16:58:32 #startmeeting network team meeting, October 15 16:58:32 Meeting started Mon Oct 15 16:58:32 2018 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:58:32 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:58:36 Hi folks! 16:58:54 Hi! 16:59:07 hi 16:59:13 https://pad.riseup.net/p/tor-netteam-2018.1-keep is the pad 16:59:18 hello 16:59:27 hey team 17:00:19 So last week was rough -- lots of sick folks 17:00:23 how are we doing this week? 17:01:04 \o/ happy that there are still sunny days where i am 17:01:57 still somewhat sick but getting better 17:02:08 gaba: where are we wrt our roadmap? I think we're supposed to be starting on 0.3.6, but we still have a lot of 035 stuff 17:02:36 this week is supposed to be lots of 0.3.5 fixing too, right? 17:02:39 Yes. We are starting today. I reorganized the tickets so the ones at the top are priority. 17:02:43 great 17:03:04 So the idea would be to get the tickets at teh top first. I added lines to roughly consider estimation on tickets per month 17:03:26 We need to combine bug fixing for 0.3.5 with issues in the 0.3.6 roadmap 17:03:41 sponsor 8 is priority 17:03:52 looking at the roadmap... does everybody have something to work on in the first section? 17:03:52 please reach out if you get blocked in any of this issues. 17:04:17 Ideally we'll be dinishing this stuff this month? does that all sound plausible? 17:04:31 hello. im in a plane about to take off. will be around for 15 mins or so and then disappear. 17:04:32 what are we looking at right now? which sheet? 17:04:33 *finishing 17:04:38 the roadmap sheet 17:04:47 * asn reads backlog 17:04:57 asn: it seems you are not assigned to stuff for the first part. The stuff about wtf-pad with mike 17:05:05 https://docs.google.com/spreadsheets/d/1Ufrun1khEo5Cwd6OwngERn829wU3W3eskdrriaYfUBQ/edit#gid=856122210 17:05:30 ahf: you'll have to brief me in for #25502... I have no idea what that is :P (row 70) 17:05:30 i think right now the "merge wtf-pad patch" task is mostly me 17:05:33 Looks like I'm on "act on disable-network feedback". The problem is that we don't have any feedback on that. I think we need to move that item 17:05:35 or at least i need to do the big review 17:05:52 that is "Merge padding code (BEFORE other stuff)" 17:05:55 yes dgoulet 17:06:03 dgoulet: we need to split it into different tickets 17:06:10 multiple things are right now #25502 and it shouldn't be 17:06:21 ack 17:06:24 nickm: what does mean for that issue, is it done or we need to wait for anything else? 17:06:44 #28019 17:06:49 gaba: We need to wait to actually get feedback, and handle it if we get some 17:07:13 ok. I will move it into november 17:07:45 what does the "!" means again? 17:07:47 catalyst, ahf, teor: It looks like the bootstrapping-related stuff is scheduled for this month. Plausible? There are about 10 working days left. 17:07:49 gaba: the plan right now is that im tasked with reviewing the giga patch at "Merge padding code (BEFORE other stuff)", and then as mike does the other tasks i need to review them. 17:07:56 ! meant "I think this person needs to be on this task". 17:08:00 ack 17:08:03 I could be wrong of course 17:08:15 gaba: so i have the responsibility of "Merge padding code (BEFORE other stuff)" atm. feel free to move it to "first part" if you want. 17:08:27 nickm: i don't have much feeling with that one, except for the reporting in the PT case 17:08:31 nickm: it's the 80% issue? 17:08:43 the assigment and names (and points estimation) is always open for comments/questions/changes. 17:08:48 There's the PT reporting, and there's the improved bootstrap reporting 17:08:56 those are both interrelated 17:08:57 nickm: i'm not sure i see where the end-of-month target date is coming from 17:09:00 asn: also the proposal update and torrc serialization, which I guess can come after 17:09:04 yes 17:09:12 torrc serialization i can also be main. but that's after. 17:09:16 asn: anyway I am still fixing up stuff from your first review in mexico. hope to have that in good shape soon 17:09:23 Some of this issues start in october but will finish in november. I considered that. 17:09:33 mikeperry: i was hoping you could do the proposal update you currently have in mind, and then i do review and another pass on it. 17:10:04 mikeperry: should we switch the serialization to be me "first part"? 