19:00:30 <GeKo> #startmeeting network health
19:00:30 <MeetBot> Meeting started Mon Mar  2 19:00:30 2020 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:30 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
19:00:39 <ggus> hi!
19:00:42 <juga> hi
19:00:44 <GeKo> okay, let's get started with the weekly party
19:00:52 <GeKo> so
19:00:58 <dgoulet> hello
19:01:09 <GeKo> the meeting pad is over at https://pad.riseup.net/p/tor-networkhealth-2020.1-keep
19:01:21 <GeKo> please add your items either in the status section
19:01:23 * ggus will need to leave 19:30 UTC
19:01:26 <GeKo> if you have something to report
19:01:38 <GeKo> or in the discussion one if that fits more in that area
19:01:44 <dennis_jackson> o/
19:01:56 <GeKo> additionally, please look over the entries already written
19:02:06 <GeKo> and mark the items bold you want to talk about
19:02:15 <GeKo> or have questions about
19:02:18 <gaba> o/
19:02:38 <GeKo> juga: nice to see you here :)
19:02:59 <juga> :) (gaba suggested me :)
19:03:09 <GeKo> good suggestion ;)
19:04:31 <gaba> changed the order of the agenda so the guildelines can be discussed with ggus
19:04:37 <GeKo> yeah, thanks
19:04:49 <GeKo> i don't see any bold items
19:05:02 <arma2> (hello world, i am nearby, but also in another meeting. let me know with some latency if there's a thing i should help with. :)
19:05:09 <GeKo> so let's go with the first one in the discussion section
19:05:12 <GeKo> o/
19:05:20 <GeKo> gaba: you brought that up, right?
19:05:47 <GeKo> do you want to go with it?
19:05:48 <gaba> right but because irl wants some feedback there
19:05:59 <gaba> to see if everybody is ok with those changes he is proposing
19:06:28 <gaba> "remove the features that filter by exit policy" https://lists.torproject.org/pipermail/tor-relays/2020-February/018189.html
19:06:46 <GeKo> heh, you changed
19:06:51 <GeKo> the order of the items
19:07:04 <gaba> oops
19:07:06 <gaba> sorry!
19:07:06 <GeKo> and i thought we want to start with the code-of-conduct one first
19:07:09 <gaba> OMG
19:07:10 <gaba> haha
19:07:14 <GeKo> i am fine either way ;)
19:07:15 <gaba> yes yes
19:07:15 <gaba> sorry
19:07:32 <gaba> the guidelines
19:07:49 <gaba> this is something that came up in the tor-relays mailing list
19:08:15 <gaba> to extend the code of conduct with something related to guidelines that relay operators could follow
19:08:30 <gaba> and I added the second link as it was related to something that those guideliens could have
19:08:47 <gaba> a limit on amount of exit relays run by one person for example
19:09:52 <GeKo> ggus: was there some discussion about that at fosdem?
19:09:59 <ggus> GeKo: nope
19:10:28 <ggus> and the only reference about that is this phrase in the mailing list
19:10:36 <ggus> >we (Frënn vun der Ënn) are working on a "Tor Organization Code of Conduct" .
19:11:01 <gaba> ahf had discussion with some people at fosdem about it
19:11:11 <gaba> not sure if we should do anything or just let it be
19:11:16 <gaba> but it is something to be aware of
19:11:33 <dennis_jackson> Something to feed into a possible Guidelines: Some exit relays are behind VPNs. This dramatically increases latency and hurts Tor Browser users.
19:11:36 <ggus> i can contact Christophe and ask how things are going.
19:12:04 <gaba> ok
19:12:30 <dennis_jackson> Are technical issues like that suitable for inclusion?
19:12:44 <arma2> if that, things like 'i promise not to look at the traffic' too
19:14:02 <GeKo> yeah
19:14:19 <arma2> paul even wrote one of these, for relay operators, back when we were first thinking about a code of conduct for tor
19:14:27 <GeKo> i am not sure what actually should be in there if anything
19:14:28 <arma2> might be useful to go find his paragraph. or ask him for it.
