17:00:37 <nickm> #startmeeting weekly network team meeting
17:00:37 <MeetBot> Meeting started Mon Jun  5 17:00:37 2017 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:37 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:00:40 <nickm> hi all!
17:00:50 <nickm> the pad's over at http://5jp7xtmox6jyoqd5.onion/p/1ngmNzW2fKcl
17:00:58 <nickm> (thanks to eor)
17:00:59 <nickm> *teor
17:01:16 <ahf> hello hello
17:01:17 * dgoulet writing my update, should be on the pad shortly
17:01:23 <nickm> alternatively, you can use pad.riseup.ne
17:01:27 <catalyst> hi
17:01:36 <haxxpop> hello !
17:02:10 <isabela> hi
17:02:32 <komlo> hello
17:02:58 <nickm> so, time for us all to write updates, and read over each other's updates!
17:03:11 <nickm> and to ask questions!
17:03:53 <marrowgari> hi
17:03:59 <dgoulet> updates is in!
17:05:23 <nickm> ok.  let's read everybody's updates and ask one another questions!
17:05:48 <nickm> I will try to sort the list of discussion topics by how long they should take.
17:07:03 <ahf> we can do our sponsor4 task in #tor-project as an extension to this meeting if that makes it easier?
17:07:22 <ahf> if it's hard to get everything done in ~60 min's
17:07:34 <isabela> ahf: sounds good
17:07:40 <isabela> i have a meeting after this meeting tho
17:07:49 <isabela> :) but i could do it at any time after that too
17:07:59 <ahf> oh, ok, i'll let nickm decide this then :-)
17:08:02 <ahf> i'm on for the entire evening
17:08:32 <nickm> asn, dgoulet: Could I ask you two to try to get the "v3 hsdesc fuzzing" task done some time this month?  If it finds bugs, it would be good to have them fixed before we call 0.3.1 stable
17:08:38 <nickm> asn, dgoulet : does that sound reasonable?
17:09:34 <Yawning> hm
17:09:37 <nickm> asn, mikeperry : I have a request for prep on guard meeting: Let's keep the core of the prep focused.  IOW, it's okay to have 12 documents on the reading list, but we should distinguish between the 12 on the reading list and 1-2 documents on the must-read list
17:09:48 <dgoulet> nickm: I would say difficult on the timeline of getting prop224 minimal viable product on time *but* I see this task as essential so we have to make it happen anyway
17:10:13 <nickm> dgoulet: I would consider "doesn't crash remotely" part of MVP
17:10:17 <dgoulet> haxxpop: btw _great_ task if you are looking for one, fuzzing hs v3 code merged :)
17:10:23 <dgoulet> nickm: indeed
17:11:41 <nickm> ahf: wrt the sponsor4 stuff -- how do the measurements look so far?
17:12:11 <haxxpop> dgoulet, you mean it's already done and I will find bugs ?
17:12:24 <dgoulet> haxxpop: no meaning we need to do it before 031 stable :)
17:12:29 <dgoulet> haxxpop: and you might find bugs yes :)
17:12:37 <dgoulet> haxxpop: anyway, if you feel like it, no pressure :)
17:12:39 <ahf> nickm: it's running and i have some logs, it's been running for a week tomorrow so i'll start doing comparison there
17:12:47 <haxxpop> dgoulet, so you mean nothing is done ?
17:13:08 <dgoulet> haxxpop: nope, tor has a fuzzing framework but nothing done on v3 hs
17:13:23 <nickm> ahf: will this get us useful data, or will the fact that most relays are still <= 0.3.0 mean that it tells us "no change" ?
17:13:53 <haxxpop> dgoulet, ok I haven't taken a look at that framework. I will do it tomorrow
17:13:57 <nickm> what are the old numbers?
17:14:07 <ahf> nickm: i'm running it with chutney on one set of nodes that is only 0.3.1 and one other set of chutney nodes running 0.3.0
17:14:11 <ahf> so it'll be the ideal case
17:14:33 <nickm> haxxpop: see doc/HACKING/Fuzzing.md for an introduction
17:14:46 <nickm> ahf: ok. so, we'll need to extrapolate to guess real-world data.
17:14:47 <nickm> that's fine
17:14:53 <ahf> yes, that was my idea
17:15:05 <ahf> we have the case before this and the case where the entire network is being nice and is upgraded :-)
17:15:48 <haxxpop> nickm, ack
17:16:02 <mikeperry> my must-read links are the agenda pad, prop247, and rpw's Trawling for Tor Hidden Services paper I think. but that is 3 :)
17:16:14 <mikeperry> nickm: ^ for the guard discovery meeting
17:16:56 <dgoulet> mikeperry: maybe send it on the meeting email thread?
17:17:09 <ahf> yes please
17:17:11 <mikeperry> yeah, I will.
17:17:15 <dgoulet> neat
17:17:42 <mikeperry> I am going to re-read the trawling paper today. I think a lot of it is covered by Prop#224. Only one or two sections are specifically about guard discovery
17:17:58 <mikeperry> so I will find those and highlight them for folks
17:18:07 <dgoulet> prop#247 you mean?