17:10:08 catalyst: I think gaba is talking about that target date here 17:10:23 nickm: probably not in 10 days, especially not if we're prioritizing 0.3.5 fixes 17:10:33 ok, i think now that we know we can do the windows part with timers i think the better PT reporting part can be done mostly at the end of this month, minus having the protocol ready to what PT's can signal where they have some knowledge they would like to relay 17:10:42 so "log" messages at the end of this month should be doable 17:10:42 gaba: so, you mean that if something is listed in October i n the roadmap, we should be starting now and not necessarily getting it done till Nov? 17:10:50 yes 17:10:57 catalyst: ^ that 17:10:59 asn: possibly? 17:11:00 nickm: ok i'm not sure if i had heard that target date before and it wasn't obvious from the spreadsheet 17:11:04 mikeperry: ok i change it. 17:11:11 mikeperry: done. 17:11:15 there are some issues that the estimation is 20 days, there is no way that we finish taht in october 17:11:34 ahf, catalyst: let's try to at least break down the steps on these tickets this week so we can plan better. They're listed as 5-20 points because IMO I think we have lingering uncertainty on the sub-tasks 17:11:49 agreed 17:11:58 nickm: ok 17:12:02 If we have a plan, we can be less scared. 17:12:45 gaba: Remember, these points estimates are pretty much "wild guess" territory; I hope we can make them better as we move 17:12:45 17:10 <+mikeperry> asn: anyway I am still fixing up stuff from your first review in mexico. hope to have that in good shape soon 17:13:02 mikeperry: yeah i was knees deep in proposals and backlog all this time. still havent crawled out yet, but i should be in good shape late this week. 17:13:05 hm where does the sums come from for capacity estimation right now? 17:13:07 mikeperry: so i think we can time well. 17:13:22 i think my google docs fu stops when i see the labels that is being used in that expression 17:13:25 ahf: Mostly, I made them up. 17:13:27 mikeperry: that is, when your revisions are done, i should be ready to start the second review. 17:13:31 nickm: ah, ok! 17:13:32 asn: ok yeah, no worries. I am also trying to gather stuff from researchers 17:13:41 mikeperry: yeah i have a thread from you that i still havent read. 17:13:41 i thought they were pulled from trac or something 17:13:44 yes nickm on points. Right now is what we have but I really hope we can improve it. 17:15:24 next topic is reviews. Wow, that's a lot of reviews! 17:15:39 Looks like we need to do them though... 17:15:43 ahf: if you get a task and you think the estimation could be better, please do it. We are assuming the max of the estimated time. 17:15:46 yep i left a disclaimer on the top 17:16:00 everyone has >= 4 reviews atm 17:16:10 Could everybody please try to get to 0.3.5 reviews early in the week this week? 17:16:17 ack 17:16:30 ahh, i see the estimation column now 17:16:33 That would be really helpful, and would let us get an alpha out and move on 17:16:35 my screen is too small, sorry 17:16:50 also i was not on the previous monday meeting, but dgoulet told me that it was semi-decided that it's the reviewer's task to take care of missing PRs 17:16:58 which makes sense i think 17:17:07 gaba: yeah, i'm gonna flesh the #25502 ticket out into a lot of children and then we can do estimates on those 17:17:08 I don't think we decided that? I think we talked about it though... 17:17:12 so if you are a reviewer, and a PR is missing, feel free to mark it as needs-revision, or make the PR, up to your disgression. 17:17:23 nickm: ack 17:17:27 nickm: Ah, wait, no, I think we did. 17:17:30 sorry, I misunderstood 17:18:01 Also IMO I think reviewers should make the PR if it is an externally submitted ticket 17:18:11 right 17:18:19 i think so too 17:18:22 let's try to do it this way, and see how it works :) 17:19:21 For this week's rotations, we have teor on bug treach, nickm on ci + coverity 17:19:47 I think we have decided that the other 2 things aren't really happening, right? 