15:58:55 <GeKo> #startmeeting network-health 07/10/2023
15:58:55 <MeetBot> Meeting started Mon Jul 10 15:58:55 2023 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:58:55 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:59:00 <GeKo> hello everyone!
15:59:22 <RishadBaniya[m]> hey there everyone, hope i'm not late
16:00:04 <GeKo> not at all
16:00:16 <GeKo> https://pad.riseup.net/p/tor-nethealthteam-2023-keep is our pad
16:00:20 <hiro> o/
16:00:23 <GeKo> hiro: ggus: ^
16:00:35 <GeKo> i have one item i want to mention here at least
16:00:41 <RishadBaniya[m]> btw GeKo i wanted feedbacks on what i can add to the application or how you want the current application to behave
16:00:49 <RishadBaniya[m]> and about that expensive call
16:00:57 <RishadBaniya[m]> i will have it fixed using a hashset approach
16:01:03 <RishadBaniya[m]> sorry for that
16:01:05 <GeKo> yeah
16:01:10 <GeKo> no worries
16:01:15 <GeKo> we can talk about that in a minute
16:01:39 <GeKo> let me just get the "process" question sorted out
16:01:41 <GeKo> :)
16:01:47 <RishadBaniya[m]> Sure :)
16:02:03 <GeKo> so, i'll be out for about 4 weeks from next monday on doing vacations
16:02:13 <hiro> nice
16:02:22 <GeKo> we need to figure out what we should do with our weekly syncs here
16:02:29 <GeKo> during that time
16:03:04 <hiro> I'd be happy to do sync w people, but that also depends if others will be out as well
16:03:15 <GeKo> it seems keeping those running while i am away could be useful
16:03:33 <GeKo> in particular as it allows coordinating our gsoc work i think
16:03:50 <hiro> ok I can call the meetings then
16:03:54 <GeKo> hiro: would you be up for faciliting these syncs until im back?
16:04:01 <GeKo> okay, nice
16:04:02 <GeKo> thanks
16:04:33 <GeKo> my last vacation day is 08/15 and i plan to catch up that week while i am at camp
16:04:50 <GeKo> and i can take over the usual tasks again on 08/22
16:04:56 <GeKo> hiro: would that work?
16:05:08 <hiro> yep sounds good
16:05:21 <GeKo> awesome, thanks
16:05:38 <GeKo> actually last vacation day is 08/14 and i am back 08/21
16:06:27 <GeKo> that was easy :)
16:06:41 <GeKo> okay any updates worth discussing here?
16:06:55 <GeKo> i guess we can start with gsoc
16:07:05 <hiro> matt is traveling this week
16:07:16 <GeKo> RishadBaniya[m]: 2023-07-10T15:43:23Z INFO  master::master::tor_network] Finished filtering in 1614.019235026s
16:07:20 <GeKo> is where i am stuck now
16:07:23 <GeKo> hiro: ack
16:07:36 <GeKo> RishadBaniya[m]: so, the filtering took almost half an hour
16:07:41 <GeKo> which is...long
16:07:41 <RishadBaniya[m]> woah!
16:07:42 <RishadBaniya[m]> was that on debug mode
16:07:49 <RishadBaniya[m]> its too long
16:07:56 <RishadBaniya[m]> for release mode
16:08:00 <GeKo> cargo run --bin master
16:08:07 <RishadBaniya[m]> that's on debug mode
16:08:08 <GeKo> is what i've been running
16:08:09 <GeKo> yeah
16:08:19 <RishadBaniya[m]> on release it gets done in 1 min
16:08:26 <RishadBaniya[m]> cargo run --release --bin master
16:08:37 <RishadBaniya[m]> sorry for not mentioning that debug takes way long
16:08:49 <RishadBaniya[m]> in the WiKi
16:09:22 <GeKo> no worries
16:09:27 <RishadBaniya[m]> even around 1 minute of time is too much overhead here, i realized where i messed up, i'm fixing it ASAP
16:09:45 <GeKo> let me see how fast this is in release mode
16:09:53 <GeKo> okay
16:09:58 <RishadBaniya[m]> Sure :)
16:10:05 <GeKo> it's worth adding that in the ticket
16:10:14 <GeKo> that is what went wrong and what the proposed fix is
16:10:17 <GeKo> so we have some context
16:10:46 <RishadBaniya[m]> Sure i'll raise an issue and add contexts
16:10:48 <RishadBaniya[m]> right now
16:10:52 <GeKo> RishadBaniya[m]: i've looked over https://gitlab.torproject.org/rishadbaniya/erpc/-/wikis/home
16:10:53 <RishadBaniya[m]> after the meeting
16:10:58 <GeKo> thanks for the write-up
16:11:18 <GeKo> i'd like us to focus on scenario (i) right now
16:11:22 <GeKo> and get that one right
16:11:30 <GeKo> there is no need to back out code or anything
16:11:49 <GeKo> just make (ii) and (iii) as experimental for now?