17:18:57 <mikeperry> no, 224. the first half of the trawling paper is about hsdir positioning, which is addressed by 224. only the later sections are relevant to prop#247
17:19:12 <dgoulet> ah in that sense ok
17:21:00 <nickm> mikeperry: What I mean is: there are 3 things on your list, there are 5 things on asn's, and etc
17:21:15 <nickm> though there is overlap
17:21:24 <nickm> and we have a whole week to come up with more things to say everybody should do first
17:21:58 <nickm> more questions on updates? More updates?
17:22:02 <nickm> If not, let's discuss things
17:22:05 <mikeperry> maybe we try to converge on the canonical set of reading material on the hackfest pad, then
17:22:25 <nickm> mikeperry: sounds good
17:22:34 <nickm> asn: ok with you? if so please coordinate
17:22:39 <nickm> (same goes for everybody else with a list)
17:23:08 <nickm> discussion topic 1: I think that teor wanted us to make sure we were doing something about TROVE-2017-004 and -005.  I believe we are; thanks to everybody who's looked at those.
17:23:37 <nickm> I think that releases thursday or friday are reasonable.
17:23:40 <nickm> anybody disagree?
17:23:41 <catalyst> do we have regression tests? if not, how badly do we want them?
17:23:57 <nickm> no; but it would be nice?
17:24:16 <nickm> (remember, don't tell people how to trigger the bugs in public, until the fixes are out)
17:25:03 <nickm> I think we should do another bug retrospective in montreal, much smaller than the previous one
17:25:07 <nickm> it should be informative
17:25:44 <nickm> next is a note: google docs doesn't work for teor on TB, so let's avoid it when possible, or provide an alternative.
17:25:53 <nickm> next is: should isabela go to wilmington?
17:26:06 <nickm> For the people maintaining the schedule: how does that look?
17:26:35 <nickm> (Also, do we have somebody filling an 'organizer' role to try to make sure we stick to the schedule, remind us to take breaks, etc?)
17:26:54 <dgoulet> I assume that could be the session's moderator
17:27:22 <catalyst> seems like there aren't really times-of-day specified yet?
17:27:24 <nickm> do we have anybody responsible for herding moderators?  I usually need a little hearding
17:27:28 <nickm> *herding
17:27:33 <Yawning> use a laserpoiner
17:27:37 <Yawning> *pointer
17:27:40 <Yawning> it works for cats
17:27:57 <isabela> nickm: re:google docs - i am not sure what to do about that. it wont work for anyone on high security level on tor browser for sure
17:28:06 <dgoulet> nickm: not that I know of but we can definitely apoint someone on the first day
17:28:25 <isabela> nickm: but I also dont know what to use instead as i need somethign i have to add images to it, and can take ppl comments etc
17:29:02 <nickm> isabela: I think we'll have to just remember "it doesn't work for everyone", so when using it, make sure to provide some other means to read (eg, offer to send a pdf or openoffice document or something)
17:29:13 <isabela> well
17:29:15 <nickm> dgoulet: that would be good.  you?
17:29:18 <isabela> i would just change if possible
17:29:19 <nickm> :)
17:29:34 <dgoulet> nickm: ok it will be me :)
17:29:35 <catalyst> are there obvious alternatives to google docs?
17:29:45 <isabela> but i dont know of another solution that does what i need and work on high level security
17:30:06 <isabela> i could download what is there as pdf to share or something but still, its a live doc right now
17:30:16 <dgoulet> nickm: there it's on the pad therefore official :)
17:30:23 <nickm> dgoulet: woo!
17:30:54 <nickm> isabela: yeah, I'm not saying I see a great solution right now either
17:31:13 <nickm> but let's keep it in our minds as a thing-we'd-like
17:31:21 <isabela> sounds good
17:31:47 <nickm> so, isabela-in-wilmington: yes or no?  I say "Yes if the schedule is not already too  tight for her to do what she wants"
17:32:35 <catalyst> ok with me
17:32:44 <dgoulet> scheduled seems to still have some holes in it but let'S keep in mind that maybe a "half-day/day off could be wise"
17:32:46 <dgoulet> -d
17:32:50 <nickm> true.
17:33:13 <ahf> i think it would be good. i think she's the only one from the team i haven't met yet, so if it fits in somehow i'd say yes
17:33:18 <nickm> dgoulet: also, this week, we really need to figure out working hours for wilmington before we start
17:33:38 <dgoulet> nickm: indeed
17:33:42 <nickm> otherwise somebody will want to start at 8 (me), and somebody will want to keep going till 7pm (not me)
17:33:55 <komlo> catalyst: maybe we can talk about regression testing also in the testing session on thursday
17:33:55 <dgoulet> nickm: it makes two of us, great :P
17:33:56 <nickm> dgoulet: is that something that you can be in charge of too?
