weekly_community_meeting_11may2016
LOGS
12:07:05 <kshlm> #startmeeting Weekly community meeting 11/May/2016
12:07:05 <zodbot> Meeting started Wed May 11 12:07:05 2016 UTC.  The chair is kshlm. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:07:05 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
12:07:05 <zodbot> The meeting name has been set to 'weekly_community_meeting_11/may/2016'
12:07:13 <kshlm> #topic Rollcall
12:07:24 * kshlm here
12:07:26 * kkeithley is here
12:07:30 <post-factum> \o/
12:07:31 <jdarcy> Good morning, esteemed colleagues.
12:07:31 * overclk is here
12:07:38 * aravindavk is here
12:07:41 * karthik___ is here
12:07:49 <kshlm> Good morning to you too jdarcy :)
12:07:56 * jiffin is here
12:08:01 * ndevos is here, but would rather be on the beach
12:08:06 <post-factum> afternoon here, but good morning
12:08:16 <post-factum> ndevos++
12:08:17 <glusterbot> post-factum: ndevos's karma is now 11
12:08:36 <kshlm> ndevos, You could be at both places ;)
12:08:42 <jdarcy> ndevos: Nah, the wifi's lousy and I'd get sand in my keyboard.
12:08:42 <kshlm> Let
12:08:48 <ndevos> kshlm: it is very tempting!
12:09:01 <kshlm> Let's start
12:09:02 <post-factum> anyone made underwater commit ever?
12:09:24 <kshlm> #topic Next weeks meeting host
12:09:33 <post-factum> rastar!
12:09:50 <kshlm> I won't be available to pick up hosting duties for the next 3 weeks
12:10:02 <jdarcy> I second post-factum's suggestion.
12:10:07 <kshlm> I vote rastar as well.
12:10:21 <kshlm> #agreed rastar is next weeks host
12:10:29 <kshlm> That's settled
12:10:36 <kshlm> #topic Last weeks AIs
12:10:45 <kshlm> #topic kshlm & csim to set up faux/pseudo user email for gerrit, bugzilla, github
12:10:48 <kshlm> Not done.
12:11:03 <kshlm> Maybe with nigelb on board now, we can get this done some time soon
12:11:42 <kshlm> I'll carry this forward until it's done
12:11:47 <kshlm> #action kshlm & csim to set up faux/pseudo user email for gerrit, bugzilla, github
12:11:51 <jdarcy> Welcome aboard, nigelb.
12:11:55 <nigelb> let me know later what you need and how I can help :)
12:12:12 <kshlm> nigelb sure
12:12:29 <kshlm> #topic jdarcy to provide a general Gluster-4.0 status update
12:12:34 <jdarcy> Done.  :)
12:12:46 <kshlm> jdarcy, link?
12:13:14 <jdarcy> #link http://www.gluster.org/pipermail/gluster-devel/2016-May/049367.html
12:13:27 <kshlm> Thank you
12:13:53 <post-factum> #link https://www.gluster.org/pipermail/gluster-devel/2016-May/049375.html
12:14:21 * msvbhat arrives bit later
12:14:29 <kshlm> Thanks post-factum
12:14:45 <kshlm> #topic hagarth to take forward discussion on release and support strategies (onto mailing lists or another IRC meeting)
12:14:46 * atinm feels that we are discussing old topics, mind to cross check?
12:15:09 <kshlm> atinm, I'm just using what was in the agenda.
12:15:17 <kshlm> I didn't prepare it.
12:15:20 <post-factum> #link https://www.gluster.org/pipermail/gluster-devel/2016-May/049402.html
12:15:45 <kshlm> post-factum, thanks again
12:16:05 <kshlm> We need more responses on that thread
12:16:37 <post-factum> also
12:16:40 <post-factum> #link https://www.gluster.org/pipermail/gluster-devel/2016-May/049455.html
12:17:23 <kshlm> I think the 2 are slightly different
12:17:36 <kshlm> But we need more responses on both anyway.
12:18:43 <kshlm> The first one talks about our supported releases and how we manage branches
12:18:53 <kshlm> The second one is about how we do perform releases.
12:19:16 <kshlm> Not much comments on either.
12:19:26 <kshlm> I think I'm the only one to have responded.
12:19:45 <kshlm> Can the other maintainers please add your opinions on the threads?
12:19:57 <kshlm> ndevos, jdarcy, kkeithley, overclk
12:19:59 <kshlm> ^
12:20:04 <kshlm> Who else is around?
