15:06:05 <asn> #startmeeting SponsorR
15:06:05 <MeetBot> Meeting started Tue May  5 15:06:05 2015 UTC.  The chair is asn. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:06:05 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:06:33 <asn> universities are weird!
15:06:34 <asn> anyway
15:06:41 <isabela> asn: +1
15:06:42 <asn> i'm not going to really chair this meeting
15:06:50 <isabela> I can do it :)
15:06:57 <asn> i also don't have much to report, i've just been chatting on IRC for thep ast wek
15:07:00 <karsten> asn: you should focus on university. talk to you in 11 days! :)
15:07:00 <asn> that's my contribution
15:07:12 <asn> my understanding is that on #15744 we are starting to collect data
15:07:33 <asn> we already have 1 week of into circ data, and now we arestarting to collect lifetime circ data.
15:07:40 <asn> in a few weeks we will be able to visualize stuff.
15:07:54 <karsten> yep
15:08:05 <asn> i have also been running the collector script of #15515 on a box of mine, and i currently have 6k descriptors
15:08:17 <asn> when the time comes (in a month or so) we can incorporate that data on the exisiting graphs
15:08:28 <karsten> is that for those five services?
15:08:32 <asn> yes
15:08:39 <asn> i'm just contiuing the script that donncha run.
15:08:46 <karsten> ok. do you expect any signifant changes?
15:08:52 <asn> who knows. probably not.
15:08:55 <karsten> significant ones even?
15:09:08 <asn> maybe for the agorahoo hs because we don't really know  what's going on there.
15:09:13 <asn> the others should be approx the same.
15:09:24 <asn> so that's for the stats front.
15:09:33 <isabela> cool
15:09:37 <asn> tbh, i think #15744 and #15515 are pretty good stats for the july meeting
15:09:55 <asn> i mean, even with those i would be ok personally. but we could think of more on the months to come.
15:10:02 <asn> (and we still haven't really explored the data of #15744)
15:10:17 * asn stops
15:10:25 <isabela> :)
15:10:28 <isabela> who can go next?
15:10:49 * karsten wrote some quick-and-dirty code for #15744 last week.
15:10:59 <asn> karsten++
15:11:00 <karsten> that's all.
15:11:11 <dgoulet> karsten: I'm running it btw since 2 hours ago
15:11:24 <karsten> dgoulet: cool!
15:11:35 <karsten> (the unit tests don't pass anymore, btw. but I figured that's okay.)
15:11:52 <asn> and we already have like a week+ of introductions histograms data
15:12:01 <asn> since that histogram didn't change on the latest #15744 branch
15:12:02 <dgoulet> I mostly did work on the hs health, improved logging and different tests on HS but the fact that 35% of HSDir are not usable right now makes the result a bit off :S..
15:12:22 <asn> dgoulet: omg yes
15:12:25 <dgoulet> some patches went in Tor also about hs
15:12:33 <asn> about the 35% of HSDirs?
15:12:38 <dgoulet> the hsfetch/hspost command is now upstream
15:12:51 <dgoulet> asn: yeah... the fix is upstream, we just need dirauth to deploy it!
15:13:09 <karsten> oh wow
15:13:14 <asn> hihi. has any deployed it?
15:13:23 <dgoulet> asn: no... we need a tor release for that
15:13:27 <asn> oh yes
15:13:32 <dgoulet> nickm is planning an alpha this week I think
15:13:35 <asn> ok
15:14:02 <dgoulet> I need to fix my branch after nickm's comments on #15881 (needed for hs health)
15:14:29 <dgoulet> I wrote (on paper) what the hs health does right now, I want to make a more detailed "spec" about it and send it asap
15:14:34 <dgoulet> *maybe* this week
15:14:43 * dgoulet done
15:15:00 <isabela> cool
15:15:01 <dgoulet> oh and please cast your vote for #4862 :)
15:15:12 <dgoulet> because I really want to make the patch for it :P
15:15:31 <isabela> anyone else around who wants to go next?
