14:00:10 <karsten> #startmeeting metrics team
14:00:10 <MeetBot> Meeting started Thu Oct 20 14:00:10 2016 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:10 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:00:12 <karsten> hi iwakeh!
14:00:33 <iwakeh> pad https://pad.riseup.net/p/3M7VyrTVgjlF
14:00:35 <teor> (thanks metrics team, you do great work)
14:00:45 <karsten> thanks, teor! :)
14:00:49 <iwakeh> thanks :-)
14:02:38 <karsten> okay, I don't have more items for the agenda.
14:02:44 <iwakeh> me neither.
14:02:53 <karsten> okay. let's start:
14:02:54 <karsten> * CollecTor 1.1.0 release plan (iwakeh)
14:03:01 <iwakeh> https://trac.torproject.org/projects/tor/query?milestone=CollecTor+1.1.0&group=status&order=priority
14:03:16 <qbi> nickm: Someone asks if there is a buffer overflow in compat_libevent.c: https://tor.stackexchange.com/q/12960/88 Could you have a look?
14:03:28 <iwakeh> some of the assigned tickets
14:03:36 <iwakeh> should be moved to 1.2.0
14:03:53 <karsten> yep.
14:04:02 <nickm> qbi: taking it over to #tor-project because they're having a meeting now
14:04:12 <iwakeh> all except #19822
14:04:38 <karsten> I was just wondering about that. it's not exactly something that needs to have a milestone.
14:04:54 <iwakeh> that's true.
14:05:06 <iwakeh> close?
14:05:19 <iwakeh> shall I move the others?
14:05:22 <karsten> yep, probably safe to do.
14:05:26 <karsten> we'll do it anyway.
14:05:29 * nickm adds an agenda item, if there's time.
14:05:45 <karsten> nickm: for this meeting? if so, please do.
14:05:51 <nickm> did. :)
14:06:42 <karsten> iwakeh: moving the others sounds fine.
14:06:59 <iwakeh> please update your view.
14:07:21 <karsten> what about #20128?
14:07:48 <iwakeh> postpone.
14:07:52 <karsten> ok.
14:08:03 <karsten> then we have "just" #18910 left.
14:08:09 <karsten> which seems fine.
14:08:25 <iwakeh> yes, it
14:08:37 <iwakeh> could have been ten tickets.
14:08:44 <karsten> heh
14:08:54 <iwakeh> almost 70 comments.
14:09:19 <karsten> but I believe we're pretty close there.
14:09:32 <iwakeh> yes, did you see the tests?
14:09:38 <karsten> yep, looks good!
14:09:46 <karsten> merged to my branch a few minutes ago.
14:09:49 <iwakeh> release date.
14:09:59 <iwakeh> next Monday
14:10:04 <iwakeh> ?
14:10:17 <karsten> hmmmm, I won't have time to do any testing this weekend.
14:10:27 <iwakeh> Wednesday?
14:10:29 <karsten> oh, what about the operator's guide?
14:10:37 <karsten> we should include that in 1.1.0, too.
14:10:54 <iwakeh> Well, it should be finished in October
14:10:57 <karsten> because it's an X deliverable.
14:11:12 <iwakeh> sure.
14:11:15 <karsten> yes, but are we going to put out another release in october?
14:11:27 <iwakeh> that can be finished until wednesday?
14:11:40 <iwakeh> I think yes.
14:11:42 <karsten> yes. it waits for review. ... where is the ticket....
14:11:53 <iwakeh> I know :-)
14:12:01 <karsten> #20380
14:12:14 <karsten> oh, and #20162 can go in 1.1.0, too.
14:12:39 <iwakeh> oh yes!
14:13:02 <karsten> and maybe #20179. (these are all tiny.)
14:13:45 <iwakeh> isn't that merged yet?
14:14:00 <iwakeh> I mean, used on the main instance?
14:14:52 <iwakeh> please, update again.
14:15:12 <karsten> the main instance uses absolute paths only in that script.
14:15:33 <karsten> I guess we could just require that for all four paths.
14:15:45 <karsten> that script will go away in the future anyway.
14:15:46 <iwakeh> so the proposed change is tested. and, yes.
14:15:58 <karsten> yes, it's tested.
14:15:58 <iwakeh> right.
14:16:34 <karsten> looks good.
14:16:37 <karsten> (the query)
14:16:40 <iwakeh> I think I can use the merged code
14:16:51 <iwakeh> on the mirror starting tomorrow.