17:19:48 bug treachery 17:19:54 triage , arg 17:19:58 my fingers 17:20:05 :-) 17:20:07 yes, the other 2 are things everybody should take some time of the week for it. 17:20:48 Proposal that i will move to the wiki by next week: https://pad.riseup.net/p/team_rotation 17:21:41 looks reasonable, maybe we should clean up a bit 17:21:45 * ahf looks 17:22:06 yeah stable roles, 1/2 is *not* decided yet... 17:22:14 so lets not put it on the Wiki just yet ? ;) 17:22:31 mark that on the pad so we know? 17:22:36 ok, i would like everybody to comment on it and make a decision by the end of the month 17:22:52 ack 17:23:02 ok 17:24:18 If there is anything else that you will be working on that is not on the roadmap (or weekly tasks) please let me know. thanks! 17:24:46 ack 17:25:02 shall we move on to discussions? 17:25:30 the first one is that I need everybody to also let me know if there's something assigned to you in 0.3.5 that you won't be able to finish up Really Soon Now 17:26:09 ok 17:26:21 oh, RSN 17:26:23 hehe 17:26:31 It's okay, and I expect that we'll defer a bunch of things like we usually do, but we should know about it and plan for it... 17:26:35 ok this plane is taking off. will check backlog when im back home tonight. peace. 17:26:46 ... not just get to Dec 15 and notice "oh hey that one was important" 17:27:22 aye 17:27:51 If there's anything that's super-big or hard, we should postpone it from 0.3.5 if we can.... 17:28:05 speaking of if-we-can... 17:28:29 https://trac.torproject.org/projects/tor/query?priority=Immediate&priority=Very+High&priority=High&status=new&status=reopened&milestone=Tor:+0.3.5.x-final&group=status&order=priority 17:28:44 that's the new and reopened high/very high tickets for 0.3.5 17:29:22 I think all of these will need a couple of people working on them to figure them out... 17:29:41 the new Very High tickets seem kind of heavy 17:30:05 as in, a lot of work? 17:30:10 yeah 17:30:15 that memleak is something I can't figure out since last week ... so I've kind of gave up for now on waiting more feedback from people getting it 17:30:27 catalyst: I agree. 17:30:31 yes. can people pair (maybe with voice) on the heavy ones? 17:30:35 dgoulet: Is anybody besides you working on that? 17:30:47 dgoulet: is there specific information you're waiting gor? 17:30:49 *for 17:30:50 from within the team, no 17:31:23 I just can't reprod that ... no where, whatsoever! ... it almost appears like a kernel leak ... difficult to say 17:32:01 Sounds like it might be necessary to try it on a VM with the same OS. 17:32:04 hmmm is 27915 really that Very High for 035? ... we aren't shipping rust at all so why 035? 17:32:10 nickm: I did that ;) 17:32:15 nickm: more than once 17:32:22 dgoulet: ouch 17:33:18 #27915 is currently High and maybe it doesn't need to be on 035 because it's waiting on Alex Crichton for info? 17:33:22 dgoulet: so what info would help? Knowing what's different? Knowing where the memory went? etc... 17:33:59 catalyst: I think we can take it in 036, sure. I had it High because it seemed CI/Testing related. 17:34:02 Let's defer it.... 17:34:22 nickm: I honestly do not know... Hello71 said that once they hit it again, they will poke more stuff... but apart from that, I have _no_ clue now what or how to look for here without spending serious time in possible dead end :S 17:34:39 nickm: is it something we're not testing, or something that breaks the build? i thought it was the first 17:34:48 It's the first, yeah. 17:36:04 dgoulet: okay, let's leave it as "new" then, since we have no idea what to do.... 17:36:32 ack 17:36:41 I'm monitoring it closely so at least that is that 17:37:05 so as for the other ones... 17:37:23 I think rl1987 is doing some good with #27808 17:37:31 that leaves #27800 , #27331, and #27750 17:38:01 I know that code pretty well, but I think if I try to track them all down myself, it will eat me 17:38:10 #27331 looks interesting. we have only gotten one report on that? :o 17:39:01 ahf: yeah. We've worked on that code together before. Want to try to look at this one? 17:39:08 (together, I mean) 17:39:19 i think we should, yeah 17:39:27 ok 17:40:03 I can try #27750 ... I know that code but without reprod. it is best effort eheh :) 17:40:12 i'd like to fix my two rust 0.3.5 tickets this week as well, but i have no idea how we should attack this. being able to reproduce it would be good 17:40:16 anybody want to help with the other two? The nodelist stuff is a #27800 data-structure invariant thing, and #27750 is a state machine thing.. 