15:15:53 <Yawning> I did hs stuff but dgoulet knows about it >.>
15:16:11 <isabela> share with us!
15:16:35 <Yawning> I fixed a trivial bug in the new GETINFO (merged), and got the HSPOST stuff merged
15:16:47 <Yawning> the first one is R, not sure about the 2nd
15:17:13 <Yawning> but our control port interface to HSes now looks something resemblign sane in master, which is good for everyone
15:18:41 <isabela> ok
15:19:07 <isabela> do we have topics for discussion?
15:19:14 <Yawning> unless HSes catch on fire, I probably won't poke on it as much, since PT stuff is looming over my head
15:19:19 <Yawning> along with other core stuff
15:19:51 <isabela> (sorry Yawning didn't mean to cut you)
15:20:05 <Yawning> it's fine
15:20:08 <Yawning> done anyway
15:20:51 <isabela> discussion?
15:20:57 <dgoulet> hrm I have no topic to discuss in mind...
15:21:03 <karsten> so, are there things left on our todo list?
15:21:52 <isabela> can I ask something?
15:21:53 <nickm> I have a topic, which is what do folks most need from me on HSs/SponsorR over the next week?
15:22:06 <isabela> can we try to update the roadmap? we missed April check ins
15:22:07 <nickm> I can't do _too_ much, but what time I can find I should share.
15:22:10 <Yawning> (what happened to that e-mail convo that asn was having with that dude re intro points)
15:22:40 <dgoulet> nickm: I would say releasing tor so dirauth can deploy #15801, that would be one important thing
15:22:45 <nickm> ok
15:22:48 <asn> Yawning: that's stalled after i told him "hey. i'm really going to make a ticket. please ack." and then i didn't make it and he didn't ack.
15:22:55 <Yawning> oh
15:23:00 <dgoulet> also end of May DonnchaC will start! (woot)
15:23:14 <asn> nickm: maybe give some love to our SoP candidates?
15:23:16 <isabela> \o/
15:23:36 <asn> nickm: like check donncha and kernelcorn (who has been missing past days) proposal, and say what you think?
15:23:43 <dgoulet> nickm: a small comment on #14917 would be nice so I can make a patch :)
15:23:45 <asn> nickm: do you have a SoP suudent this year?
15:23:45 <Yawning> kernelcorn traveling
15:23:59 <karsten> isabela: should we try to update the roadmap together now?
15:24:04 <isabela> sure
15:24:15 <Yawning> no, though my stuff touches core so nickm will get dragged into discussion at some point
15:24:28 <karsten> well, not interrupting other discussions, just throwing it out as discussion/agenda item.
15:24:37 <Yawning> where my is kernelcorn's thing
15:24:56 <asn> Yawning: it's an onion, don't have it saved i think :(
15:25:34 <Yawning> his stuff is all in his github somwhere too
15:25:54 * asn disappears for now. ttyl.
15:26:01 <karsten> ttyl asn
15:26:12 <isabela> cheers
15:26:14 <isabela> ok
15:26:19 <nickm> asn: proposal location?
15:26:30 <isabela> should we try to review the roadmap?
15:26:40 <dgoulet> nickm: https://gist.github.com/DonnchaC/03ad5cd0b8ead0ae9e30
15:26:46 <dgoulet> isabela: yup sure
15:26:49 <isabela> cool
15:27:06 <isabela> lets first look at what was marked as red on march
15:27:19 <karsten> where are you looking?
15:27:35 <isabela> oh sorry
15:27:36 <isabela> !!
15:27:42 <isabela> https://trac.torproject.org/projects/tor/wiki/org/roadmaps/HS
15:27:49 <nickm> dgoulet: that's what i should look over? ok
15:28:10 <isabela> #8243
15:28:16 <dgoulet> nickm: it's DonnchaC's proposal we commented on it and all but feel free to provide love about it :)
15:29:06 <isabela> and "Evaluate PeerFlow design. "
15:29:19 <dgoulet> isabela: discussion have been going on, we have a consensus on that so we should make the patch actually (imo)
15:29:26 <dgoulet> isabela: regard to 8243
15:30:01 <isabela> dgoulet: I will move it to may and maybe asn can pick it up when he is done with exams or should I move it to june?
