12:00:26 <kshlm> #startmeeting Weekly community meeting Nov 2, 2016
12:00:26 <zodbot> Meeting started Wed Nov  2 12:00:26 2016 UTC.  The chair is kshlm. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:00:26 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
12:00:26 <zodbot> The meeting name has been set to 'weekly_community_meeting_nov_2,_2016'
12:00:33 <kshlm> Okay. Let's start.
12:00:44 <kshlm> I'll wait for 2 minutes for people to filter in.
12:00:50 <nigelb> o/
12:01:03 <kshlm> In the meantime, if anyone wants to add last minute updates to the agenda, please do it.
12:02:09 * kkeithley is here
12:02:19 * jdarcy o/
12:02:29 <kshlm> kkeithley, nigelb, jdarcy Hey!
12:02:51 <kshlm> kkeithley, nigelb, you guys can add your updates on Ganesha and infra if you have any.
12:03:27 <nigelb> I have just done so.
12:04:11 <kshlm> nigelb, Thanks.
12:05:03 <kshlm> Let's start with a quick rollcall.
12:05:07 <kshlm> #topic Rollcall
12:05:19 * kshlm _o/
12:05:28 <nigelb> o/
12:05:33 * rastar is here \o
12:05:41 * jdarcy \o
12:05:49 * anoopcs is here
12:05:50 <jdarcy> It's important to exercise both arms evenly.
12:06:05 * ppai is here \o
12:06:20 <kshlm> jdarcy, \o/
12:06:42 <kshlm> Let's continue.
12:06:49 <kshlm> #topic Next weeks host
12:06:58 <kshlm> Do we have a volunteer to run the next meeting?
12:08:05 <kshlm> No one?
12:08:48 <kshlm> Okay then. I'll continue as the host.
12:09:07 <kshlm> Let's move on to...
12:09:11 <kshlm> #topic Open Floor.
12:09:25 * rafi is here
12:09:26 <kkeithley> I'll do it if nobody else will.  I'll have chocolates next time I come to BLR (first week December) for volunteers and people who review my 6 or 7 patches
12:09:46 <kshlm> Just one new topic apart from the ones carried over from last week.
12:10:22 <kshlm> kkeithley, You can take the week after the next.
12:10:27 <kshlm> I'll add a note about it.
12:10:28 <kkeithley> sure
12:11:02 <kshlm> Okay.
12:11:14 <kshlm> So which topic do we want to discuss first today.
12:11:18 <kshlm> Options,
12:11:18 * rafi is looking into kkeithley patches ;)
12:11:28 <kshlm> - Recognising community contributors(Manikandan Selvaganesh)
12:11:38 <kshlm> - Sending up EasyFix bugs(assigned with a owner) on mailing list so that new contributors can get started(May be?) (Manikandan Selvaganesh)
12:11:49 <kshlm> - FSFE pad service is about to be decomissioned. Need to find an alternative.
12:12:05 <kshlm> Is Manikandan here today?
12:12:19 <nigelb> not that I can see.
12:12:37 <kshlm> nigelb, Neither can I.
12:13:02 <nigelb> For the last one, there's http://beta.etherpad.org/
12:13:13 <kshlm> We were discussing the 1st topic last week, we can take that up last.
12:13:40 <kshlm> I'll start with
12:13:47 <kshlm> #topic FSFE pad service is about to be decomissioned. Need to find an alternative.
12:13:56 * samikshan joins late
12:14:08 <kshlm> nigelb, http://beta.etherpad.org/ it says beta in its name.
12:14:27 <kshlm> How stable would it be?
12:15:09 <nigelb> As long as we use it during meetings and not as primary data storage, I don't see a big problem with it.
12:15:14 <jdarcy> I assume we don't want to run our own Etherpad server.
12:15:39 <nigelb> Having helped run one, let's just say, we don't.
12:15:57 <jdarcy> OK, now that's part of the meeting record.  :)
12:16:03 <kshlm> We do use the etherpad as a rolling log of our meetings.
12:16:09 <anoopcs> https://pad.riseup.net/
12:16:22 <jdarcy> Other stupid idea: Google Docs
12:16:53 <kshlm> A new pad there comes with a big warning about things getting destroyed frequently.
12:17:05 <nigelb> That's the only way to deal with the spam :)
12:17:37 <nigelb> The best way is to have rolling logs on a wiki page (on github/gluster/gluster) and current meeting logs on etherpad.
12:18:03 <kshlm> That's more work for me :(
12:19:05 <nigelb> The other option, as jdarcy said, is google docs. Either way, I don't think we should waste a lot of time on this.
12:19:24 <kshlm> But yeah. I agree that we shouldn't be using the pad as log.
