14:59:18 <karsten> #startmeeting metrics team meeting
14:59:18 <MeetBot> Meeting started Thu Jun 11 14:59:18 2020 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:59:18 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:59:20 <karsten> hi gaba!
14:59:20 <phw> o/
14:59:24 <karsten> hi phw!
14:59:25 <jnewsome> o/
14:59:30 <karsten> hi jnewsome!
14:59:58 <phw> i keep coming for the friendly welcome. no other meeting has that
15:00:06 <acute> hi!
15:00:10 <karsten> hi acute!
15:00:13 <karsten> phw: yay. :)
15:00:31 <kreyren> tomcat, :p
15:00:51 <karsten> https://pad.riseup.net/p/tor-metricsteam-2020.1-keep <- pad
15:01:19 <karsten> please add topics!
15:02:38 <acute> nothing to add from me
15:05:01 <karsten> okay, we have one more topic on the agenda. let's start.
15:05:06 <karsten> Setting up op-??3 instances
15:05:24 <karsten> we're still running a quite old onionperf version on op-??2 instances.
15:05:44 <karsten> which I think is okay, because we shouldn't update that all the time to have some continuity.
15:05:57 <karsten> still, it may be about time to upgrade.
15:06:14 <karsten> should we start three new instances?
15:06:31 <karsten> I put some more thoughts on the pad.
15:06:43 <karsten> one thing we need to decide is which tor version we're going to run.
15:07:04 <karsten> so far we have been running tor from debian stable. but if we do that, we won't have the #34303 fix.
15:07:22 <karsten> we can either run 0.4.3 or 0.4.4 or master.
15:07:43 <karsten> or wait until that patch gets backported to 0.3.5 and included in debian stable.
15:08:08 <phw> i like the idea of retiring op-??2 one week later. it should give us enough time to notice and react to issues
15:08:16 <acute> I'd go for changing one thing at a time
15:08:41 <karsten> phw: sounds good. fortunately we have anough VMs available to do that.
15:08:55 <karsten> acute: hmm, or maybe groups of things?
15:09:40 <karsten> ideally, I'd like to run these instances for a month without changing them, unless they're broken.
15:09:44 <karsten> again, for continuity.
15:09:55 <acute> it would be good to spot if we're introduced any bugs
15:11:12 <acute> but if there not enough VMs, switching to a newer version of tor is reasonable
15:11:49 <karsten> we can run several VMs.
15:12:08 <karsten> but the idea was to keep op-??2 running for that reason.
15:12:51 <acute> ah, I misread
15:12:54 <karsten> how about we include the first two changes in op-??3 and wait for #34303 to be backported for op-??4?
15:12:56 <acute> sounds like a good plan
15:13:35 <karsten> okay, great!
15:14:04 <karsten> alright, that's all on this topic from me.
15:14:15 <karsten> moving on:
15:14:19 <karsten> Trac/GitLab migration
15:14:44 <karsten> if you can, please take a look at the test migrations and see if anything's going to be really bad for us in the future.
15:15:03 <karsten> other than that, expect not to have trac available over the weekend.
15:15:39 <karsten> any questions on this migration?
15:15:49 <karsten> (not that I could answer them, but I could forward them.)
15:16:03 <gaba> I'm here to answer any question
15:16:22 <gaba> Is there anything right now in the metrics org in gitlab that people are using?
15:16:35 <gaba> http://gitlab.torproject.org/tpo/metrics
15:16:54 <gaba> or we can start from zero from trac into a new Tor Metrics org in gitlab?
15:17:15 <gaba> I see that all the repos there were used 9 months ago
15:17:36 <gaba> so I assume we can remove them. acute and irl were using onionperf but then it got moved
15:17:49 <karsten> I'm not using anything there. acute?
15:17:59 <acute> have not used anything there in a while
15:18:13 <gaba> ok
15:18:15 <karsten> sounds like starting from zero is the way to go.
15:18:21 <acute> yes
15:18:22 <gaba> sounds good
15:18:28 <karsten> cool!
15:18:51 <karsten> anything else on the trac/gitlab topic?
15:19:11 <acute> I look forward to the new workflow :)
15:19:21 <karsten> yes! :)
15:19:52 <karsten> alright, moving on:
15:19:58 <karsten> Any blockers?
15:20:05 <karsten> anyone blocked by anything or anyone?
15:20:19 <karsten> maybe we can discuss something here to unblock you?
15:21:12 <acute> nothing from me
15:21:26 <phw> same
15:21:36 <karsten> okay, great!
15:21:52 <karsten> last topic:
15:21:53 <karsten> How can I tell what trac ID introduced a given line of code? (should we merge branches with --no-ff?)
15:22:00 <karsten> yes, maybe!
15:22:09 <karsten> which project does that? so that I can look?
15:22:28 <phw> some context: i stumbled upon a bug the other day and found it difficult to find out in what ticket i should comment on it
15:22:34 <phw> bridgedb does
15:22:56 <phw> i merge branches with --no-ff, which introduces a merge commit
15:23:07 <karsten> I had so far avoided them. :D
15:23:12 <karsten> but I see your point.
15:23:18 <phw> i don't mean to bikeshed, fwiw
15:23:29 <karsten> no, this is fine.
15:23:52 <karsten> I'll take a look at bridgedb and also read a bit more about this.
15:24:15 <karsten> it might be that I'll do the next merge to some task-xy-merged branch for you to take a look before it goes to master.
15:24:18 <karsten> does that sound okay?
15:24:19 <phw> thanks karsten. we're following this development model with bridgedb: https://nvie.com/posts/a-successful-git-branching-model/
15:24:52 <phw> yes, sounds good!
15:25:35 <karsten> great! thanks for the suggestion. always welcome.
15:25:48 <karsten> okay, do we have anything else for today?
15:26:32 <gaba> not from me
15:26:41 <acute> not from me
15:26:47 <phw> not from me
15:26:59 <jnewsome> do we need to talk about my ticket I handed back?
15:27:18 <karsten> acute: do you have any questions on that?
15:27:36 <karsten> if not, are you still cc'ed on that ticket, jnewsome?
15:27:42 <jnewsome> karsten: yup
15:27:50 <acute> not just now, at the moment I'm collecting tgen 1.0.0 data before moving on to the analysis bit
15:28:02 <karsten> yay :)
15:28:21 <jnewsome> ok, just wanted to make sure it actually was handed off to someone and not lost between the cracks :)
15:28:32 <karsten> ah, jnewsome, do you still need the VM? it's okay to keep it, but if you don't need it anyway, we can turn it off as well.
15:29:16 <jnewsome> hmm, it could be handy, but if it's a scarce resource i can hand it back for now
15:29:32 <karsten> no, it's okay to keep it. if it becomes a scarce resource I'll ask again. :)
15:29:41 <karsten> thanks for taking care of the hand-off!
15:30:04 <karsten> okay, anything else?
15:30:34 <karsten> if not, talk to you next week. thanks, everyone! o/
15:30:39 <jnewsome> o/
15:30:40 <phw> o/
15:30:44 <acute> thank you! bye o/
15:30:55 <karsten> #endmeeting