16:58:46 #startmeeting Network team meeting, 25th July 2022 16:58:46 Meeting started Mon Jul 25 16:58:46 2022 UTC. The chair is micah. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:58:46 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:58:52 hihi! 16:59:04 o/ 16:59:05 o/ 16:59:07 Hi everyone, welcome to the network team weekly meeting! I'm your guest facilitator while ahf is off hacking. 16:59:09 o/ 16:59:45 Apologies in advance if I deviate from what you are used to in these meetings, I'll try to do my best ahf impression 16:59:58 o/ 17:00:01 As a reminder, the meeting pad is https://pad.riseup.net/p/tor-netteam-2022.1-keep 17:00:20 As it says there, if you have updates, please enter them on the pad, under your name. If you have things to talk about, please make them bold. 17:01:21 I don't see any announcements in the pad, so if you have any announcements to make, now would be the time! 17:01:36 reminder that there's no work thu and fri 17:02:16 also I have a dentist appointment tomorrow, shouldn't interfere with anything 17:02:17 indeed! thanks for that nickm - 28th and 29th are TPI holidays 17:03:27 last time nickm made the announcement that he will be AFK August 18-22nd, so just re-reminding about that 17:04:13 and I believe ahf will back on the 15th of August 17:05:55 ok, I don't see any updates from folks, or specific agenda items. The updates of what folks worked on in the last week and what is planned the next week is not something you folks do at these meetings? 17:06:38 yeah, we do that via the issue board 17:07:15 ok, then we should switch to the boards, does anyone have things to add to the boards: https://gitlab.torproject.org/groups/tpo/core/-/boards ? 17:08:07 I think we're okay. Next monday we should go over the arti 1.0.0 milestone and see where we're at, since we're hoping to turn it in on 1 September 17:09:34 ok, great, I'll add that to the pad for next week 17:10:00 dgoulet: does it make sense to talk about release status this week? 17:10:05 yes 17:10:36 so we have a couple ticket that we believe will help with the DoS and so we need to release soon those 17:10:38 arti 1.0.0> scary 17:10:40 _and_ the fallbackdirs 17:11:00 and so I expect a release soon here in at least all maintained versions 17:11:02 -- 17:11:06 dgoulet: do you want me to work on the OOS thing on any particular timeline? Please let me know if so 17:11:21 dgoulet: do you have specific tickets you need people to review to move that forwards? 17:11:57 https://gitlab.torproject.org/tpo/core/tor/-/issues/40383 i'm guessing 17:12:21 but perhaps there is one related to the memory consumption, or other DoS issues? 17:12:36 no not that one 17:12:46 so 1 ticket is confidential... a couple others need to be opened 17:13:04 another one has been reviewed by nickm 17:13:19 ok, you will add them to the release milestone when you've opened them? 17:13:20 so once I have those, I'll get reviewers to help likely a combination of mikeperry and nickm 17:13:23 yes 17:13:34 excellent. can the confidential one be added there, or does that not work 17:13:38 please let me know if there's anything I should hack on there too. 17:13:46 nickm: sure thanks! 17:13:54 nickm: as for OOS, I don't see it as urgent for now no 17:14:05 as a data point 17:14:07 ok 17:14:17 I'm still heavily investigating the gigantic memory pressure our relays are seeing 17:14:21 I am still looking at the confidential one.. it may take a few more sims to get right. might not be worth blocking the releases if we think https://gitlab.torproject.org/tpo/core/tor/-/issues/40623 will help things 17:14:53 mikeperry: if we can get reasonable data in the coming days, I think we can wait but if you say like weeks, maybe we can skip indeed 17:14:55 well there are two confidential ones 17:15:17 I can do one of them quickly, but it is unlikely to be involved. the other will require more sims 17:15:34 I can discuss via bbb after if you like 17:15:38 yeah lets do that 17:16:13 ok 17:16:28 great, perfect... doing a break-out after the meeting is great, so we don't suck everyone into something only a few need to dig into 17:16:35 +1 17:17:10 ok, besides these DoS/OOS issues (to be created) and the confidential issues, are there other issues that are being tracked for this release, or just those? 17:17:43 i dont see any on the milestone, so perhaps its just these? 