mikeperry: correct me if i am wrong, but it's the negotiation part that needed to land right? 17:10:21 right 17:10:37 ahf: yes, which I am currently testing in chutney 17:11:07 nickm: right. i don't think there is a plan for not doing CC in 0.4.7 because parts of the roadmap on that depends on it going out in reality for feedback 17:11:21 so i think waiting for the changes here while the simulation work is ongoing is acceptable 17:11:37 I think so too, but: 17:12:28 our regular schedule has been to put out release approximately every 4-5 montsh 17:12:49 *months 17:12:52 right 17:13:20 If we freeze 047 in November, we get it out in Jan/Feb. That's probably fine. 17:13:39 i think even if we freeze a bit into december that will be fine too 17:14:02 but the part that i don't get is that we had a meeting 1.5-2m ago where we decided that releases for tor.git can be way more flexible now and we also agreed that the *big* thing in 0.4.7 is CC 17:14:04 If we freeze in March, though, we don't get it out till may 17:14:08 oh, i agree 17:14:14 it should be in before, i agree on that 17:14:24 we are not moving to a -current format :-) 17:15:01 if we are frozen before we go on holiday i think we'll be good 17:15:11 if we are not frozen before we go on holiday, then we have a bit of a problem probably 17:15:46 mikeperry, dgoulet: IYO is "freeze before holidy" more of a "definitely will happen", "will happen with a lot of effort", or something else? 17:16:45 I am confident exit-side negotiation will be done by then, yes 17:17:12 Done, and merged, and nothing else has to go in 047? 17:17:56 yes negotiation is the last piece 17:17:58 by nothing else you mean features right? because we merge bugfixes post stable release 17:18:05 right 17:18:11 and i am quite sure we'll have bugs to fix 17:19:24 was that all the information for this discussion point? :o 17:19:41 I think so. Should I update the wiki accordingly? 17:20:22 sure, freeze ~15 december? 17:20:43 sure! 17:20:53 cool, then we have a week to deal with fire if anything 17:21:20 I've given us 3 months for release, per usual. 17:21:28 yeah, i think that is fine 17:22:00 ok 17:22:05 thanks, and sorry for the diversion! 17:22:08 i think that was all discussion items and there are no announcements here 17:22:14 no worries, good to clear things up 17:22:20 mikeperry: the s61 floor is yours 17:23:34 yeah so the main thing is I have been working on exit-side negotiation. it sometimes works in chutney, but sometimes it hits cases where parts of the code still check/verify the originsl ntor handshake type 17:23:44 so I have to hunt the rest of those down 17:24:49 I also altered nicks mockup branch to only send ntorv3 to exits rather than every hop 17:25:03 and right now I only bump one protover instead of two 17:25:19 nice 17:25:48 we're also fine tuning the shadow gitlab situation. some minor struggles there 17:26:16 good. you're getting whatever help is needed there or is it all something you can handle yourself? sounded like we got some more hw for the runners last week? 17:26:19 hiro is finishing up a graphing piece, and then with that I think we will be ready tun run a lot more sims in parallel 17:26:27 nice \o/ 17:26:36 * dgoulet is doing the HS part of negotiation also 17:27:00 \o/ 17:27:03 (please put me at some point in the review chain on these negotiation patches before they merge?) 17:27:31 jnewsome: are there still pain points in gitlab storage, etc? I think we worked around most of the blockers yeah? 17:27:48 jnewsome: or is it still possible we randomly run out of storage? 17:28:16 mikeperry: I think we're basically good now. eventually we'll fill up the disk on the runner, but I think we have on the order of 100 GB to go 17:28:31 i'll start looking into some form of garbage collection, hopefully before we hit that 17:29:22 391 GB to go 17:30:01 are we able to manually prune runs we don't care about? 17:30:51 if it's artifacts or CI things in gitlab i think they can be pruned from the API - i don't think docker images can be pruned via the api 17:31:10 not super easily at the moment. on the host machine the files are owned by root. I'm currently making them read-only from inside the CI; maybe I can +w the directories 17:31:48 ahf: we're using a docker volume for the bulk storage 17:32:05 since we kept hitting the limit and crashing and burning with artifacts 17:32:24 ah 17:32:28 smart 17:34:17 how is the sbws thing going w/ the exit pinning? 17:34:33 juga is making progress 17:34:50 they can't make this meeting, alas 17:34:54 I have not checked in on that this past week, but I saw juga and dgoulet on irc talking about it 17:35:09 but as far as i understood chances are good they'll get it working soon 17:35:09 I recall juga being finally able to make that .exit sitaution work! 17:35:15 yeah 17:35:23 there are some tests that still need to get done 17:35:33 but it looks pretty much promising 17:36:25 ok nice 17:36:43 anything else to cover? 17:36:54 * ahf good 17:37:02 i am good 17:37:07 okay with me 17:38:29 aight I think we're good then 17:38:29 shall we end the meeting? 17:38:32 perfect 17:38:35 thanks everybody o/ 17:38:37 #endmeeting