16:12:10 <GeKo> those are interesting optimizations but we should not focus on them for now as they make things only more complex
16:12:32 <GeKo> oh, the master/slave terminology made me cringe
16:12:35 <RishadBaniya[m]> Sure :)
16:12:43 <GeKo> we should replace that with something like parent/child or so
16:12:51 <GeKo> i can file a ticket for that later
16:13:24 <RishadBaniya[m]> Sure :), it does sound cringe
16:13:27 <RishadBaniya[m]> parent child seems nice
16:14:35 <GeKo> i do think some instructions to run just scenario (i) in the README file are good
16:14:52 <GeKo> i saw you moved those into the wiki
16:15:05 <GeKo> while i think the wiki is fine for detailed information
16:15:14 <RishadBaniya[m]> I had trouble adding images, so i had moved
16:15:20 <RishadBaniya[m]> to WiKi
16:15:26 <GeKo> yeah
16:15:32 <GeKo> it's fine to have that part in the wiki
16:15:53 <GeKo> but something like `cargo run --release --bin master`
16:16:08 <RishadBaniya[m]> noted! i will add that
16:16:10 <GeKo> should be in the README as well
16:16:42 <GeKo> i guess essentially everything in https://gitlab.torproject.org/rishadbaniya/erpc/-/wikis/home#running-master-worker could be mentioned there
16:17:24 <RishadBaniya[m]> I'll add it properly in the README after the meet
16:17:30 <GeKo> thanks
16:17:36 <GeKo> okay, now we have: 2023-07-10T16:16:52Z INFO  master::master::tor_network] Finished filtering in 100.408107542s
16:17:41 <GeKo> much better :)
16:18:05 <GeKo> now i am stuck there. is there supposed to be further output about what's going on?
16:18:23 <RishadBaniya[m]> actually it's now producing
16:18:25 <RishadBaniya[m]> 2 hop circuits
16:18:26 <RishadBaniya[m]> in the database
16:18:28 <RishadBaniya[m]> file
16:18:44 <GeKo> is there some more output i could get on the terminal to see what's going on?
16:18:48 <RishadBaniya[m]> i had no idea what to log now, in the master worker only mode, because if i were to log all circuit creation attempts
16:18:55 <RishadBaniya[m]> log would be huge!
16:19:21 <GeKo> well. you could create an option for a log file
16:19:45 <GeKo> which would log circuit creation attemps and what there results are
16:20:12 <RishadBaniya[m]> i will add that too
16:20:12 <GeKo> together with some information like: scheduling X circtuis
16:20:16 <GeKo> *circuits
16:20:22 <RishadBaniya[m]> currently the sqlite3 database has the data saved
16:20:25 <RishadBaniya[m]> i'll also add the table name and other information in the WiKi
16:20:28 <GeKo> or donex out of y circuits
16:20:29 <RishadBaniya[m]> sure, that would be better
16:20:47 <RishadBaniya[m]> i will have that added and juga had also mentioned to add
16:20:54 <RishadBaniya[m]> arguments for path to config file
16:21:01 <GeKo> some way for follow what's going on without investigating the database
16:21:01 <RishadBaniya[m]> i'll also add a log file argument
16:21:07 <GeKo> great
16:21:18 <GeKo> i think that's my feedback for now
16:21:26 <RishadBaniya[m]> thanks for the feedback, now i know what i should fix
16:21:26 <GeKo> i'll add more tomorrow and will file some tickets
16:22:00 <RishadBaniya[m]> btw after fixing these issues, should i start working on extracting circuit informations through onionperf tool
16:22:31 <RishadBaniya[m]> is that the next step i should move on with the application
16:22:40 <GeKo> if things are working well, we could think about that
16:22:56 <GeKo> let me test what you have right now a bit more
16:23:07 <GeKo> and let's file some issues following from that
16:23:16 <RishadBaniya[m]> Sure :)
16:23:23 <GeKo> and once we are happy with the current state we can consider the onionperf part
16:23:29 <GeKo> does that make sense?