15:31:05 <karsten> I'd say move to may, and if that's too soon, move to june later.
15:31:10 <isabela> ok
15:31:11 <dgoulet> isabela: realistic to have it in May, I can work on it if he can make it
15:31:12 <karsten> difficult to say without asn.
15:31:21 <dgoulet> can't*
15:31:25 <isabela> cool
15:31:33 <isabela> the second one was assinged to armadev
15:31:58 <isabela> if no one knows ssup with it I can ping him later
15:32:03 <dgoulet> isabela: Peerflow has a critical issue so it's back in reasearch ;)
15:32:15 <dgoulet> because armadev found a show stopper issue :)
15:32:22 <dgoulet> isabela: so I would say "GREEN" :) ?
15:32:26 <isabela> hehehe
15:32:32 <isabela> sure :)
15:32:36 <nickm> dgoulet: wrt 14917 .... hm. let's talk once through with this thread
15:32:49 <dgoulet> nickm: awesome!
15:32:51 <nickm> actually I have to move my car in 7 minutes, so.
15:32:55 <nickm> after that?
15:33:01 <dgoulet> nickm: no problem, I'll be here all day :)
15:33:03 <isabela> this "Think of future stats using top-down approach" is green as well
15:33:08 <isabela> ok we can move to april
15:33:19 <dgoulet> yup
15:33:23 <karsten> hmm
15:33:32 <karsten> what does it mean if something is GREEN but not done?
15:33:41 <karsten> for the march table, that is?
15:33:54 <isabela> it should not be green if is not done
15:34:05 <isabela> well
15:34:11 <karsten> ok, I think we'll have to ungreen #13192 then.
15:34:23 <isabela> maybe the person marked it thinking it would get done by EOM and it didnt happen and the person forgot to update it
15:34:28 <dgoulet> karsten: huh that thing is still problematic?
15:34:33 <karsten> every time I suggest a fix, some crazy person comes up with another floating point edge case..
15:34:43 <Yawning> wut
15:34:48 <isabela> ah
15:35:09 <dgoulet> teor is the last one finding stuff lol...
15:35:22 <dgoulet> ok yeah I guess Yellow then :(
15:35:31 <karsten> I wonder how to proceed.
15:35:46 <dgoulet> karsten: unleash a Yawning on it? :)
15:35:57 <karsten> interesting!
15:36:00 <karsten> hello Yawning!
15:36:01 <Yawning> oh
15:36:02 <dgoulet> ahah
15:36:10 <Yawning> it's test case stuff
15:36:21 <dgoulet> we can ask teor probably also
15:36:25 <Yawning> looks like he posted the solution, someone should write the patches
15:36:52 <Yawning> the actual code itself is fine I think
15:37:12 <karsten> is that something you'd enjoy just a little tiny bit?
15:37:23 <Yawning> maan
15:37:45 <Yawning> it's still part of the "3 days of bank holiday" fun here >.>
15:37:56 <isabela> hehe
15:38:12 <Yawning> suprised teor didn't just attach a branch
15:38:23 <dgoulet> yeah let's ask teor and see his/her answer
15:38:43 <isabela> sounds good
15:38:46 <karsten> ok.
15:39:07 <dgoulet> ok so April, I can comment on " Enable HS stats collection by default. "
15:39:12 <nickm> (okay gotta run; back soon; will ping when back)
15:39:13 <isabela> go for it
15:39:25 <dgoulet> been vetoed in Valencia so it's RED, ticket still open but deffered to 0.2.???
15:39:57 <isabela> should we close it?
15:40:01 <isabela> and take it out
15:40:13 <isabela> (of the roadmap)
15:40:20 <karsten> or only remove from the roadmap?
15:40:23 <karsten> and leave it open?