17:18:17 iirc, no outstanding ones 17:18:32 quick look confirms it... some I need to work on but not sure I'll make it for this needed release 17:18:48 sounds good 17:19:57 one of the things ahf does is to check if there are any updates in issues in other teams (anti-censorship and network health) that might need net-team help 17:20:37 I'm unsure how he identifies those, although I do have a link into their issues: 17:20:38 https://gitlab.torproject.org/groups/tpo/core/-/issues?scope=all&utf8=%E2%9C%93&state=opened&label_name[]=For%20Anticensorship%20Team 17:21:13 If one of the other teams has said that one is particularly important, then he usually flags it for us 17:21:14 do those issues that need net-team input have a 'For Network Team' label? 17:21:31 usually yes 17:21:44 yah, one is pending that they need which is on my short list 17:21:51 but likely won't be in this release :S 17:22:02 https://gitlab.torproject.org/tpo/core/tor/-/issues/40635 17:22:27 ok, its also in the Icebox 17:22:59 this is the network health list of issues: https://gitlab.torproject.org/groups/tpo/core/-/issues?scope=all&utf8=%E2%9C%93&state=opened&label_name[]=For%20Network%20Health%20Team 17:23:37 ok, I'm not aware of any flagged issues from either team, so we can move on from there 17:24:07 it seems the next item here on the agenda is to remind people of a few things that you are probably very much already aware of: 17:24:11 * Remember to "/me status: foo" at least once daily. 17:24:15 * Remember that our current code reviews should be done by end-of-week. 17:24:18 * Make sure you are in touch with everybody with whom you are doing work for the next releases. 17:24:20 and finally: 17:24:43 * Check other's people call for help in their issues: please remember to help volunteers ! 17:25:05 any other reminders that folks want to pass on? 17:25:38 * dgoulet has none 17:25:39 if you do, please go ahead and say them :D 17:26:06 after reminders, it appears to be discussion items 17:26:15 there are none listed on the pad currently 17:26:46 so if you have anything that needs to be discussed, we can open it up here for that 17:27:58 if there is silence, then I take it there are no discussions that need to happen, and we can close the meeting 17:28:09 we got s61 stuff 17:28:26 go ahead on s61 mikeperry 17:29:18 ok. so hiro reduced the guard rotation for the op6 'a' instances so they match op7a. thanks hiro 17:29:58 hiro: wrt the Guard+Fast stuff, I have not looked deeply that that yet, since this OOMing problem is impacting Guards.. we're actually seeing relays lose the guard flag now :/ 17:30:10 so that is being back burnered 17:30:29 jim is out on vacation 17:30:36 Ok 17:31:43 juga: I am doing my best to follow https://gitlab.torproject.org/tpo/core/tor/-/issues/40623. I guess the rough plan can be to test it locally, and then I can review it when you think its ready for that. not sure I will be as good as geko at review but I can try 17:32:32 mikeperry: ok, i'll let you know when i think is ready 17:32:32 then I think we can test it on gabelmoo and/or longclaw 17:32:42 yup 17:32:58 juga: one thing to consider is if bwauths don't look at their logs, we may want to put this XOFF status in the bwfiles so we can see it 17:33:15 ah, yes 17:33:26 gabelmoo does 17:33:40 (have emails on warnings) 17:33:58 farahavar may not get back to us, tho. so we may have to resort to the files for it, long-term 17:34:08 ok 17:35:05 my stuff I'll be working on, I will mostly be chatting with dgoulet on bbb about that 17:36:20 great, thanks mikeperry for the S61 stuff 17:36:42 eta: just as an fyi, one of the things I am working on may change the TOR_VEGAS algs a bit. I will cc you with that after some sim results, and will update the spec 17:37:16 I think that's it 17:37:44 excellent, anyone else have anything else to update folks on? 17:37:51 mikeperry: ack, thanks 17:38:14 mikeperry: also, are there test vectors available for Vegas? I think the branch I was using didn't have any yet 17:38:43 eta: yeah not yet, because I was concerned there might be something like this. so the test vectors won't come until I figure this piece out 17:40:33 ok, great. if there isn't further discussion points, I'll thank everyone for coming and putting up with the substitute ahf. As a reminder if you have questions on roadmap/team related things feel free to poke me, i'm happy to talk about anything 17:41:11 and am available, so just ask! 17:41:41 with that we'll do... 17:41:42 #endmeeting