12:19:40 <kshlm> I'd like to get a decision on this right now.
12:19:40 <ppai> there's this too - a long list of options: https://github.com/ether/etherpad-lite/wiki/Sites-that-run-Etherpad-Lite
12:19:44 <amye> How many things will we lose as the pad service is discontinued?
12:20:45 <kshlm> amye, We've been using it as a dumping ground for a lot of things.
12:20:52 <amye> Yeah, I know. :D
12:20:53 <jdarcy> I have three Etherpad tabs open right now.  Sometimes it's a dozen.
12:21:17 <kshlm> not a lot of that is indexed anywhere either.
12:21:43 <amye> That's why I'm a little concerned that it's going to get reaped and we're never going to know what we lost - it's the lack of indexing that makes me have grouchy feelings about etherpad. :/
12:22:16 <jdarcy> Likewise.
12:22:26 <jdarcy> At least with GDocs, you have a list of what's there.
12:22:52 <jdarcy> Also, GitHub wiki.
12:23:11 * obnox arrives late
12:23:14 <nigelb> Github wiki is where I'd like things on etherpad to end up after the need for live editing is over.
12:23:39 <kshlm> +1 to nigelbs plan.
12:24:06 <amye> I'll vote for github wiki - it's public, it's something we can index -- what do we do for live editing in the meantime?
12:24:10 <jdarcy> So use $whatever when live, move to GitHub wiki when a doc stabilizes?
12:24:55 <nigelb> I'm happy to put today's meeting notes on the wiki.
12:24:58 <nigelb> And see how that works out.
12:25:40 <kshlm> nigelb, I'll help with that.
12:25:48 <rastar> github wiki ++
12:26:41 <kshlm> We need to make sure this is documented.
12:26:47 <jdarcy> On the wiki.
12:26:50 <nigelb> :D
12:26:55 <obnox> jdarcy: ;-)
12:27:30 <kshlm> nigelb and I can take care of this.
12:27:39 <kshlm> nigelb, What do you say?
12:27:50 <jdarcy> Well, Etherpad's not going away *today* so we can put something there as well.
12:28:03 <nigelb> sounds good.
12:28:40 <kshlm> jdarcy, I can put a notice on the meeting etherpad, but we don't know how many more there are.
12:29:21 <amye> Hmm. That, I don't know how we solve.
12:29:34 <kshlm> #action nigelb, kshlm, will document and start the practice of recording etherpads into Github wikis.
12:29:34 <amye> How long do we have?
12:29:40 <kshlm> amye, Dec 1 it turns read only
12:29:46 <kshlm> Feb 1 everything is gone.
12:29:57 <amye> Ok then.
12:30:21 <kshlm> I'll also send out an email about this, so everyone else is informed.
12:31:11 <kshlm> #action kshlm to send out an email about Etherpad+Wiki after it's done for the first time.
12:31:26 <kshlm> We can find our etherpad alternative later.
12:32:02 <kshlm> Let's try to first preserve what ever we can find.
12:32:10 <amye> Thank you
12:32:25 <kshlm> I'll move on to the next topic, if there is nothing else to discuss about this.
12:33:00 <kshlm> Okay.
12:33:10 <kshlm> nigelb, Is the new topic yours?
12:33:19 <kshlm> #topic Glusto Job on Centos CI
12:33:21 <nigelb> Yep.
12:33:28 <kshlm> You have the floor.
12:33:49 <nigelb> In followup to Jonathan and ShwethaHP from the last meeting, we have the job setup on centos CI.
12:34:07 <kshlm> That is awesome news.
12:34:16 <nigelb> It doesn't work fully yet. I belive it needs a bit of tweaking, but we should have a working setup this week or so.
12:34:27 <nigelb> https://ci.centos.org/job/gluster_glusto
12:34:34 <amye> Sweeeeet
12:34:46 <rastar> Cool.
12:34:48 <nigelb> My hope is that we can leverage the learning from the test day into a few tests before the next cycle.
12:34:57 <nigelb> test day + check lists that Pranith was driving.
12:35:46 <nigelb> Anyone have questions on this?
12:35:48 <kshlm> nigelb, You'd require help with that?
12:36:24 <nigelb> I will need help with more people writing tests, but we're not there yet :)
12:36:51 * post-factum is late, but memory management is addressed by jdarcy, so i don't care atm :)
12:37:19 <kshlm> nigelb, Let us know when we're there.
12:37:41 <kshlm> I'm ready to write some of my release testing as glusto tests.
12:37:54 <nigelb> That's where I want to get us to.
12:38:10 <kshlm> This has been good news nigelb.
12:38:15 <jdarcy> At some point I should convert the test I've been using for multiplexing into a Glusto test.