14:17:02 <karsten> do you want me to squash commits first?
14:17:20 <iwakeh> yes, that'll be better.
14:17:22 <karsten> because, who knows what squashing will break...
14:17:30 <karsten> okay, I'll do that after the meeting.
14:17:32 <iwakeh> yes ;-)
14:18:16 <karsten> okay. I'll also set up a test instance this afternoon/evening.
14:18:41 <iwakeh> as the current mirror is not yet
14:19:13 <iwakeh> public; I think it is save to use it as System-Test-Environment in this case.
14:20:08 <karsten> you mean we don't have to worry too much about breaking it?
14:20:28 <iwakeh> as long as it is not official.
14:20:36 <karsten> sure.
14:20:58 <iwakeh> I'll run a local version first though.
14:21:09 <karsten> the milestone is just to set up one mirror that syncs from another one.
14:21:19 <iwakeh> ok, so it can
14:21:30 <iwakeh> still be an 'alpha' mirror?!
14:21:31 <karsten> we can still think about smarter setups in november.
14:21:37 <karsten> sure.
14:21:44 <iwakeh> that's fine
14:21:58 <karsten> okay, great!
14:22:06 <iwakeh> so wednesday?
14:22:10 <iwakeh> release day?
14:22:12 <karsten> ah, that.. :)
14:22:19 <karsten> yes, sure!
14:22:25 <iwakeh> if all goes fine.
14:22:50 <karsten> yep.
14:23:28 <karsten> okay, anything more on collector 1.1.0?
14:23:39 <iwakeh> that's it.
14:23:53 <karsten> great! let's talk about nickm's item first:
14:23:57 <karsten> * Nickm asks: periodic review on which metrics Tor collects? Lots of stuff in rephist. And geoip. And elsewhere.
14:23:58 <iwakeh> sure.
14:24:09 <karsten> so, where's the milestone we have for little-t-tor stats...
14:24:30 <karsten> 2017-01: Perform an analysis on reducing the amount of sensitive, potentially personally identifying data stored in memory of Tor relays and bridges or reported to the directory authorities. (Sponsor X 4.1. Tor daemon)
14:24:46 <nickm> So, the background here is that I've been going module-by-module and looking at all the functionality.
14:24:53 <karsten> I'd guess that we'll look closely at rephist.c and geoip.c for that.
14:24:56 <nickm> and trying to document it
14:25:21 <nickm> What I'm asking about in particular is not only "is there anything that is risky that we store or report"...
14:25:31 <nickm> .. but also, "is there anything that we just don't use" ?
14:25:51 <iwakeh> what do we store and were is it used?
14:26:10 <nickm> Are you asking me? :)
14:26:17 <iwakeh> rephrasing.
14:26:29 <nickm> Yes, that's a good paraphrasae.
14:26:31 <nickm> Or even
14:26:41 <nickm> "what do we store, where is it used, and is it doing any good?"
14:27:04 <karsten> yes. we're going to look at those questions.
14:27:07 <nickm> great
14:27:13 <karsten> is january 2017 soon enough?
14:27:25 <nickm> if we get answers in january, that should be on time to remove stuff in 030...
14:27:26 <karsten> or would an earlier date be better for some release schedules?
14:27:36 <karsten> what's the 030 schedule?
14:27:57 <karsten> we also have this milestone: 2017-04: Reduce the amount of sensitive, potentially personally identifying data stored in memory of Tor relays and bridges by implementing at least one suggestion from the earlier analysis document. (Sponsor X 4.2. Tor daemon)
14:27:57 <nickm> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam -- completely guesswork.
14:28:25 <karsten> okay, we can probably move around some things in 2017 to better match that schedule.
14:28:32 <nickm> removing stuff is easy; implementing complex stuff is hard.
14:28:49 <iwakeh> better we start dec 2016
14:29:03 <karsten> yep.
14:29:27 <nickm> I would bet a beverage that you find at least one thing we can just remove
14:29:35 <karsten> hehe, yes.
14:29:45 <nickm> Let me know if I can help with code archaeology :)
14:30:13 <iwakeh> is there no git ;-)
14:30:37 <nickm> if git has all the info you need, no need to bug me :)
14:30:42 <karsten> heh. thanks, I guess we'll send a draft of whatever we write to the network team and/or tor-dev@.
14:30:48 <nickm> but if there's stuff that doesn't make sense, just assk
14:30:53 <karsten> sounds good!
