16:59:06 <ahf> #startmeeting Network team meeting, 11th April 2022
16:59:06 <MeetBot> Meeting started Mon Apr 11 16:59:06 2022 UTC.  The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:59:06 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:59:09 <ahf> hello hello
16:59:12 <mikeperry> o/
16:59:17 <ahf> pad is at https://pad.riseup.net/p/tor-netteam-2022.1-keep
16:59:28 <GeKo> o/
16:59:35 <jnewsome> o/
16:59:36 <nickm> hihi
17:00:38 <ahf> i am semi afk this week to finish some work on our flat but i will be around for some of the post-15 utc meetings this week, though no 1:1 as per my email
17:00:39 <ahf> ok
17:00:44 <ahf> how are people doing with their boards?
17:00:55 <ahf> https://gitlab.torproject.org/groups/tpo/core/-/boards
17:01:06 <dgoulet> o/
17:01:11 <ahf> o/
17:02:10 <ahf> you can bluff the hell out of me on this question today since i cannot access gitlab for some reason
17:02:11 <ahf> lol
17:02:12 <ahf> lame
17:02:32 <ahf> i will do a pass later today and see if anything looks messy
17:02:36 <dgoulet> it worked!
17:02:38 <ahf> dgoulet: how are we doing with releases?
17:02:43 <nickm> hmm. I think I have enough to carry me through the end of the week (when I go on break)
17:03:11 <juga> o/
17:03:22 <ahf> very nice
17:03:29 <dgoulet> The -rc is out since last week! So far so good! Seems quite stable. health team, mikeperry and I are working on enabling CC network wide before stable
17:03:41 <ahf> VERY nice
17:03:43 <dgoulet> goal is really stable before end of month so fingers crossed! :)
17:03:54 <ahf> \o/
17:03:59 <ahf> very excellent
17:04:00 <dgoulet> done
17:04:23 <ahf> i cannot do any passes on incoming tickets but will check that once i am home
17:04:27 <ahf> mikeperry: wanna do s61?
17:04:33 <mikeperry> ok
17:04:38 <eta> o/
17:04:48 <ahf> o/
17:05:03 <mikeperry> yeah, so to turn on congestion control, we need the dirauths to set cc_alg=2. we can do that now, I believe
17:05:33 <mikeperry> we need to make sure none of the bwauths upgrade to 0.4.7 before also updating to sbws with https://gitlab.torproject.org/tpo/network-health/sbws/-/issues/40125
17:05:58 <mikeperry> but otherwise, it should be fine
17:06:16 <ahf> this means the dirauth's needs to upgrade to -rc too?
17:06:29 <mikeperry> they should not need to
17:06:34 <ahf> ok, great
17:06:37 <mikeperry> just set the consensus param
17:06:43 <ahf> *nods*
17:07:00 <jnewsome> dirauths: "do as I say not as I do"
17:07:35 <ahf> XD
17:07:46 <mikeperry> the other piece is that hiro's 10MB sims look good. I was worried that throughput might fall with larger downloads; it did not. it increased
17:08:04 <mikeperry> though we need to fix the onion svc graphing it seems
17:08:28 <ahf> awesome
17:08:31 <mikeperry> and probably try 20/50MB just to check things out
17:09:19 <mikeperry> the next step for me iunit tests and test vectors
17:09:49 <mikeperry> let's hope this week has less admin/grant surprises than the last 7 weeks. fingers crossed
17:09:54 <ahf> nice, will probably make the arti gang very happy too
17:10:01 <ahf> hah
17:10:03 <jnewsome> mikeperry: I implemented the simulation option for adding legacy clients last week. sent to you for review
17:10:24 <mikeperry> oh yeah I saw that. I will look that over
17:10:26 <jnewsome> https://gitlab.torproject.org/jnewsome/sponsor-61-sims/-/merge_requests/35
17:11:03 <mikeperry> thanks. that should allow us to test a few things wrt tuning post-stable as well
17:11:37 <ahf> nice
17:11:37 <mikeperry> we have our report due at the end of the month, remember
17:11:45 <mikeperry> so we will need to add that stuff to the report pad
17:13:02 <mikeperry> I still need to check in with ggus about comms and outreach for the exit upgrade
17:13:16 <mikeperry> GeKo,juga: anything else from network-health/sbws?
17:13:38 <juga> mikeperry: For sbws#40132: how many exits or bandwidth we consider enough?
17:14:08 <GeKo> nothing from my side
17:14:51 <mikeperry> juga: I think we just need the fastest exits to upgrade, so there are exits that meet that 2X threshhold you mentioned last week, right?
17:15:04 <juga> yeah, ok
17:15:53 <mikeperry> since we can control the bwauth use of congestion control separately, we can also watch for overload at those exits when we flip that switch
17:16:08 <juga> ok
17:16:13 <mikeperry> so there's less risk, so long as we're watching that stuff
17:16:33 <mikeperry> dgoulet: we should backport that overload-general ntor fix before the sbws switch too
17:16:53 <dgoulet> ok and by that you likely mean do an 046 release in the process?
17:16:53 <mikeperry> dgoulet: https://gitlab.torproject.org/tpo/core/tor/-/issues/40560 -> 0.4.6
17:17:12 <mikeperry> yeah, somewhere around when 0.4.7-stable is released is likely fine
17:17:27 <dgoulet> ack! I'll plan that
17:17:33 <ahf> thanks
17:18:13 <mikeperry> I think that is it. any other questions?
17:18:22 * juga is good
17:19:02 * ahf good
17:19:13 <dgoulet> all good
17:20:18 <ahf> let's call it then
17:20:21 <ahf> thanks all o/
17:20:26 <ahf> #endmeeting