16:59:54 <ahf> #startmeeting network team meeting, 17th of may 2021 16:59:54 <MeetBot> Meeting started Mon May 17 16:59:54 2021 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:59:54 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:01 <ahf> hello hello 17:00:04 <GeKo> o/ 17:00:06 <ahf> https://pad.riseup.net/p/tor-netteam-2021.1-keep is our pad 17:00:16 <dgoulet> o/ 17:00:16 * nickm running late, sorry! Adding to pad now 17:00:26 <ahf> no rush nickm, you can also do it after 17:00:36 <nickm> not for proposed discussions I can't :) 17:01:04 <ahf> that is true 17:01:35 <ahf> dropping link to our board: https://gitlab.torproject.org/groups/tpo/core/-/boards - are folks doing alright there? 17:01:54 <nickm> I think I am... 17:02:39 <dgoulet> yes 17:03:18 <ahf> excellent 17:03:29 <asn> o/ 17:03:59 <ahf> how do we think the 0.4.5-post-stable and 0.4.6-stable situation is? 17:04:00 <ahf> hello asn 17:04:30 <nickm> hmmm 17:05:01 * ahf needs to revise the patch for tor#40333 17:05:14 <nickm> so I think we're okay; I don't currently see anything on those lists where I would shout "release blocker!" 17:05:21 <ahf> ya 17:05:38 <nickm> ahf: tor!338 is ready for your review again :) 17:05:53 <ahf> perfect, will get to it today 17:06:05 <ahf> ok, looking at discussion items now 17:06:13 <ahf> i wonder why it is mixed up in two places right now in the pad 17:06:42 <dgoulet> yeah I have one in 045 but I have no clue how to address it and it is not critical at all so we are fine on my side 17:06:50 <nickm> dgoulet: which one? 17:06:50 <ahf> perfect 17:07:10 <dgoulet> https://gitlab.torproject.org/tpo/core/tor/-/issues/40302 17:07:24 <dgoulet> operators have weird setup I can't reproduce... 17:07:36 <nickm> Do they have repro instructions? 17:07:36 <dgoulet> I had an idea in mind lately that I need to try on 17:07:40 <dgoulet> sort of yeah 17:07:50 <nickm> It might be worthwhile to ask for easier repro instructions if you can't follow theirs :/ 17:07:51 <dgoulet> just that it is a mixed of public IP + firewall + private IP ... 17:08:01 <dgoulet> problem is setup 17:08:09 <dgoulet> but I have an idea so I will try 17:08:17 <nickm> ok 17:08:19 <dgoulet> but worst case, they found a workaround 17:08:37 <ahf> first discussion item today: 17:08:40 <ahf> 2021-05-17 [nickm] Let's divide up the new TROVE stuff, or make a plan to talk about it? And let's plan release dates? 17:09:14 <asn> sounds good about dividing! 17:09:22 <nickm> Yeah, did everybody see those new tickets? 17:09:29 <dgoulet> yes 17:09:51 <asn> i can take 006 17:10:15 <nickm> asn: much appreciated! I'll assign to you. 17:10:25 * asn took it 17:10:27 <nickm> I've got a patch for 004; I could do either or both of the others 17:11:00 <ahf> i have not looked at those tickets yet 17:11:17 <dgoulet> btw, 004 is not private 17:11:22 <nickm> yes, it is not 17:11:24 <dgoulet> ah first comment :) 17:11:55 <dgoulet> I can try 003 17:11:56 <nickm> I'll take 003 since it's a bit tricky and I know the code. 17:11:59 <dgoulet> roger 17:12:06 <nickm> oh! Or i can review dgoulet's fix 17:12:13 <nickm> dgoulet: which would you prefer? :) 17:12:16 <dgoulet> I would feel more comfortable on 003 than 005 lol 17:12:29 <nickm> huh, ok. 17:12:33 <nickm> then take 003 and make me reviewer? 17:12:37 <nickm> let me know if you get stuck at all 17:12:37 <dgoulet> ack 17:12:39 <nickm> I'll do 005 17:12:43 <nickm> unless ahf wants it 17:13:09 <nickm> assigning to myself. 17:13:10 <ahf> if you want it done deterministically fast then please take it - i know these things usually can get a bit stressful 17:14:13 <nickm> ok! 17:14:53 <ahf> ok, next item? 17:14:55 <ahf> 2021-05-17 [nickm] Please remember to look for ways to do _less_ in C tor, especially if it comes with a maintenance risks. 17:15:05 <ahf> this is a bit like an announcement/reminder? :-) 17:15:20 <nickm> we missed one thing: 17:15:20 <ahf> or maybe it is also a topic for discussion if anybody have any questions/discussions around that area? 17:15:23 <nickm> on the last topic 17:15:27 <ahf> oh 17:15:27 <ahf> we did 17:15:27 <nickm> "and let's plan release dates" 17:15:33 <ahf> yes, sorry, reverting back to the TROVE item 17:16:10 <nickm> We should be aiming for a final 046 RC around the end of May, I think. Maybe mid-June would be the right time to put out 046 as well as all these fixes? 