12:38:27 <jdarcy> It's *really* good at stressing the system, if I do say so myself.
12:38:45 <nigelb> we run these over 4 nodes.
12:39:59 <kshlm> nigelb, Those are 4 nodes from duffy? (Duffy is the centos-ci provisioner that provides bare metal centos machines)
12:40:10 <nigelb> Yep.
12:40:14 <nigelb> I actually request 5 nodes
12:40:35 <nigelb> But the gluster setup is on 4 nodes and the first one is the "management node" where the cluster is managed.
12:41:03 <kshlm> nigelb, Couldn't the gluster slave be the management node?
12:41:37 <nigelb> I wanted to isolate them so gluster could max out CPU/RAM.
12:42:17 <kshlm> The management node shouldn't require a lot of resources.
12:42:31 <kshlm> We can discuss this later.
12:42:35 <nigelb> Yeah.
12:42:41 <rastar> that makes sense, sometimes you might have to reboot all Gluster nodes
12:42:45 <nigelb> At this point my priority is getting it to run correctly.
12:42:51 <nigelb> and then figure out optimizations.
12:42:58 <kshlm> nigelb, Cool.
12:43:05 <kshlm> I'll talk with you later.
12:43:10 <nigelb> sure
12:43:21 <kshlm> If there are no more questions, I'll move to the next topic.
12:44:17 <kshlm> Thanks again nigelb
12:44:31 <kshlm> #topic Recognizing contributors.
12:44:38 <kshlm> This is a carry over from last week.
12:44:57 <kshlm> A summary from the last time:
12:45:31 <kshlm> Stats are not the best for recognition, If it is required they should be thourough, fine-grained
12:45:39 <kshlm> coverage of stats on bitergia is not complete or enough
12:45:46 <kshlm> there are well identified additional ascpects to be considered for recognition
12:45:52 <kshlm> Everyone agrees we need to find a better method for recognizing
12:46:09 <amye> I'm still doing the 'top 5 contributors' but I'm pretty sure those metrics are wrong.
12:46:32 <amye> That being said, I will take suggestions for how we should be changing our dashboards to reflect contributions.
12:46:59 <amye> https://glusterfs.biterg.io/app/kibana#/dashboard/Overview is the new Kibana based dashboard
12:47:55 <obnox> If it is made clear that these metrics are not the only proof for being an important contributor, i.e. if it is made clear that these are only aspects (albeit very important ones) then nothing is wrong with them per se. imho
12:48:24 <obnox> Only there is danger in overestimating the metrics - both on the inside and on the outside perception
12:48:57 <nigelb> I think we're missing a piece in this discussion. You commit code or review code, and then what?
12:49:05 <amye> Jazz hands?
12:49:22 <amye> No, I kid, but at the moment, we don't have a community way to recognize that.
12:49:25 <nigelb> Most projects reward contributors with commit access/ownership.
12:49:40 <amye> Even glusterbot's karma doesn't work as expected
12:50:01 <kkeithley> are we overthinking this? Nothing is going to be perfect. A first order approximation is probably going to be accurate enough to capture the top five or ten or twenty contributors
12:50:11 <kkeithley> s/capture/identify/
12:50:54 <kkeithley> and the metrics can be refined over time if we think they somehow skewed
12:50:55 <obnox> captivate
12:51:52 <nigelb> Let me rephrase my question.
12:51:56 <nigelb> What is the problem we're trying to solve?
12:52:22 <kkeithley> my 2¢
12:52:27 <obnox> nigelb: not sure. I think people want recognition based on measureable performance data
12:52:40 <nigelb> "There are not enough community contributions" / "Contributions are not visible" / "Contributions are not rewarded" <-- There's plenty of similar but different problems we could solve.
12:52:41 <kshlm> Yup.
12:53:27 <obnox> nigelb: curiously it seems to be the engineers who are asking for such 'ranking' not employer's management .. ;-)
12:53:38 <kshlm> I feel it's more along the lines of people don't feel motivated to participate in the community because of lack of recognition.
12:53:51 <kkeithley> top contributors that aren't Red Hat employees? ;-)
12:54:19 <obnox> kshlm: lack of recognition from whom?
12:54:35 <kkeithley> from the Community is what I inferred
12:54:55 * obnox shouts "Developers! Developers! Developers!" in excitement...
12:54:56 <nigelb> Probably a stupid question: Why are we not rewarding measurable contribution with maintainership? (i.e. do good work and you get more work :P)
12:55:07 <kkeithley> lol
12:55:13 <ppai> he he
12:55:15 <kkeithley> we dont' want to scare them away
12:55:53 <obnox> nigelb: yeah, that might be a next step. first problem to solve: come up with meainingful measures
12:56:01 <obnox> nigelb: like commit counts is not enough
12:56:05 <amye> No, I think it's a valid question - how do we currently decide maintainership?
