16:05:10 <GeKo> #startmeeting nethealth 08/08/2022 16:05:10 <MeetBot> Meeting started Mon Aug 8 16:05:10 2022 UTC. The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:05:10 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 16:05:27 <GeKo> hello everyone 16:05:35 <GeKo> we gonna resume our weekly syncs 16:05:43 <GeKo> i am sure everonye missed them already :) 16:05:51 <GeKo> juga: hiro: ggus: ^ 16:06:20 <ggus> o/ 16:06:26 <GeKo> we have our pad at http://kfahv6wfkbezjyg4r6mlhpmieydbebr5vkok5r34ya464gqz6c44bnyd.onion/p/tor-nethealthteam-2021.1-keep 16:06:39 <GeKo> please add stuff you have in mind 16:06:45 <GeKo> updates, where you need help with etc. 16:08:06 <hiro> :) 16:08:09 <hiro> I had forgot 16:08:14 <hiro> putting the stuff 16:08:14 <GeKo> see? 16:08:17 <GeKo> :) 16:09:38 <GeKo> i guess i can start with my stuff 16:09:51 <GeKo> i plan to mostly catch up on all sorts of backlog this week 16:10:20 <GeKo> ggus: are you ready for doing the EOL dance for 0.4.6.x? :) 16:10:52 <GeKo> network-health/team#252 is the ticket 16:11:34 <GeKo> if all goes well i might have a list with relay operators to contact this week 16:11:45 <GeKo> and we could start the usual contacting 16:12:01 <ggus> *sigh*, yes i'm ready 16:12:25 <ggus> how many relays are running 0.4.6.x? 16:13:40 <GeKo> 855: 0.4.6 [11.72 %] [12.18 %] (MAJOR) 16:14:03 <ggus> it's a lot. 16:14:08 <GeKo> yep 16:14:27 <GeKo> we could wait a while and then it will be less ;) 16:14:59 <GeKo> but then network team folks told me 0.4.6.x is indeed eol 16:15:06 <GeKo> as nusenu claimed on tor-relays* 16:15:13 <GeKo> s/*/@/ 16:15:18 <ggus> maybe we could push an announcement on social media first, wait one week, then start contacting by email 16:15:26 <GeKo> wfm 16:16:27 <ggus> the problem is not sending 800 emails, but the following up :/ 16:17:02 <GeKo> i know... 16:17:24 <GeKo> so, you do the social media thing this week 16:17:36 <GeKo> and i gather the list at the end of next week? 16:17:48 <ggus> yes, sounds good 16:18:00 <GeKo> great 16:19:39 <GeKo> apart from that i don't have anything to talk about 16:19:41 <emmapeel> like a last round on the bar 16:19:57 <GeKo> yep, exactly :) 16:20:09 <GeKo> anything else from your side, ggus and hiro? 16:20:23 <hiro> I am ok 16:20:32 <hiro> aka groot 16:20:38 <GeKo> :) 16:20:57 <GeKo> there are some drl things upcoming but i guess you have that on your radar 16:21:05 <ggus> GeKo: we can discuss in the week the tor relay op meetup 16:21:10 <ggus> *next week 16:21:16 <GeKo> if you can't make them today/tomorrow i am fine playing a metrics dev, hiro 16:21:25 <hiro> it's today 2000 cest right? 16:21:30 <GeKo> yes 16:21:36 <hiro> yeah I'll be there 16:21:44 <GeKo> ggus: the upcoming, yes 16:21:55 <GeKo> do you know how the one at mch went? 16:22:39 <ggus> GeKo: ahf shared about it in the last community team meeting 16:23:26 <ggus> 14:09:06 <ahf> it went very well 16:23:27 <ggus> 14:09:14 <ahf> we were ~15-20ish people in the c-base village 16:23:31 <ggus> http://meetbot.debian.net/tor-meeting/2022/tor-meeting.2022-07-25-13.59.log.txt 16:23:50 <GeKo> ah, great. i can read that piece later then 16:24:31 <hiro> GeKo, actually can we have a few minutes to discuss check.tpo ? 16:24:45 <GeKo> sure, how are things? 16:24:51 <GeKo> i just saw the tpa ticket today 16:24:54 <hiro> it seems there are memory issues there too 16:24:59 <hiro> I have to look up whats up 16:25:07 <hiro> but the service doesn't have a mantainer 16:27:18 <GeKo> right 16:27:33 <hiro> is that something browser people want to maintain? 