17:16:39 <dgoulet> yeah sounds reasonable 17:16:44 <nickm> ok 17:17:14 <ahf> mid-June it is 17:17:41 <nickm> Also if you haven't done a private branch and private MR before: please test with a non-sekrit commit before you make a private MR for one of these fixes :) 17:18:15 <dgoulet> interseting, I've never done a private MR 17:18:16 <nickm> IIUC you have to make a private fork of the tor repository, and make your MR relative to that project's own master branch 17:18:22 <dgoulet> ack 17:18:34 <nickm> (If there's a better way I will eagerly learn) 17:19:08 <ahf> yeah, need an additional fork of tor.git than the one you usually use 17:19:27 <ahf> so you can fork to another name on in your user dir on gitlab 17:20:35 <ahf> ok 17:20:39 <dgoulet> hmmm wait 17:20:44 <dgoulet> I'm unable to fork tor.git again... 17:20:51 <dgoulet> it keeps bringing me to the current one I have :S 17:20:54 <dgoulet> fork again* 17:21:05 <ahf> ah! 17:21:10 <ahf> then create a new project in your user dir 17:21:16 <ahf> called tor-private or whatever, mark it secret 17:21:22 <ahf> then push locally to the empty repo 17:21:24 <dgoulet> ok! 17:21:35 <dgoulet> but I can make MRs that way against core/tor ? 17:21:43 <dgoulet> if it was not "fork" from core/tor? 17:21:57 <ahf> hm. i hope gitlab isn't so silly that it requires that the social graph of the repos are intact 17:22:02 <ahf> nickm: how did you create yours? 17:22:12 <nickm> I don't remember. 17:22:14 <ahf> ok 17:22:21 <dgoulet> I,ll try 17:22:29 <ahf> sounds good 17:22:47 <ahf> ok, going to next item 17:22:56 <ahf> 2021-05-17 [nickm] Please remember to look for ways to do _less_ in C tor, especially if it comes with a maintenance risks. 17:23:16 <ahf> anything we want to discuss about this here and now? it seems quite open 17:24:34 <ahf> i take that as a no 17:24:35 <nickm> yeah, I guess I'd just like folks to think about it and try to invent ways we can make this more of a policy in the future 17:24:40 <nickm> that's all 17:24:40 <ahf> ya 17:24:44 <ahf> makes sense 17:25:02 <ahf> ok, let's talk hackathon on Thursday 17:25:04 <dgoulet> I'm personally keeping it very narrow to sponsors and reported bugs tbh 17:25:11 <ahf> dgoulet: nice 17:25:12 <dgoulet> and always keeping in mind to improve arti first 17:25:39 <dgoulet> go ahf 17:26:10 <ahf> we need to talk about if we need to prepare anything first. we talked about last week a bit about talking about some ideas we have for the hackathon too 17:26:31 <ahf> like specific project ideas that allows us to play a bit around with the arti api and what is possible to do with the current codebase and its api 17:26:52 <nickm> I wonder if anybody outside this team has any ideas for quick hacks they'd like somebody to build 17:27:14 <ahf> GeKo, dgoulet, and arma2 might have some ideas from the health team? 17:27:24 <dgoulet> yeah but I'm in this team :P 17:27:24 <GeKo> dgoulet has :) 17:27:28 <ahf> they always seem to be working with a ton of smaller tools 17:27:36 <ahf> dgoulet: XD you are welcome to say them too 17:28:05 <dgoulet> well they are not quick hacks 17:28:27 <ahf> are they 3-4 hour hacks? 17:28:42 <dgoulet> no idea tbh... maybe something nickm would be able to estimate 17:28:43 <nickm> just say and maybe we can build a tiny piece :) 17:28:45 <GeKo> i feel i don't know enough about arti at that point to suggest any quick hacks 17:29:05 <nickm> like, ,maybe quick little tools you wish you could build with stem or tor or something 17:29:20 <ahf> i got a little bit hooked on trying to write something that could scan for inter-relay reachability 17:29:28 <dgoulet> but mostly what I think we need is a way for arti to be designed in a way that as an API user, I can have fine grained control on let say a circuit on how it is built and what goes on it for instance... 