17:34:08 <dgoulet> nickm: yup, I'll herd the schedule maker! :D
17:34:16 <nickm> great, ty
17:34:27 <catalyst> komlo: sounds good; i'll start sketching an outline in the pad
17:34:55 <dgoulet> nickm: one thing that would be I would say *essential*, if for you, me and asn to tak on Thursday about prop224 032 merge so if possible, I'm booking you ;) shouldn't take very long
17:35:24 <nickm> on to the next question?
17:35:30 <dgoulet> nickm: (won't be like in AMS, this time no open issues, just logistics)
17:35:30 <nickm> or more on wilmington?
17:35:43 <dgoulet> about Wilmington and isabela
17:36:04 <dgoulet> as I said, if isabela is there, good for *me* because I can then have a DRL meeting pitch ;)
17:36:23 <nickm> I think isabela is usually pretty smart about helping us with good outcomes.
17:36:28 <nickm> isabela: is there a most logical day IYO?
17:37:48 * isabela will crash the party on wed then, if ppl are tired and burned out we dont need to do meetings
17:38:01 <isabela> i could also crash on thursday but wed seems more open
17:38:04 <isabela> from the schedule
17:38:07 <nickm> ok. if no objetions, moving on?
17:38:20 <dgoulet> \o/
17:39:10 <nickm> ok
17:39:25 <nickm> next question is from teor, wanting to know if we should add noise to the padding stats before publishing them
17:39:56 <nickm> mikeperry: do you agree with teor on #22422 ?
17:39:58 <isabela> dgoulet: you going to drl thing right?
17:40:09 <dgoulet> isabela: I am yes
17:40:13 <isabela> :D
17:40:16 <isabela> tx
17:41:23 <nickm> coming back to the sponsor4 question later...
17:41:34 <nickm> there are two things we should do for June prep as noted on the pad.
17:41:36 <nickm> one is take rotations
17:41:51 <nickm> one is making sure that the set of 0.3.1.x tickets we still have is realistic
17:42:09 <nickm> please let's remember to do both?
17:42:48 <mikeperry> nickm: hrmm.. karsten and I had a fair amount of discussion on the rounding. I am not sure it needs noise too, but I can go over the mails with him and add him to the ticket
17:43:16 <nickm> ok.  In my ideal world, you and teor and karsten would reach a consensus about some reasonable thing to do, and I would just do it
17:43:23 <nickm> s/do it/take a patch for it
17:43:30 <nickm> do you think that's likely?
17:45:05 <nickm> Does anybody see any "needs_revision" or "new" or "reopened" tickets on 0.3.1 that it would be a Bad Idea to defer?  If so let's make sure they are assigned, and/or "high" priority.
17:45:57 <catalyst> i think #17605 can probably get downgraded; doesn't seem to be much progress on it either?
17:47:00 <nickm> catalyst: fine by me.  My plan is that everything that isn't "high" or assigned or needs_review will get deferred.
17:48:04 <mikeperry> nickm: yeah. I think that is likely. I will dig up the analysis we already did and see if people agree that the reasonable thing is already done.
17:48:15 <mikeperry> and if not, what we should do instead
17:48:28 <nickm> mikeperry: thanks; I'd be quite grateful
17:48:42 <nickm> ok. I think our last question is the sponsor4 thing?
17:49:04 <ahf> yep
17:49:22 <nickm> our remaining must-do items are to fix whatever bugs we find, and produce some measurement results in chutney, and some estimates in the real network, and a report on what we did.
17:49:32 <nickm> Our remaining can-do items are to make it use even less bandwidth.
17:49:52 <ahf> how much time left do we have on the grant?
17:50:30 <nickm> I believe we have a few more months, but the question is also, "when do we bill".
17:50:56 <nickm> isabela: any guidance here?
17:51:10 <ahf> ok. not sure i understand that? i had the impression we billed them once a month?
17:51:18 <ahf> maybe that is different from sponsor to sponsor?
17:51:39 <isabela> we have till november on the grant
17:51:53 <isabela> my guidance would be to review that plan you wrote back in december
17:51:56 <dgoulet> haxxpop: oh also that's the hs fuzzing ticket: #21509
17:52:23 <haxxpop> dgoulet, ok :)
17:52:40 <isabela> if we are calling it done and i have to write a report to them
17:52:52 <isabela> i would probably report answering each of those phases
17:52:59 <isabela> measure, design, implement
17:53:04 <nickm> isabela: okay.  let's look at https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/Sponsor4Design again after the meeting
17:53:09 <isabela> i guess our 2 designs was 140 and 278
17:53:13 <nickm> and at https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/Sponsor4Plan
17:53:19 <isabela> k
17:53:20 <nickm> any more for this meeting?
17:54:00 <nickm> Thanks, everybody, for another excellent week!
17:54:00 * dgoulet is good
17:54:32 <nickm> everybody please remember to get some rest this weekend, esp if you'll be coming to wilmington
17:54:33 <ahf> looking forward to see a lot of you IRL again next week!
17:54:40 <nickm> likewise
17:54:43 <nickm> #endmeeting