12:20:16 <jdarcy> I'll try to think of something semi-intelligent to say.
12:20:59 <ndevos> I was hoping others would post some ideas and opinions before I do...
12:21:17 <post-factum> ndevos: everyone waits for you first
12:22:02 <kshlm> ndevos, what post-factum said :)
12:22:03 <kkeithley> besides stick to the release schedule (https://www.youtube.com/watch?v=NnP5iDKwuwk)
12:22:20 <ndevos> post-factum: yeah, and I REALLY hope some of our maintainers will speak up at one point, thats part of why they are maintainers too
12:23:06 <kshlm> kkeithley, one of the threads is about changes to the schedule
12:23:18 <kkeithley> I know
12:23:45 <kshlm> I thought that was question.
12:23:51 <kshlm> :D
12:23:58 <kkeithley> any schedule we have....  If we don't stick to it, is pointless. :-/
12:24:21 <kshlm> Well the other discussion should help up stick to schedules
12:24:37 <post-factum> ...or go with rolling releases...
12:25:21 <kshlm> post-factum, as in just 1 supported release all the time?
12:25:37 <post-factum> yeah, origin/master, and nothing else matters
12:26:27 <post-factum> picking lts each year or so
12:26:32 <ndevos> I can guarantee that we'll break things a *lot*
12:26:44 <post-factum> that's for sure
12:27:00 <ndevos> oh, "nothing else matters" and a LTS?
12:27:30 <kshlm> That's nearly what aravindavk suggested.
12:27:36 <ndevos> stable == LTS at the moment, we dont have a distinction there
12:27:38 <post-factum> kshlm: correct
12:29:02 <aravindavk> ndevos: non LTS stable != LTS
12:29:40 <aravindavk> frequent release of non LTS stable without updates
12:29:43 <post-factum> and the question is whether it is desirable to support multiple stable branches besides lts
12:29:45 <ndevos> aravindavk: sure, and at the moment I would not even call 3.7 a stable release :-/
12:30:08 <post-factum> ndevos: 3.7 looks like natural rolling release
12:30:17 <post-factum> ndevos: or at least looked before 3.8 branching
12:30:52 <post-factum> ndevos: and noone make complaints
12:30:56 <ndevos> post-factum: yeah, that was done because the 3.8 release was not done, and there was no way to get features out to users
12:31:13 <post-factum> ndevos: no guarantee it won't repeat
12:31:55 <kshlm> Alright. Time to moderate.
12:31:57 <ndevos> post-factum: no, but we can at least try to stick to the schedule that we made, and treat release-* branches as stable
12:31:59 <post-factum> :)
12:32:05 <kshlm> Can we continue this discussion on the mailing lists?
12:32:12 <post-factum> kshlm: you are right
12:32:23 <ndevos> sure!
12:32:25 <kshlm> It's nice to have everyone opinions in one place.
12:32:30 <kshlm> Thanks guys.
12:32:43 <kshlm> #topic amye to check on some blog posts being distorted on blog.gluster.org, josferna's post in particular
12:33:07 <aravindavk> kshlm: link to post? I can have a look at it
12:33:27 <kshlm> aravindavk, josferna's post?
12:33:50 <aravindavk> kshlm: whichever distorted
12:33:53 <ndevos> http://planet.gluster.org/ -> joelearnsopensource
12:34:12 <kshlm> Was it planet or blog.gluster.org?
12:34:17 <post-factum> #link http://blog.gluster.org/author/josephaug26/
12:34:19 <post-factum> this?
12:34:27 <kshlm> We're syncing both.
12:34:39 <ndevos> I thought blog.gluster.org is deprecated and is scheduled for deletion?
12:35:09 <aravindavk> kshlm: ok, I will check
12:35:25 <kshlm> The blog.gluster.org post is http://blog.gluster.org/2016/03/nos-conf-2016/
12:35:45 <aravindavk> kshlm: 404 error
12:35:49 <kshlm> ndevos, I thought so too earlier. But I think amye has other plans.
12:36:09 <ndevos> kshlm: oh, ok
12:36:17 <kshlm> aravindavk, I think that was the supposed fix.
12:37:02 <kshlm> I should probably just bring this up with amye, she's gonna be in Bangalore for next few days.
12:37:16 <kshlm> And close this.
12:37:34 <kshlm> #action amye to check on some blog posts being distorted on blog.gluster.org, josferna's post in particular
12:37:43 <kshlm> Carrying it forward anyways
12:38:02 <kshlm> #topic pranithk1     sends out a summary of release requirements, with some ideas
12:38:08 <kshlm> I don't think I've seen this.