15:40:26 <dgoulet> take it out yes, I would prefer leaving it open but just deferred
15:41:08 <isabela> ok
15:41:17 <dgoulet> ok that, "Tor EFF hidden service challenge.", I have no idea of the status ahha
15:41:30 <karsten> ah, who asks teor about #13192?
15:41:31 <isabela> hehe
15:41:48 <isabela> I think we can move that to may and then move again.. will follow up with armadev
15:41:49 <karsten> I can ask, but it will be just a short question.
15:41:52 <dgoulet> karsten: we can simply do it on the ticket or when I see teor online next
15:42:04 <karsten> what's your preference?
15:42:05 <dgoulet> isabela: sounds good
15:42:06 <Yawning> if it stalls, I'll fix it
15:42:11 <Yawning> just poke me when that happens
15:42:12 <dgoulet> karsten: ticket
15:42:22 <karsten> Yawning: cool!
15:42:23 <dgoulet> and poke when he/she appears
15:42:34 <karsten> dgoulet: okay, you or me?
15:43:18 <dgoulet> karsten: if you have time, make the comment :)
15:43:23 <karsten> ok, will do.
15:43:32 <isabela> "Add control events for Tor Perf."
15:43:35 <dgoulet> now for Tor Perf...
15:43:59 <Yawning> (wtf is tor perf)
15:44:00 <dgoulet> so we had this idea of trying to use a bit of R time to make torperf 2 but we have not enough people lol
15:44:02 <isabela> lol
15:44:12 <dgoulet> but Rob did start writting something that look like it
15:44:13 <karsten> dgoulet: :(
15:44:29 <dgoulet> but then backed down for ETOOCOMPLEX engeenering problem
15:44:36 <karsten> :(
15:44:47 <dgoulet> karsten: I wish I could fork myself :)
15:44:56 <karsten> so, okay,
15:45:02 <dgoulet> I simply have not enough bw right now to work on that :(
15:45:09 <karsten> in order to get some results for july we can set up the #1944 thing permanently.
15:45:13 <dgoulet> maybe later in May ...
15:45:24 <karsten> I can do that.
15:45:33 <isabela> ok
15:45:48 <karsten> and also try to add results to metrics.
15:46:07 <isabela> I will move to may and add karsten on 'who'
15:46:14 <karsten> dgoulet: no worries, btw, it's not a tiny amount of work to rewrite torperf.
15:46:18 <karsten> isabela: please do.
15:46:37 <dgoulet> karsten: indeed, I read the tech report and it's not a small task
15:46:48 <dgoulet> karsten: especially if we want it maintained and documented :)
15:46:51 <Yawning> I keep getting the sub-codebases confused
15:46:56 <Yawning> is that the bwauth stuff?
15:47:02 <dgoulet> Yawning: nope
15:47:11 <karsten> that's torflow :)
15:47:16 <dgoulet> Yawning: https://gitweb.torproject.org/karsten/torperf.git/?h=inet
15:48:05 <karsten> Yawning: don't look. it's yet another time sink. :)
15:48:16 <isabela> hehe
15:48:20 <dgoulet> "Deploy short term stats." <-- this I'm unsure
15:48:27 <Yawning> a ha
15:48:44 <isabela> hmm
15:48:45 <dgoulet> could it be what we are doing with #15515 and #15744 maybe ?
15:48:51 <isabela> yeah
15:48:58 <isabela> is a follow up from the topdown list i think
15:48:59 <karsten> works for me.
15:49:12 <isabela> ok
15:49:33 <isabela> I will add 15515 and 15744 to may roadmap
15:50:12 <dgoulet> isabela: they are quite ongoing right now, we even have graphs for them so maybe Yellow?
15:50:44 <Yawning> (imma call the next pt implementation 'torstream' just to further confuse things)
15:50:58 <isabela> i can leave that item on april as yellow and add these two for may and we can grade it at the end of the month
15:51:21 <dgoulet> isabela: good idea yes
15:51:25 <isabela> cool
15:51:41 <isabela> next one 'Establish performance baseline with shadow.'