16:27:35 <GeKo> we saw that when we did the bullseye upgrade... 16:27:53 <hiro> yeah seems like we had that issue everywhere 16:28:00 <hiro> or should we maintain check? 16:28:12 <GeKo> i think we should bite the bullet and maintain it 16:28:33 <GeKo> as it's tightly coupled with other services we deal with 16:28:50 <GeKo> i don't think browser people need it/want it 16:29:01 <hiro> ok 16:29:13 <GeKo> there is a tor-check via the controller inside the browser 16:29:30 <GeKo> check.tpo is not important anymore for tor browser 16:29:38 <arma2> i think orbot et al might still use it 16:29:47 <GeKo> yes 16:29:50 <arma2> (but i also think that having it maintained inside metrics is a reasonable plan) 16:29:55 <GeKo> i am not saying no is using it anymore 16:30:04 <GeKo> and thus advocating for shutting it down 16:30:22 <GeKo> i recall a bunch of folks noticing when we had issue last time 16:30:37 <arma2> as for what is going wrong with it, i would look at either (A) is arlo's go thing still working efficiently? it was the webserver serving check. and then (2) is it getting the input files from metrics properly and efficiently? 16:30:40 <GeKo> and thus it seems it's still used by folks 16:31:00 <arma2> and then for step 3, i would try to figure out how much irl changed it when they did the tordnsel rerwrite. if any. 16:31:36 <arma2> it used to simply take in a bulkexitlist output, and do a super fast computation on it and serve an answer. it scaled super well. 16:31:45 <GeKo> yeah 16:31:56 <arma2> (this was the replacement over the original cgi python thing, which...did not scale well. :) 16:33:34 <hiro> so well the bottom line is how to we move forward maintaining it? 16:33:45 <GeKo> hiro: i wonder why we suddenly got load issues, though 16:34:07 <hiro> uhm the latest upgrades seems to have caused a lot of issues 16:34:14 <arma2> i would move forward with those three steps i just said. maybe starting with step 3, "what did irl change in the past years if anything" 16:34:44 <GeKo> yeah, i looked at that and there are no changes to the check codebase 16:34:49 <hiro> yeah those are actions, but who takes those actions was more what I was asking 16:34:52 <arma2> and then try to narrow it down between "the webserver itself is having trouble answering questions" vs "the input data is having troubles getting to the webserver" 16:34:56 <GeKo> hiro: right 16:35:06 <GeKo> i am fine dealing with code-side changes 16:35:20 <hiro> I am ok to take another service 16:35:21 <GeKo> i stared at all those bits when we had issues with the bullseye upgrade 16:35:45 <GeKo> and i can write patches and you could review them if needed? 16:36:16 <hiro> that sounds ok, I'd be happy even if you wanted to share the maintenance of the service as a team effort 16:36:27 <GeKo> yep 16:36:33 <hiro> so that if I am out you can take care of it and the other way around 16:36:39 <GeKo> that was what i was about to type into my chat window :) 16:36:53 <hiro> ok then 16:37:07 <GeKo> so, it seems we are set here? 16:37:12 <hiro> yep! 16:37:25 <GeKo> are you going to look at tpa#40842? 16:37:25 <hiro> I'll reply to gaba if that's ok 16:37:34 <GeKo> on the ticket? yes 16:37:34 <hiro> yes that's fine 16:37:49 <hiro> let's see what and why we are OOM-ing everywhere!!! 16:38:05 <GeKo> heh :) 16:38:07 <GeKo> thanks 16:38:22 <GeKo> looks like we are done here for today then 16:38:35 <GeKo> thanks everyone and have a nice week 16:38:38 <GeKo> #endmeeting