12:38:28 <kshlm> Was this releated to the release process enhancement discussions?
12:38:46 <kshlm> The one that hagarth started.
12:39:25 <ndevos> yeah, I think so
12:39:33 <ndevos> and to aravindavk's mail
12:40:08 <ndevos> we need to know what we want to offer to our users, and see what schedule/planning we can do with that
12:40:53 <ndevos> developers want new features out and tested as early as possible
12:41:07 <kshlm> aravindavk's thread is about this.
12:41:12 <ndevos> users want to have their data in a safe place that does not break
12:41:33 <kshlm> hagarth's was about how we execute releases.
12:41:49 <kshlm> Improvements to the release-process done by the release-maintainer/managers.
12:41:51 <jdarcy> If we only had a modular architecture that would let us un-bundle and release new features separately.
12:42:09 <kshlm> jdarcy, :D
12:42:11 <ndevos> I think aravindavk wrote about a solution, but without stating the requirements/expectations of different groups (but its been a while since I read the mail)
12:42:58 <kshlm> ndevos, Please bring it up on the thread!
12:43:16 <kshlm> I'll carry pranith's AI forward
12:43:16 <ndevos> kshlm: I think that was praniths AI :)
12:44:08 <kshlm> #action pranithk1     sends out a summary of release requirements, with some ideas
12:44:34 <kshlm> #topic hagarth will start a discussion on his release-management strategy
12:44:49 <kshlm> This was about improving the release process,
12:45:10 <kshlm> post-factum, already linked to it before
12:45:32 <kshlm> #link https://www.gluster.org/pipermail/gluster-devel/2016-May/049455.html
12:45:40 <kshlm> We need more comments here as well.
12:45:59 <kshlm> I'm all for the proposed change, and have already given my +1
12:46:16 * msvbhat brb
12:46:43 <kshlm> #topic kshlm to check with reported of 3.6 leaks on backport need
12:46:44 <ndevos> it might be overkill, seeing how little many of the maintainers actually reply to requests...
12:47:08 * ndevos still needs to put that in an email too
12:47:23 <kshlm> but it could be one way to get more maintainers participating
12:47:49 <ndevos> maybe, but I'm not too confident it'll work
12:48:23 <ndevos> move on?
12:48:28 <kshlm> Yeah
12:48:34 <kshlm> #topic kshlm to check with reported of 3.6 leaks on backport need
12:48:39 <kshlm> Not done.
12:48:46 <kshlm> This was an AI from 2 weeks back.
12:49:00 <kshlm> I'll dig up the thread again, and check if this is still required.
12:49:06 <kshlm> #action kshlm to check with reported of 3.6 leaks on backport need
12:49:20 <kshlm> So we're done with AIs.
12:49:53 <kshlm> AFAIK we don't have much to update about 3.7.12 and 3.6
12:49:58 <jdarcy> That leaves eleven minutes for five releases.
12:50:14 <ndevos> nothing on 3.5 either
12:50:14 <kshlm> 3.7.12 is pending an agreement about the release process.
12:50:24 <kshlm> Isn't 3.5 done?
12:50:36 <ndevos> not until 3.8 is GA
12:50:41 <kshlm> Ah okay.
12:50:47 <kshlm> On the topic of 3.8
12:50:51 <kshlm> #topic GlusterFS-3.8
12:50:57 <kshlm> ndevos, Any updates?
12:51:16 <ndevos> looks pretty ok
12:51:31 <ndevos> https://www.gluster.org/community/roadmap/3.8/ has been updated with the current features
12:51:57 <ndevos> https://www.gluster.org/community/roadmap/3.9/ now has the removed features, that is a placeholde page until we know a 3.9/4.0 schedule
12:52:08 <kshlm> Unfortunately pranith and I couldn't get the sub-dir mount feature in.
12:52:14 <ndevos> and we need to get notes in https://public.pad.fsfe.org/p/glusterfs-3.8-release-notes
12:52:32 <kshlm> I atleast hopt that I can get the encrypted transport fixes I'm working on in.
12:52:56 <ndevos> many features did not make it in, thats fine, as long as we do regular releases in time
12:53:13 <aravindavk> ndevos: gluster eventing is not merged in 3.8. We need to remove from roadmap page?
12:53:33 <kshlm> aravindavk, pull-request?
