18:00:54 <GeKo> #startmeeting network health
18:00:54 <MeetBot> Meeting started Mon Apr  6 18:00:54 2020 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:54 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:01:01 <GeKo> okay.
18:01:05 <GeKo> hello everyone!
18:01:15 <GeKo> we have our first weekly meeting in april 2020
18:01:30 <GeKo> which means we are in a new quarter, which in turn means a new meeting pad url:
18:01:42 <GeKo> https://pad.riseup.net/p/tor-networkhealth-2020.2-keep
18:01:58 <GeKo> so, please add your items if you have any
18:02:01 <GeKo> ggus: ^
18:02:13 <GeKo> dgoulet: stayed in the room, so i guess we have everyone
18:02:15 <dgoulet> o/
18:02:19 <GeKo> gaba: ^
18:03:14 <GeKo> dgoulet: there is a small item for you on the pad
18:03:30 <GeKo> not sure how that fits into your schedule but if you have time for a small review
18:03:31 <gaba> hi!
18:03:33 <GeKo> that would be neat
18:04:31 <GeKo> i wonder whether we have ggus today. if not i'll catch him after the meeting i guess
18:04:53 <dgoulet> GeKo: sure will review
18:05:00 <GeKo> thanks
18:05:41 <ggus> GeKo: hi
18:05:45 <ggus> i was in another meeting
18:05:48 <GeKo> hola
18:05:53 <GeKo> it's cool
18:06:15 <GeKo> ggus: so the item for you i had: should i get back to the eff folks with what we have for now?
18:06:24 <GeKo> or should i wait a bit longer?
18:06:26 <GeKo> or?
18:06:59 <juga> hi
18:07:15 <ggus> GeKo: i think we should go ahead and reply to eff.
18:07:43 <GeKo> okay, i'll put that on my plate then, thanks
18:07:57 <juga> o/
18:08:09 <GeKo> do we have any other status updates? or comments on any items?
18:08:11 <GeKo> hi!
18:08:27 <ggus> i'll add some topics to discuss
18:08:29 <ggus> one sec
18:09:26 <ggus> ok, done
18:09:53 <GeKo> ggus: okay, i guess we can start with your items
18:10:42 <ggus> for sponsor30 we have an activity to do a workshop on how to install bridges during tor meetups
18:11:01 <gaba> \o/
18:11:03 <ggus> since things are upside down
18:11:10 <ggus> i was thinking if we could have some tor relay meetups on IRC
18:11:46 <ggus> we used to have last year, when colin was around, but we stopped after some issues with timezones
18:12:10 <GeKo> sounds like a good idea
18:12:15 <GeKo> how would that work?
18:12:29 <GeKo> is there an agenda or just an ask me anything style thing?
18:12:37 <ggus> this time i'd try two kind of activities: 1. ask anything, network health topics;
18:13:02 <ggus> 2. one day to support new relay operators; like a hackathon
18:13:38 <gaba> how 2 would work?
18:14:28 <ggus> come to irc and learn how to setup a bridge
18:14:40 <ggus> if you get stuck, we help you
18:14:53 <ggus> if you don't know where to host a bridge, we can suggest some ISPs
18:15:02 <GeKo> i am up for testing that format, sounds interesting
18:15:58 <GeKo> ggus: do you already have a date for that planned?
18:16:37 <ggus> GeKo: i'd like to sync with phw and pili, but we should start this month
18:16:43 <gaba> yes, I like #2
18:17:13 <pili> o/ (I'm in another meeting :) )
18:17:45 <GeKo> ggus: sounds good. let me know if i should help with anything
18:18:11 <ggus> i'll need help on giving support during european timezone :)
18:18:26 <GeKo> count me in ;)
18:18:31 <ggus> wooo! :D
18:19:41 <GeKo> ggus: okay, your other item i guess?
18:20:01 <ggus> yes
18:20:43 <ggus> so our form about relay association is ready. i can start sending to people. or should i itinerate with an organization first and see how it goes and then ask others?
18:21:27 <GeKo> it might be worth starting with one organization and see how it goes
18:21:35 <GeKo> or is it time-sensitive?
18:21:52 <ggus> no, and it's not linked with sponsor activity.
18:22:14 <GeKo> good. yeah, at least that's what i would do then
18:22:23 <GeKo> but i don't have strong feelings here
18:23:25 <ggus> ok! i'll send to one organization this week and see what happens
18:23:36 <GeKo> great, thanks
18:23:51 <GeKo> are we good with that item?
18:24:31 <ggus> yup
18:24:36 <GeKo> thanks
18:24:56 <GeKo> so, we have issues with our bw measurements resulting in tickets like #33775
18:25:08 <juga> yup :/
18:25:15 <GeKo> the question is what to do about those
18:25:27 <GeKo> for context
18:25:50 <GeKo> we have those issues because we don't have enough torflow instances running currently and sbws has still bugs
18:26:01 <GeKo> which we knew about and are working on fixing
18:26:25 <GeKo> but so far they did not hit us because the torflow instances prevented that
18:26:52 <GeKo> ggus: if you hear relay operators complaining about bw issues that's likely that problem
18:27:17 <GeKo> now, we have i think two categories of solutions for that
18:27:37 <GeKo> one short-term hacky one: we try to get this mitigated on the dir auth side
18:27:57 <juga> and temporal
18:28:01 <GeKo> arma2 wrote a mail to the tor-relays list to inform the folks about the options:
18:28:26 <GeKo> https://lists.torproject.org/pipermail/tor-relays/2020-April/018337.html
18:28:51 <GeKo> and the one that takes slightly longer, fixing the sbws bugs
18:29:04 <GeKo> i think we'll end up with both
18:29:25 <GeKo> but maybe there are more options we have not considered yet...