17:40:31 but my money is on that it is linked to the OOM or/and mem. pressure we are seeing :S 17:40:44 I think we might wind up just adding better logs for these cases and hoping we can fix them when we know more 17:40:56 dgoulet: Okay, let's try #27750 17:41:32 * catalyst will take #27800 17:41:46 catalyst: ok, let's talk after the meeting. 17:42:10 we're running towards the end of our time and of the discussion topics too.... 17:42:56 re PTO, please don't link the non-public calendar to a public doc 17:43:06 catalyst: you haven't gotten anti-censorship CV's from Erin, right? i think she might have only sent them to gaba and me 17:43:19 the link in the pad is to the general meeting calendar 17:43:27 ahf: correct 17:43:42 catalyst: fixed... 17:43:52 catalyst: ok, gonna write to her 17:43:59 How do we feel about moving the first meeting of each month to the patch-party time for now? 17:44:05 nickm: thanks. i think that meeting calendar is public but not sure 17:44:06 that's Tuesday 2300 17:44:13 ahf: already wrote to her, no response yet 17:44:22 catalyst: ah! ok, then i wont write, cool 17:44:34 catalyst: ok, I put the link back then, sorry. Please remove any links that shouldn't be there? 17:44:41 nickm: that is ok with me 17:44:53 i think it would be good for a meeting where teor can be present too 17:44:59 and tuesday at 23 sounds good 17:45:06 i might miss it sometimes, but i'll aim for being there 17:45:13 the proposal is to have the meeting with teor once a month 17:45:14 I'd like all the people who aren't in Europe to try to make that time. If you're on europe time, that's probably the middle of the night 17:46:19 so Nov 5th will be the first one at this time ^ 17:46:35 yup. Let's circulate reminders closer to then 17:46:42 yeah 17:46:45 and also we should tell tor-dev and adjust the schedule on the wiki 17:46:46 so it's the regular patch party time, except it's an official meeting once a month? 17:46:54 yup 17:47:05 wait that is not what I understand ^ 17:47:13 our weekly meeting is moved for the first of the month 17:47:29 meaning on Nov 4th, we do not have a weekly meeting? or we do and we duplicate the next day? 17:47:42 I think we move the Nov 4 meeting to nov 5. 17:47:50 (Is what I had understood) 17:48:05 good ok 17:48:20 so not a patch party ;) but our meeting 17:48:20 yeah 17:48:21 * catalyst sees a Monday Nov 5 and a Tuesda Nov 10 17:48:42 oops. That should be the 6th, then. 17:48:45 yes, I will change it in the calendar. 17:49:14 anything else for this week? A bunch of folks have used the "ask-for-help" area on their updates .... 17:49:34 .... so please check to see if anybody has mentioned you there, and talk to them :) 17:49:45 great :) 17:50:04 nickm, dgoulet, gaba: should we talk PT status reporting subtasks tomorrow? 17:50:17 i need to finish the s19 stuff for asn tonight, so him and i can talk about it tomorrow 17:50:25 ahf: that's fine with me... 17:50:29 yeah same here... need to finalize s19 17:50:31 ahf: so yes 17:50:42 great, cool 17:50:43 yes 17:50:45 sounds good 17:50:59 ok. I'll mark the 035 tickets we talked about, get to any of my 0.3.5 reviews 17:51:10 and talk to folks about the 035big-ticket stuff 17:51:12 nickm: you assigned us to the consdiff one? 17:51:25 ahf: I mentioned us on it; there's no way to dual-assign afaik.... 17:51:51 cool 17:52:06 any more for today? 17:53:08 okay, folks. Let's get 0.3.5 out and 0.3.6 staretd. If you get stuck on anything, please talk to me and gaba right away and we'll make it all work :) 17:53:11 peace, all! 17:53:15 #endif 17:53:17 bye! 17:53:19 um. that's not it 17:53:19 o/ 17:53:20 #endmeeting