15:52:26 <dgoulet> right so we have a unresolv(able?) issue with shadow and tracing....
15:53:11 <kottman> quickie, after TOR client establishes SSL, and exchanges VERSION, CERTS, NETINFO, the next command have to be CREATE_FAST? or it can be CREATE?
15:53:23 <dgoulet> I still want it, I just don't have it in sight for the next let say 3 weeks :S
15:53:29 <kottman> sorry for bumping in like that :D
15:54:17 <Yawning> *poofs*
15:54:18 <karsten> dgoulet: can you switch to shadow and logging?
15:54:25 <dgoulet> karsten: yeah that would be the way to do it
15:54:45 <dgoulet> karsten: take the tracepoints we have and simply hook a logger to disk on them
15:54:49 <isabela> dgoulet: lets move it to june
15:55:20 <dgoulet> the thing is that with the last DoS attack on HS, we were able to get profile on some HS that were being attacked
15:55:24 <dgoulet> so we have very good results
15:55:40 <dgoulet> and we have a very good idea now where to act on HS performance
15:55:58 <dgoulet> so having shadow results is good but we I don't think we are going to learn much more
15:56:01 <dgoulet> (for now)
15:56:04 <kottman> any good people in :(, little help here?
15:56:22 <dgoulet> kottman: meeting going on here, maybe try in #tor for now, thanks
15:56:27 <isabela> ok
15:56:30 <kottman> ok, thanks :)
15:56:38 <isabela> dgoulet: maybe leave it in april as red then
15:56:54 <dgoulet> isabela: yeah I think so
15:56:57 <isabela> ok
15:57:09 <dgoulet> isabela: things have changed since Valencia and I'm unsure it's really needed for now thus makes sense
15:57:33 <isabela> dgoulet: that is totally fine, this is just to keep track, things does changes all the time
15:57:41 <dgoulet> " Update torproject website to mention Ahmia." <--- that's GREEN
15:58:14 <isabela> by leaving it in april we can go back to it when we think it might make sense again, moving it to june would be 'should be done now or soon
15:58:19 <isabela> cool
15:58:20 <dgoulet> isabela: ok
15:58:23 <dgoulet> isabela: hrm wait
15:58:32 <dgoulet> isabela: there is a website patch but I don,t see it on the website lol
15:58:38 <dgoulet> trying to find the ticket
15:58:47 <isabela> haha
15:59:17 <dgoulet> #15008
15:59:35 <dgoulet> "will wait for some more feedback before pushing." <-- 13 days ago from asn
15:59:38 <dgoulet> that explained it :)
15:59:50 <dgoulet> so yeah GREEN - 1% :P
15:59:59 <isabela> cool
16:00:00 <isabela> :)
16:00:36 <isabela> #3521
16:00:49 <isabela> #13209
16:00:59 <dgoulet> "Deploy HS health measurements." <--- that's Yellow I would say, it's running on my server and collecting data, it's in good shape, I need to add some stuff on it (which I've been working lately on)
16:01:10 <dgoulet> we found 15801 with it
16:01:13 <dgoulet> so you know, progress :)
16:01:24 <isabela> ok
16:01:30 <dgoulet> code is here: https://gitlab.com/hs-health/hs-health
16:01:41 <dgoulet> once I have a "stable" thing, I might request a tpo.org git repo
16:01:49 <isabela> I will move to may, and we can mark it green at the end of the month or when its done
16:02:10 <dgoulet> ok
16:02:29 <isabela> lol
16:02:30 <dgoulet> "Build and execute plan to tell world about Tor and SponsorR" <-- I guess that's done with armadev's email on tor-talk? :)
16:02:34 <isabela> hahaha
16:02:36 <isabela> exactly
16:02:45 <karsten> true
16:03:16 <isabela> Decide whether to allow port-scanning or not.