14:30:56 <iwakeh> good to know.
14:30:57 <nickm> woo
14:31:12 <karsten> thanks for bringing that up here. :)
14:31:21 <karsten> moving on?
14:31:27 <iwakeh> yes
14:31:28 <nickm> fine by me
14:31:32 <karsten> * Quick tasks that make the October report prettier (karsten)
14:31:48 <iwakeh> we found these tickets for the release .
14:31:55 <karsten> so, I think we're doing a lot with the sync feature.
14:32:03 <karsten> and those smaller features.
14:32:17 <iwakeh> right in addition to the
14:32:23 <iwakeh> operator guide or as
14:32:29 <iwakeh> a link from it
14:32:48 <iwakeh> https://trac.torproject.org/projects/tor/wiki/org/teams/doc/CollecTor/Operation
14:32:55 <iwakeh> just a suggestion
14:32:58 <iwakeh> yet.
14:33:06 <karsten> yep. we do have good stuff there.
14:33:11 <iwakeh> whoever mirrors could enter their mirror.
14:33:19 <karsten> I'm just wondering if we have other almost-done stuff that we could finish easily.
14:33:34 <iwakeh> first draft for
14:33:51 <iwakeh> reproducible build, I took some
14:34:10 <iwakeh> notes today when verifying the metrics-lib 1.5.0 release.
14:34:15 <karsten> neat!
14:34:19 <karsten> very curious about that.
14:34:24 <iwakeh> where shouöd that go?
14:34:35 <iwakeh> 'should'
14:34:48 <karsten> hmmmm. wiki?
14:35:00 <iwakeh> or mail the pre-draft?
14:35:06 <karsten> sure, or that.
14:35:24 <iwakeh> assuming readers know
14:35:37 <iwakeh> the command line and gpg and the like.
14:35:44 <iwakeh> diff
14:35:45 <karsten> here's something else we might do after the CollecTor 1.1.0 release: think about handling bad input.
14:35:48 <iwakeh> ...
14:35:53 <karsten> ah, sure, we can assume that.
14:36:03 <iwakeh> ah, the onionoo ticket?
14:36:07 <karsten> yes.
14:36:09 <karsten> so bad.
14:36:29 <karsten> and the related CollecTor ticket for bridgedescs.
14:36:44 <iwakeh> ok.
14:36:54 <iwakeh> but that won't
14:37:00 <iwakeh> be finished this month.
14:37:02 <karsten> similarly, we could think about log levels. maybe as alternative if we don't come up with anything good for bad input.
14:37:06 <karsten> possible.
14:37:21 <iwakeh> ok. logging will do.
14:37:33 <karsten> mostly brainstorming here.
14:37:36 <iwakeh> as fast solution.
14:37:47 <karsten> yet important.
14:37:58 <iwakeh> shall we discuss that monday again?
14:38:15 <iwakeh> with some more code reading as background?
14:38:18 <karsten> you mean have another meeting on monday?
14:38:26 <iwakeh> or discuss on the tickets?
14:38:57 <karsten> my issues is that I won't have much time to prepare something between now and monday.
14:39:01 <karsten> issue*
14:39:26 <karsten> so maybe ticket for now and talk more next thursday?
14:39:29 <iwakeh> ok, right then just discuss on
14:40:04 <iwakeh> the tickets and if the solution(s) are clear we don't need the meeting.
14:40:32 <iwakeh> depends also on what shows up during the collector tests.
14:40:44 <karsten> oh, true. let's not underestimate that.
14:40:58 <iwakeh> if they run smoothly, there'd be time. if not ...
14:41:34 <karsten> yep!
14:41:36 <iwakeh> the items I copied from last weeks
14:41:45 <iwakeh> action items
14:42:08 <iwakeh> these are all tickets somewhere?
14:42:21 <karsten> we should check. :)
14:45:10 <karsten> alright.
14:45:19 <karsten> we have a couple of action items it seems.
14:45:26 <iwakeh> right.
14:45:42 <iwakeh> back to work?
14:45:45 <iwakeh> :-)
14:45:48 <karsten> sounds good! :)
14:45:53 <karsten> let me squash some poor commits..
14:46:04 <iwakeh> have a nice weekend!
14:46:11 <iwakeh> bye, bye.
14:46:15 <karsten> thanks! you, too, and good luck with testing.
14:46:20 <karsten> bye!
14:46:25 <karsten> #endmeeting