19:14:40 <GeKo> and whether we should try to drive that or not
19:15:10 <GeKo> ggus: maybe that is a good next step, yes
19:17:01 <ggus> ok, i can do that today
19:17:03 <GeKo> dennis_jackson: that is interesting
19:17:11 <GeKo> do you have data about that?
19:17:20 <dennis_jackson> Actually - it came from Arthur's dataset
19:17:32 <dennis_jackson> It was something I looked it whilst I was at Mozilla
19:17:33 <GeKo> i think i'd be happy to contact relay operators
19:17:36 <GeKo> aha
19:17:53 <irl> counter point, exits are scarce and removing the ones behind vpns might actually make things worse for everyone
19:18:15 <dennis_jackson> That's true and I'm not suggesting removal
19:18:18 <GeKo> yeah, which is why i'd like to talk to those folks (first)
19:18:33 <dennis_jackson> Just in guidelines, asking people not to would be a good step. They won't be aware of the impact it has
19:18:39 <irl> this might also be a western world viewpoint, and actually the increased latency doesn't matter much compared to access network latency in a lot of the world
19:19:22 <GeKo> well, "not mattering much" does not seem to be a good criteria for not fixing things, though
19:19:28 <dennis_jackson> I have data on this - it is significant
19:19:29 <GeKo> if we think they should be fixed
19:20:23 <GeKo> i mean if we can talk to those folks and get things corrected on their end that seems like a good way of spending a bit of time
19:20:56 <irl> i doubt those exits are using vpns by mistake
19:21:16 <irl> but certainly worth talking to the operators
19:21:16 <GeKo> yeah, maybe. i  guess we'll see
19:22:22 <GeKo> okay, anything else for this point?
19:22:58 <dennis_jackson> So to check: Are these guidelines ethical / social? Or is it general best practice?
19:22:59 <ggus> i'm good
19:23:31 <dennis_jackson> If the former, easy to put these issues in separate tickets, just want to check where to file them
19:23:57 <GeKo> dennis_jackson: i think it's not just best practices
19:23:57 <ggus> i believe the first option
19:24:12 <GeKo> at least from what i understand when reading the mail conversation
19:24:16 <GeKo> yes, what ggus said
19:24:52 <dennis_jackson> okay, thanks :)
19:24:58 <GeKo> okay, irl, you are up
19:25:07 <GeKo> or gaba :)
19:25:12 <gaba> irl :)
19:25:36 <irl> tordnsel has a dnsbl that is used by irc networks and other things to determine if connections come from an exit relay
19:26:01 <irl> it's written in haskell, no one knows haskell, the thing runs on debian oldoldstable and is the last thing on a server we want to turn off
19:26:06 <irl> so we're not keeping it
19:26:11 <irl> we have a replacement thing
19:26:17 <irl> but it's different
19:26:55 <irl> instead of allowing fine grained knowledge of exit policies to determine if both a connection comes from an exit relay *and* it would be allowed in the exit policy, we now only tell you it came from an exit relay
19:27:15 <irl> i've spoken with one irc network and a staff member tells me this makes zero difference to them
19:27:47 <irl> the only people that are going to notice are people that run irc bouncers on their exit relays and deliberately excluded an irc network from their exit policy to not get treated as a tor client
19:27:55 <irl> i can't think of any other examples that would actually happen
19:28:08 <irl> but, maybe someone else can, so here is the question
19:28:31 <irl> is there any reason we should try to do the exit policy handling that tordnsel did before, or is it fine to just have the dnsbl for all exit relays?
19:29:17 <dennis_jackson> Has there been any analysis of the exit policies on the network to upper bound how many people this could impact?
19:29:23 <irl> no
19:29:43 <irl> but i really don't recommend running your irc bouncer on your exit relay host
19:30:08 <irl> i posted to tor-relays@ on the 27th, no responses
19:30:37 <GeKo> from my pov i am fine with the plan
19:30:43 <dennis_jackson> Okay - I will add this to my never ending list of things to look at sometime. (How many non-default exit policies there are)
19:30:54 <irl> GeKo: that's excellent news because i already implemented it
19:30:56 <dennis_jackson> It impacts walking onions as well.