12:56:18 <obnox> nigelb: line counts, review counts, ML posts ...
12:56:37 <obnox> amye: sure it's a valid question, but not the one we are presently discussing
12:56:49 <obnox> it may get related though ;-)
12:56:51 <amye> Table it for the moment.
12:57:21 <nigelb> Yes, it's quality of work that matters more than quantity.
12:57:31 <nigelb> The problem with numbers is that quality is incredibly hard to measure.
12:57:45 <nigelb> Unless we want to do meta reviews.
12:57:49 <obnox> nigelb: indeed. exactly my criticism
12:58:07 <obnox> and who counts mentoring?
12:58:16 <obnox> and design discussions? ...
12:58:27 <obnox> super tough to measure
12:58:46 <amye> Ok, as we're nearly at time, take this to ML?
12:58:57 <ppai> many other communities vote to decide maintainers
12:59:02 <nigelb> We have 2 minutes to go, but I think we should look at other communities.
12:59:07 <nigelb> And see how they've solved this problem.
12:59:12 <obnox> if at all
12:59:31 <nigelb> I'll write to the ML.
12:59:38 <obnox> e.g. in samba we had discussed to have maintainers years back, and abandonded
12:59:41 <obnox> -d
12:59:51 <kshlm> nigelb, Thank you for that.
13:00:09 <kshlm> This discussion can keep going on forever.
13:00:18 <kshlm> I'll stop it here..
13:00:21 <obnox> +1
13:00:39 <obnox> regarding ML - i need someone to look into ML setup - why does heketi-devel not archive?
13:00:44 <obnox> pretty please
13:00:45 <kshlm> nigelb has volunteered to take it to the mailing lists, so let's keep it going on for ever there.
13:00:56 <amye> obnox: sounds like a bug for misc
13:01:05 <obnox> amye: right, i mailed him last week
13:01:05 <amye> plz to bugzilla
13:01:09 <kshlm> obnox, Open a issue. nigelb or misc will look into it.
13:01:18 <sankarshan> it might be good notion to have a different archive for that list as well
13:01:19 <amye> nah, mailing lists are misc
13:01:40 <obnox> ok thx
13:02:06 <kshlm> obnox, https://bugzilla.redhat.com/enter_bug.cgi?product=GlusterFS&component=project-infrastructure
13:02:17 <nigelb> ♥
13:02:25 <obnox> maybe mail was not enough, I was not aware that there is a bz comp for that - thx!
13:02:32 <nigelb> (I'm glad more people have it at their fingertips)
13:02:48 <kshlm> Let's end this meeting with updates and announcements.
13:02:54 <kshlm> #topic Updates & Announcements
13:03:07 <kshlm> Not a lot of updates this week, possibly owing to the holidays in India.
13:03:50 <kshlm> I had sent out a design + RFE for volgen for GD2. I'd like to get more feedback on it.
13:04:00 <kshlm> https://www.gluster.org/pipermail/gluster-devel/2016-October/051297.html
13:05:02 <kshlm> amye, Sent out out monthly newsletter.
13:05:14 <kshlm> https://www.gluster.org/pipermail/gluster-users/2016-November/028920.html
13:05:29 <amye> Yeah, sorry I missed that one, but the newsletter went out Oct 31.
13:06:02 <kshlm> amye, Do you have a plan for the bootstrapping challenge ready?
13:06:21 <amye> I followed up on ML, we're moving to 'after 3.9'
13:06:29 <kshlm> Okay.
13:06:37 <amye> Mostly because that release gives us a great opportunity
13:06:46 <amye> as well as holidays in India this week which I spaced on. :)
13:07:02 <kshlm> Okay.
13:07:06 <amye> So yes, there is a plan and it will be put in place with more warning. :)
13:07:42 <kshlm> The update had me slightly confused. I thought you wanted to do it before the 3.8 release.
13:07:49 <kshlm> s/.8/.9/
13:07:59 <amye> Oh, no, sorry.
13:08:08 <kshlm> I'll end with our regular announcements.
13:08:20 <kshlm> If you're attending any event/conference please add the event and yourselves to Gluster attendance of events: http://www.gluster.org/events (replaces https://public.pad.fsfe.org/p/gluster-events)
13:08:20 <kshlm> Put (even minor) interesting topics on https://public.pad.fsfe.org/p/gluster-weekly-news
13:08:20 <kshlm> Remember to add your updates to the next meetings agenda.
13:08:26 <kshlm> Thanks everyone.
13:08:35 <kshlm> Let's catch up again next week.
13:08:38 <amye> thank you!
13:08:43 <kshlm> #endmeeting