16:23:37 <RishadBaniya[m]> yeah i got you
16:24:00 <GeKo> so, the debug output part is highest prio right now
16:24:15 <GeKo> so we can see what is going on and figure out whether something more needs fixing
16:24:33 <GeKo> no worries about the size of the logging file
16:24:48 <RishadBaniya[m]> i will set the scheduling X circuits
16:24:50 <RishadBaniya[m]> in the debug
16:25:04 <GeKo> you could make that approach smarter by having some things in INFO mode and some in DEBUG mode
16:25:25 <GeKo> ideally, with the INFO mode allowing to see what is going on without taking too much space
16:25:25 <RishadBaniya[m]> sure i'll will consider that as well
16:25:42 <GeKo> and all the nitty-gritty details for bug hunting behind DEBUG
16:26:01 <GeKo> and then one can consider on how large a log file one wants to have :)
16:26:15 <RishadBaniya[m]> that seems nice
16:26:20 <RishadBaniya[m]> so the log level decides
16:26:20 <GeKo> great
16:26:22 <RishadBaniya[m]> the size eventually
16:26:24 <GeKo> yeah
16:26:32 <RishadBaniya[m]> i love this approach
16:26:48 <GeKo> good, good :)
16:26:55 <GeKo> RishadBaniya[m]: anything else you need from us?
16:27:20 <RishadBaniya[m]> i guess that's everything for now, i will ping you if i have any query
16:27:47 <GeKo> i think we can revalidate the filtering approach you have, which takes some time
16:28:16 <GeKo> i am in particular interesting in understanding exactly how a new consensus fits into that picture
16:28:24 <GeKo> *interested
16:28:36 <GeKo> i'll look closer at that this week as well
16:28:37 <RishadBaniya[m]> should i document
16:28:39 <RishadBaniya[m]> that in the WiKi?
16:28:43 <RishadBaniya[m]> i can do that too
16:28:43 <GeKo> yes, please
16:28:55 <RishadBaniya[m]> sure, i'll have it documented
16:28:57 <GeKo> that's actually a crucial part of the tool
16:29:19 <GeKo> and one of the main tasks imo: to figure out a good strategy here, balancing different goals
16:29:39 <RishadBaniya[m]> yeah, about that as of right now
16:29:44 <RishadBaniya[m]> i've decided the most generic
16:30:06 <RishadBaniya[m]> only by considering
16:30:15 <RishadBaniya[m]> if two relay can create circuits
16:30:26 <RishadBaniya[m]> it still lacks repeated testing of same circuit
16:30:37 <RishadBaniya[m]> i.e same relay combination but i'll have that design ready too
16:31:06 <RishadBaniya[m]> i'm yet to utilize the flags here, now that the entire framework for the application has been laid, adding things on top of it is a bit easier
16:31:14 <GeKo> documenting what is missing in that context and what you plan in the wiki is  important as well
16:31:22 <GeKo> okay
16:31:50 <GeKo> oh, another important thing: last week was the bi-weekly update due to our mailing lists
16:31:52 <RishadBaniya[m]> Sure :) i'll make sure to document crucial parts, please do provide me feedbacks on things that i should document
16:31:56 <GeKo> please do that this week
16:32:11 <RishadBaniya[m]> i have it ready, few changes to be made
16:32:17 <RishadBaniya[m]> i'll mail
16:32:21 <RishadBaniya[m]> it tommorrow
16:32:22 <GeKo> and update the timeline as well with what you did last week
16:32:33 <RishadBaniya[m]> Sure :)
16:32:37 <GeKo> (the accomplished part is missing)
16:32:44 <GeKo> oookay
16:32:56 <GeKo> do we have anything else for today?
16:33:18 <RishadBaniya[m]> i guess that's it from my side...
16:33:26 <GeKo> hiro: our usual sync time this week works?
16:33:39 <hiro> yep
16:33:51 <GeKo> for the s112 parts: note we have our monthly sync tomorrow 1600utc at our bbb room
16:33:56 <GeKo> nice
16:34:07 <GeKo> hiro: you good otherwise?
16:34:29 <hiro> yep all good
16:35:07 <GeKo> okay, let's finish this sync then. thanks eveyone
16:35:10 <GeKo> #endmeeting