19:31:01 <GeKo> irl: :)
19:31:12 <irl> ok i will write up the service retirement announcement
19:31:21 <irl> this is going to happen in weeks, not months
19:31:21 <GeKo> dennis_jackson: in what way?
19:32:10 <dennis_jackson> To clarify: The uniformity of exit policies impacts Walking Onions, not this decision on dnsel
19:32:21 <GeKo> aha, yes
19:32:35 <GeKo> that's true
19:32:43 <dennis_jackson> Sorry - I am not great at IRC - too spoiled by slack and threading etc
19:33:05 <GeKo> nah, it's cool :)
19:33:28 <GeKo> okay, anything else for this topic?
19:33:35 <GeKo> or do we have other discussion points?
19:34:04 <GeKo> juga: i wonder whether we should talk a bit about sbws?
19:34:12 <GeKo> i wonder if ahf would be around for that
19:35:06 <juga> GeKo: as you prefer
19:35:22 <juga> not sure what to comment from my side
19:36:09 <GeKo> juga: i was just wondering if ahf and i could/should more help than doing code reviews or so
19:36:30 <GeKo> like do you feel there is too much work for you so we should step up here
19:36:35 <GeKo> and help with that?
19:36:55 <juga> GeKo: code reviews are really needed, so that's great help for now
19:37:05 <gaba> there are 3 tickets for review: https://trac.torproject.org/projects/tor/query?status=needs_review&status=needs_revision&keywords=~sbws-roadmap&order=priority
19:37:07 <GeKo> okay.
19:37:29 <gaba> actually, only 2
19:37:33 <juga> (oh, 1 shouldn't be in revewi)
19:37:35 <gaba> #33076 is for mikeperry
19:37:35 <juga> yes
19:37:59 <juga> 1 only :)
19:38:13 <juga> (will create more this week)
19:38:16 <gaba> which one?
19:38:24 <juga> #30196
19:38:31 <GeKo> #30196 seems to be in need for revision
19:38:32 <juga> ahf was on that
19:38:36 <GeKo> according to teor
19:39:04 <GeKo> (which is why i did not look closer at it yet)
19:40:01 <GeKo> or maybe i misread their comment?
19:40:25 <juga> all fine ahf was going to make some comments on that
19:40:35 <GeKo> okay
19:41:04 <GeKo> juga: so, i am fine with poking more at sbws and doing reviews
19:41:05 <gaba> yes. #30726 is for review. Maybe you can do that one? Also there are 2 tickets that are blockers for sbws #30735 and #30735 that needs to be done.
19:41:19 <GeKo> gaba: i already did that one today :)
19:41:28 <gaba> nice :)
19:41:29 <GeKo> or did you mean working on it?
19:41:53 <juga> i just changed #30726 to needs revision
19:42:06 <juga> since GeKo answered a question
19:42:20 <GeKo> juga: and if there are things i can help with beyond doing code review, let me know
19:42:24 <juga> re. tor#30735, i already started working on it, but ist not for review
19:42:36 <juga> GeKo: ok, cool
19:42:37 <GeKo> nice
19:42:52 <gaba> juga: what about #3009? is that something geko can start on?
19:42:57 <gaba> oops
19:43:00 <gaba> #33009
19:43:15 <GeKo> juga: i am fine, too, getting a personal introduction to sbws :)
19:43:30 <GeKo> i can ping you for that
19:43:31 <gaba> :)
19:43:35 <juga> GeKo: yes, we can do that this week
19:43:47 <GeKo> i am not sure when this week is a good time for me
19:43:58 <juga> gaba: re. #33009, yes, if GeKo feels like
19:44:08 <juga> GeKo: just ping me and we see
19:44:15 <juga> or next week if needed
19:44:20 <GeKo> thanks
19:44:57 <juga> idem :)
19:45:11 <GeKo> alright, i might take a look at #33009 this week, we'll see.
19:45:38 <GeKo> okay, anything else for today?
19:46:15 <gaba> not from me
19:46:23 <juga> neither me
19:46:30 <GeKo> let's close this meeting then. thanks everyone! and have a nice week
19:46:37 <GeKo> #endmeeting