16:03:23 <isabela> ssup with that
16:03:28 <dgoulet> "Decide whether to allow port-scanning or not." <-- so the fix that dissallow it is upstream but asn reoppened it for consideration
16:03:33 <dgoulet> sooooooo not sure what color :)
16:03:58 <dgoulet> it'S #13667
16:04:00 <isabela> ok, I will move to may and we can add color at EOM when asn is back from exams
16:04:08 <isabela> I will add the ticket as well
16:04:13 <isabela> (thanks dgoulet)
16:04:38 <dgoulet> so there is a bunch of stuff we did in April that is not in that roadmap
16:05:00 <isabela> yes
16:05:08 <dgoulet> all SponsorR ticket like after #15463 in https://trac.torproject.org/projects/tor/wiki/org/sponsors/SponsorR#Tickets
16:05:39 <dgoulet> so the overall "Improvement of HS reachability/scalability/security" went forward quite a bit I think
16:05:47 <dgoulet> just nothing on the roadmap with that title :)
16:06:41 <isabela> true - eventually this roadmap will actually be milestones for each month and this stuff will be there
16:06:46 <isabela> this is a very manual way of doing this
16:06:49 <isabela> is just temporary
16:06:56 <dgoulet> sure np
16:07:01 * isabela needs to move on with updating how we use trac
16:07:18 <dgoulet> isabela: yeah I put the "points" field lately on the new tickets
16:07:26 <isabela> aha! nice
16:07:31 <dgoulet> according to the "legend" you sent
16:07:34 <isabela> that's cool
16:07:38 <dgoulet> https://trac.torproject.org/projects/tor/ticket/15881
16:07:41 <dgoulet> example :) ^
16:07:53 <nickm> dgoulet: I'm back now
16:07:58 <dgoulet> isabela: I guess you have new rules for Keywords? :)
16:08:02 <karsten> what's "small" points?
16:08:15 <dgoulet> nickm: hello!, we are not done with R meeting :(
16:08:25 <nickm> ok. I'll be around
16:08:26 <isabela> we should be tho
16:08:33 <karsten> yeah, I think we're done.
16:08:39 <isabela> #endmeeting
16:08:42 <dgoulet> aaaahh... asn left but opened the meeting!
16:08:43 <dgoulet> lol
16:08:46 <isabela> !!!!
16:09:05 <karsten> nice trick
16:09:11 <dgoulet> he did it on purpose! :P
16:09:23 <isabela> aiaiai
16:09:38 <isabela> oh well, we should be done even tho the bot isnt
16:09:43 <dgoulet> isabela: ok note down the time and strip it from the log latter? lol
16:09:55 <karsten> btw, you know that we added the "Points" and "Actual Points" fields years ago for a PM method we gave up years ago?
16:10:06 <isabela> karsten: that is not official yet, but small should be less than a day
16:10:10 <karsten> we could easily remove them and add new fields as needed.
16:10:42 <isabela> I plan on recreating the ticket form with custom fields
16:10:44 <isabela> like sponsor
16:10:50 <dgoulet> nickm: so 14917 ?
16:10:57 <karsten> isabela: sounds good!
16:10:58 <dgoulet> nickm: what,s your thought?
16:11:07 <dgoulet> isabela: ah great!
16:11:26 <isabela> i just need to get out of the report pile
16:11:27 <isabela> hehe
16:11:32 <isabela> and more coffee!
16:12:07 <isabela> talking about that, I will go get coffee and then update the roadmap
16:12:18 <isabela> ciao
16:12:21 <dgoulet> isabela: cheers
16:12:46 <nickm> dgoulet: I'm not completely sure what I think on that one.
16:13:06 <nickm> One option is that if you pick just one node as your entry, you get a warning
16:13:20 <nickm> but that probably wouldn't help that much under some circumstances
16:13:26 <nickm> what do you favor?
16:14:15 <dgoulet> either accepting exiting at your guard for RP (unsure of the effect on anonimity/security of that) or EntryNodes + HS == Denied
16:14:33 <dgoulet> I would prefer the first one if not dangerous
16:14:34 <asn> #endmeeting