14:32:04 <karsten> #startmeeting metrics team
14:32:04 <MeetBot> Meeting started Thu Mar 29 14:32:04 2018 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:32:04 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:32:32 <karsten> https://storm.torproject.org/shared/5h1Goax5eNusxjXJ_Ty5Wl7hFR1uqCReUiN8xdlBG8T <- new agenda pad
14:32:39 <karsten> by the way, iwakeh, does that pad load faster?
14:32:47 * irl 
14:32:51 <iwakeh> yes!
14:32:53 <karsten> hi irl!
14:32:58 <karsten> iwakeh: okay, great, good to know.
14:33:13 <karsten> irl: want to add something to the agenda?
14:33:34 <irl> nope
14:33:39 <karsten> okay. let's start.
14:33:44 <karsten> * metrics-web: can we speed up `ant war` and reduce file size of the generated .war? (karsten)
14:33:55 <karsten> iwakeh writes on the pad:   - This is most likely due to the humongous rs (ticket 25392 and child).  RS should become part of metrics-web.
14:33:58 <iwakeh> I commented on pad.
14:34:08 <irl> i thought 40 seconds wasn't bad for a java web thing
14:34:12 <karsten> hehe
14:34:20 <karsten> but I'm not sure if it's RS related.
14:34:26 <iwakeh> that is the ancient java days before 2k.
14:34:36 <iwakeh> ;-)
14:34:39 <irl> RS does introduce a lot of new files, but they're not particularly big
14:34:45 <karsten> I think it was slow before we added RS.
14:34:51 <karsten> it's all the jetty stuff.
14:34:56 <karsten> I guess my question is:
14:34:59 <iwakeh> Many duplications.
14:35:00 <irl> i was wondering how do we profile it
14:35:17 <irl> and is there some compression happening in the war stage that we can just turn off
14:35:25 <karsten> are there any tweaks that we can do really easily that we haven't done yet, because it didn't matter?
14:35:47 <iwakeh> Well, the integration and straightening out of duplications and different paths
14:35:56 <iwakeh> is a value in itself and might
14:36:07 <iwakeh> also help the creation speed.
14:36:17 <karsten> should I open a ticket?
14:36:23 <iwakeh> There are two
14:36:24 <karsten> (I guess that was really the question I had.)
14:36:28 <karsten> oh, ok.
14:36:31 <iwakeh> #25392
14:36:38 <iwakeh> #25393
14:36:56 <iwakeh> It's just that
14:36:57 <karsten> waiting for trac..
14:37:35 <karsten> ah, I meant tickets for making it faster/smaller if it's not related to RS.
14:37:40 <iwakeh> (trac is slow ...)
14:38:40 <karsten> let me check whether it was much faster/smaller before we added RS, and if not, open another ticket. ok?
14:38:52 <iwakeh> Right, and that way
14:39:03 <iwakeh> we have a reference system,
14:39:10 <iwakeh> your development machine ;-)
14:39:24 <karsten> which doesn't do the job in 40 seconds...
14:39:27 <karsten> okay!
14:39:41 <karsten> alright, moving on.
14:39:43 <karsten> * Onionoo release to fix the bug that nusenu is worried about: mid-April? (karsten)
14:39:52 <irl> which bug?
14:39:53 <karsten> there's an open question on the mailing list.
14:40:04 <karsten> let me find it.
14:40:27 <karsten> https://trac.torproject.org/projects/tor/ticket/22594
14:40:49 <irl> ah
14:40:53 <iwakeh> rather a friendly request.
14:40:53 <irl> yes, we should definitely fix that
14:41:01 <karsten> so, my plan was to just release that with the next major protocol version, which would be around mid-april.
14:41:16 <karsten> yes, it's fixed. but not yet released/deployed.
14:41:24 <karsten> is that too late?
14:41:33 <irl> i think that's ok
14:41:34 <karsten> (the ticket is 10 months old.)
14:41:38 <karsten> was
14:41:40 <iwakeh> I understood nusenu to just wanting to be informed
14:41:45 <karsten> right.
14:41:55 <iwakeh> so, no urgency.
14:41:55 <karsten> okay, I'll tell him that this is our plan then.
14:41:56 <irl> i had a discussion with some people in rome, and onionoo really needs to be more consistent and reliable than it has been historically
14:42:11 <karsten> ah, mind telling more?
14:42:13 <irl> in the next couple of months i want to try and move people from their crazy hacks to using onionoo as a data source again
14:42:24 <karsten> consistent how?
14:42:32 <irl> people that were previously using onionoo had come up with other systems as onionoo had been going down too often
14:42:37 <irl> i mean, consistent with a recent consensus
14:42:54 <karsten> ah ok.