18:30:03 <GeKo> ggus: i think we plan an email to tor-project about the sbws state
18:30:21 <GeKo> but i am not sure whether we should do something more for our relay operators
18:30:38 <juga> tor-project or tor-relays?
18:31:22 <GeKo> tor-relays i meant, sorry
18:31:24 <GeKo> and thanks
18:31:35 <juga> np, just wondered
18:31:38 <GeKo> ggus: but maybe there is more we can do/communicate. i am not sure
18:31:40 <ggus> i think our major relay operators noticed this, but others aren't aware
18:31:56 <GeKo> okay
18:32:16 <GeKo> so, if there are folks concerned and reach out i am fine getting looped in and explain things
18:32:29 <GeKo> including timelines for fixes egc.
18:32:31 <GeKo> *etc.
18:32:48 <GeKo> the current plan for sbws fixing is still end of this month
18:32:48 <juga> it would be too bad if they start shutting down cause of this, so appart of inform them about the situation, might be helpful clarify that is temporal and shutting down would be worst
18:33:05 <GeKo> yeah, i agree
18:33:36 <GeKo> i think arma2 started to reach out to operators and was probably saying something along those lines
18:33:49 <juga> in which channel?
18:33:51 <GeKo> (i am not sure as i did not read the mails/did not get cc'ed)
18:34:09 <juga> ok
18:34:22 <GeKo> we had a discussion today with teor and others in #tor-project
18:34:42 <juga> i read backlog (not sure all of it)
18:34:46 <GeKo> that's all i am aware of regarding irc convos
18:34:51 <juga> ok
18:35:20 <GeKo> any comments/questions? anything else for discussion today?
18:35:55 * ggus thinking
18:35:55 <juga> re. #33775 i think you made a nice brief
18:36:05 <juga> *summary
18:36:14 <GeKo> :)
18:36:32 <GeKo> is there  more we should add?
18:36:51 <juga> re. #33775 or as disscusion here?
18:37:03 <juga> (ie. other point of disscusion)
18:37:12 <GeKo> re #33775
18:37:26 <GeKo> but if there are more topics to add, please do
18:37:28 <GeKo> we still have time
18:38:02 <juga> not sure right now re. #33775, seems we should check if arma2 reached operators and write that email
18:38:36 <juga> there's kind of a 3rd approach that is start having a bwauth with nightly version of sbws
18:38:46 <juga> hopefully improves hings
18:38:46 <GeKo> he reached on of them who replied that things are getting better already
18:38:52 <juga> ah, ok
18:38:52 <GeKo> not sure about nifty
18:39:21 <GeKo> yeah, i think having a bwauth with a nightly could be good
18:39:27 <GeKo> unrelated to that
18:39:48 <juga> ok
18:39:57 <GeKo> so we can easier see whether the last critical fixes we need to deploy have an effect
18:40:01 <GeKo> and if so which :)
18:40:16 <tjr> I run SBWS off of git so I could update it....
18:40:22 <juga> and if it gets better we could change the other sbws to nightly too
18:40:26 <GeKo> i worry there are other important bugs behind those and we don't have thpse discovered yet
18:40:33 <juga> instead of chaning to torflow
18:40:42 <tjr> The main problem is my time to do active debugging is pretty limited.
18:40:43 <juga> yeah, that's totally possible
18:41:11 <juga> tjr: yeah, let's see with thte dirauths, then change if they approve
18:41:18 <GeKo> juga: well, i still think we need the torflow solution right now as we know this will stabilize things
18:41:35 <GeKo> while we might not see improvements with our patches for whatever reason
18:41:37 <juga> GeKo: ok, yeah, you're right
18:41:52 <GeKo> and we can likely not as fast deploy them compared to torflow switching
18:41:53 <GeKo> but
18:42:10 <GeKo> having at least a bw auth tracking sbws's latest nightly code from, say, now on
18:42:25 <GeKo> would give us some confidence that we really have found all blockers
18:42:33 <GeKo> and fixed them as early as we could
18:42:40 <juga> tjr: so the hacky way to have torflow the way arma2 describe is easy for you?
18:42:42 <GeKo> and don't have more outstaning issues
18:42:48 <juga> yup
18:43:05 <GeKo> i think it's easy for stefanie if i saw it right
18:43:19 <GeKo> err stefani
18:43:29 <GeKo> not sure about the other folks
18:43:34 <juga> ah, ok didn't see that, great
18:43:53 <juga> i think the nightly one should be longclow, just cause i can update it fast and debug ;)
18:43:58 <juga> *longclaw
18:44:28 <GeKo> yes, sounds good
18:45:06 <juga> ok, no more ideas on #33775 from my side
18:45:36 <GeKo> okay, it seems we have a bunch of plans
18:45:45 <GeKo> let's see how they work out in the next days :)
18:45:52 <juga> ok
18:45:57 <GeKo> any final items before i call the meeting?
18:46:11 <juga> not from my side
18:46:17 <ggus> i'm good
18:46:45 <dgoulet> all is good
18:47:48 <tjr> If Linus wants to switch to using arma's vote that is up to him. :-)
18:47:56 <GeKo> thanks! stay safe and healthy everyone
18:47:59 <GeKo> #endmeeting