17:30:12 <dgoulet> and so we can build tools on the health side that allows us to tests relays like sending them Cell A, B, C in that order and looking on how it responds and timings 17:30:18 <nickm> ahf: that part should be pretty easy 17:30:23 <dgoulet> that is why I also worked a bit on the tracing + event part laast week 17:30:26 <ahf> yeah, and it wont discover anything new 17:30:31 <dgoulet> so we could get our tools to learn what is happening 17:30:49 <dgoulet> in realtime as they are unfolding 17:30:50 <ahf> dgoulet: is there a list of the health tools somewhere and what they do? i assume most of them are implemented with stem? 17:30:57 <dgoulet> so this is the broader design idea I have in mind 17:31:10 <dgoulet> ahf: yeah they are on Gitlab 17:31:16 <dgoulet> but I just don't want to just rewrite them tbh 17:31:19 <ahf> and we can see them? 17:31:20 <ahf> ok 17:31:23 <dgoulet> I want more control with arti :) 17:31:27 <dgoulet> that I don't have now 17:31:49 <dgoulet> *if* we want tools for the hackaton, then yes, the health repository has plenty! "margot" is a great start for a lot of them :) 17:32:45 <ahf> do you have a wishlist for margot? 17:32:52 <dgoulet> I do 17:32:54 <nickm> what I want to avoid with the present discussion is that we all show up on thursday with no idea what to hack on :) 17:33:05 <ahf> yeah 17:33:17 <dgoulet> I would like an async way to describe what I want here... email thread likely 17:33:23 <dgoulet> it is not easy to outline it just here in a chat... 17:33:30 <nickm> sounds fine to me 17:33:32 <dgoulet> I want to provide examples and such so it is a longer discussion 17:33:47 <ahf> sounds good to me. if there are useful things that would be useful here and now on margot, then i am up for diving into that 17:33:59 <ahf> dgoulet: want to start a thread on ... on hm! 17:34:23 <ahf> i have no idea what list would be ideal for that. maybe the damned network team list 17:34:39 <dgoulet> hackathon is not open to public right this time? 17:34:51 <ahf> it is not, maybe just a list of the people who have singed up for the hackathon? 17:34:53 <dgoulet> ok so yeah lets just talk between ourselves then... 17:35:01 <ahf> like just a big CC list instead of an ML 17:35:11 <dgoulet> sure 17:35:52 <ahf> awesome, thank you 17:36:04 <ahf> in this mail you will also write about some of the wishlist items you have for margot? 17:36:19 <dgoulet> yes ofc, I'll outline it all 17:36:25 <dgoulet> ALL my ideas, my entire brain 17:36:57 <ahf> ok, awesome! that sounds like a long email 17:37:04 <ahf> thi means i also got an answer to my discussion item 17:37:25 <ahf> i don't see anything else on our list, so i think it's time for the s61 part of the meeting with mikeperry 17:37:36 <mikeperry> kk 17:38:03 <mikeperry> so I think the main thing going on is rob's speedtest experiment 17:38:20 <mikeperry> GeKo: you mentioned something about dennis looking into a parallel experiment in email? 17:38:29 <mikeperry> is that just to gather more datapoints, or other? 17:39:25 <GeKo> other data points 17:39:43 <GeKo> it's about how users are affected by the flooding 17:39:48 <GeKo> and whether they are at all 17:39:49 <mikeperry> ok. so it is not sth that will influence any results substantially 17:39:50 <mikeperry> kk 17:39:55 <GeKo> no 17:40:07 <GeKo> it's more to give a fuller picture 17:40:12 <mikeperry> awesome 17:40:34 <GeKo> yep 17:41:33 <ahf> so a copy of rob's experiment or? 17:41:39 <mikeperry> ahf: lmk if there are any questions or anything wrt https://gitlab.torproject.org/tpo/core/tor/-/issues/40312. I can try to keep an eye on the ticket 17:41:41 <ahf> or is it another speedtest experiment? 17:41:51 <ahf> mikeperry: sounds good 17:41:56 <mikeperry> ahf: yah, pretty much re-running the previous one 17:42:00 <ahf> excellent 17:42:59 <mikeperry> there's some minor improvements I think.. last time not all relays were tested, etc 17:43:26 <ahf> nice 17:43:38 <mikeperry> our main activity is to look for any overload or issues, and see if we can do this more often. it did help.. but it also hurt stuff and caused bugs. so we're gonna find 'em 17:44:10 <mikeperry> that DNS one is a major suspect. perhaps there will be others. we'll be doing this on and off for a while 17:44:47 <mikeperry> anyway I think that's it for s61 for now unless there are other questions 17:44:56 * ahf has none 17:45:34 <GeKo> <- is fine, too 17:46:02 <ahf> let's call it then! 17:46:04 <ahf> #endmeeting