14:42:58 <irl> so not dropping relays because their contact lines are escape characters, would be a thing
14:43:06 <iwakeh> What about the experiment with collector2?
14:43:27 <irl> in general i think onionoo has been getting better
14:43:33 <karsten> so, the contact line thing did not happen in the wild anymore. AFAIK.
14:43:39 <irl> ah cool
14:43:48 <iwakeh> @karsten didn't you start that?
14:43:50 <karsten> the collector2 experiment looked okay.
14:44:03 <karsten> it took longer than expected, because there were still quite a few bridge descriptors.
14:44:21 <karsten> so, if we switch, onionoo will be "down" for a few hours.
14:44:40 <iwakeh> (oh yes)
14:44:42 <karsten> iwakeh: unless you delete more bridge descriptors from recent/.
14:44:53 <karsten> everything except the last three days.
14:45:07 <iwakeh> ok.
14:45:12 <karsten> and we'll have to coordinate this.
14:45:15 <karsten> on both instances.
14:45:48 <karsten> unrelated: corsicum had issues, too.
14:45:51 <iwakeh> true.
14:45:56 <karsten> corsicum = collector2.
14:45:59 <iwakeh> what kind?
14:46:06 <karsten> I really hope we're not switching to a differently broken host.
14:46:18 <karsten> hmm, not sure if I can find out. let me check.
14:47:04 <karsten> no, I can't. I just see that it went down three times yesterday.
14:47:13 <karsten> longest downtime from 11:54 to 14:10 (cest).
14:47:29 <iwakeh> hmm, maybe wait with the switch then.
14:47:39 <karsten> can we fetch from two collectors?...
14:47:45 <iwakeh> Until we know what this was caused by?
14:48:04 <iwakeh> Try one and if it times out the other?
14:48:11 <karsten> try both.
14:48:16 <karsten> or, download from both.
14:48:20 <karsten> and deduplicate internally.
14:48:35 <karsten> we should try out.
14:48:46 <iwakeh> yep.
14:48:59 <karsten> I can do that, but not in march.
14:49:05 <karsten> (fully booked for sponsor 13)
14:49:24 <iwakeh> well, March is over with Easter.
14:49:25 <karsten> so, okay, I'll try out next week.
14:49:36 <karsten> right.
14:50:31 <karsten> okay.
14:50:40 <karsten> so many action items for me!
14:50:45 <karsten> alright, moving on?
14:51:04 <iwakeh> fine
14:51:10 <irl> ok
14:51:10 <karsten> * Sponsor 13 deliverable 2 timeline for the rest of March (karsten)
14:51:25 <karsten> I'm still working on the Relay users part.
14:51:47 <karsten> but, I admit, we're almost out of time this month.
14:51:52 <iwakeh> It is a lot.
14:52:34 <karsten> here's what I'll do:
14:52:47 <karsten> I'll try to finish that part and send it to our sponsor.
14:52:53 <karsten> without publishing.
14:53:16 <karsten> and then we'll do the review/revision dance next month.
14:53:20 <iwakeh> Ah, they are interested in reading the drafts?
14:53:31 <karsten> maybe, I don't know.
14:53:42 <karsten> they want to see what progress we're making.
14:53:53 <iwakeh> isabela does, probably.
14:53:55 <karsten> maybe they'll just count the pages.
14:54:08 <karsten> yes, and isabela wants to know.
14:54:23 <iwakeh> and should.
14:54:26 <karsten> !
14:54:31 <isabela> !
14:54:34 <isabela> oi
14:54:37 <karsten> hello!
14:54:41 <iwakeh> hi!
14:54:56 <karsten> okay, I'll send what I have by end of the month to isabela.
14:55:05 <karsten> and we make it (even!) better in april. :)
14:55:21 <karsten> that's all on this topic from me.
14:55:23 <isabela> yes, sponsor would be happy w drafts too because they can see the process / progress
14:55:30 <karsten> sounds good!
14:55:32 <iwakeh> well, and have another started activity by then.
14:55:33 <isabela> :)
14:55:54 <karsten> iwakeh: true! please share the status of that with me.
14:56:20 <iwakeh> #25644
14:56:48 <karsten> okay, please update that before the month ends.
14:56:54 <iwakeh> karsten: surely.
14:57:03 <karsten> perfect!
14:57:06 <karsten> out of topics?
14:57:18 <irl> i have no more topics
14:57:24 <iwakeh> me neither.
14:57:49 <karsten> great! short meeting. let's talk more next week!
14:58:01 <karsten> happy holidays, and bye, bye!
14:58:06 <irl> bye!
14:58:13 <iwakeh> Happy Easter! bye, bye!
14:58:16 <karsten> #endmeeting