12:53:36 <ndevos> aravindavk: yeah, I wasnt sure about that one, and a few others
12:53:58 <aravindavk> kshlm: ndevos ok, will send pull request
12:54:08 <ndevos> aravindavk: great, thanks!
12:54:11 <kshlm> ndevos, It's been noted that an RC1 was planned late last week
12:54:28 <ndevos> aravindavk: move it to the 3.9 page, and add links to the bugs too in case those are missing
12:54:35 <kshlm> Has there been a change to plans?
12:54:49 <aravindavk> ndevos: sure
12:54:51 <ndevos> kshlm: that is correct, but I did not have time for it, trying later this week again
12:55:07 <kshlm> ndevos, Good to know.
12:55:30 <kshlm> So that should be it regarding 3.8 for now
12:55:34 <ndevos> kshlm: I also would like to see some release-notes before doing the RC1 release and ask users to test it
12:56:25 <kshlm> So the feature-developers need to provide this.
12:56:39 <ndevos> yes, and jiffin is getting an email out to them
12:56:48 <kshlm> Will you be asking for this, or do they alreday know
12:57:00 <kshlm> Okay.
12:57:05 <kshlm> Good to know again.
12:57:18 <kshlm> I'll move onto 4.0 then.
12:57:25 <kshlm> #topic GlusterFS-4.0
12:57:35 <kshlm> Nothing's been done for GD2.
12:57:46 <post-factum> #link https://www.gluster.org/pipermail/gluster-devel/2016-May/049436.html
12:58:01 <kshlm> Thanks post-factum.
12:58:05 <jdarcy> Mostly covered in the email linked earlier.  Also, both Shyam and I have been able to make some progress on JBR and DHT2 respectively.
12:58:09 <kshlm> How do you find these links so fast.
12:58:17 <post-factum> ctrl+f-fu
12:58:40 <kshlm> I've been more interested in fixing the encrypted connections in glusterfs, and have been doing that for the past 2 weeks.
12:59:13 <kshlm> I hoped to do some refactoring in GD2, and I hope to get that done by the end of next week.
12:59:21 <kshlm> Before I go on a 2 week vacation.
13:00:14 <kshlm> We've been discussing about having show-n-tell sessions for 4.0 features,
13:00:18 <jdarcy> And (if you have a good vacation) forgetting absolutely everything about what you were in the middle of.
13:00:35 <kshlm> but that's already being discussed on the mailing lists.
13:00:55 <kshlm> jdarcy, I hope to leave nothing half-done.
13:01:08 <ndevos> kshlm: I upgraded my 3.5/3.6 home storage to 3.8, and had to disable ipv6 to get it to work - I hope your fix addresses that too :)
13:01:26 <kshlm> I need to finish 2 things, encrypted connections and refactoring GD2.
13:01:30 <post-factum> ndevos: was it downtimeless?
13:01:34 <kshlm> And I think I can get done.
13:01:42 <jdarcy> There seems to be an emerging consensus on doing the 4.0 show-n-tell at 15:00 UTC, not sure which day.
13:01:43 <kshlm> ndevos, Yes. It should be fixed.
13:02:06 <kshlm> post-factum, I don't think so.
13:02:34 <kshlm> IPv6 has problems, so ndevos would have had some downtime
13:02:48 <kshlm> But the GA release should'nt have any problems.
13:02:56 <ndevos> post-factum: no, mounting failed everywhere, until I disabled ipv6 in the dns server
13:02:59 <post-factum> i thought ipv6 was invented to silve problems and not to introduce them
13:03:06 <post-factum> *solve
13:03:20 <kshlm> :)
13:03:30 <kshlm> But glusterfs still doesn't understand it very well.
13:03:38 <jdarcy> What's the state of IPv6 regression tests?  Are there any?
13:03:42 <kshlm> It needs to learn some more.
13:04:23 <kshlm> jdarcy, I'm trying to get the current regression suite complete successfully in a IPv6 environment.
13:04:37 <kshlm> That should be good enough for a start.
13:04:37 <ndevos> oh, and also  something about afr/init asking about op-version >= 30707... and it was set to 30800 already - set/reset some volume options to regenerate the .vol files :-/
13:05:15 <kshlm> I'd like it if IPv6 were enabled on the jenkins slaves.
13:05:24 <kshlm> That would have caught a lot of the problems much earlier.
13:06:25 <jdarcy> We're over time.  How are we doing on the agenda?
13:06:32 <kshlm> ndevos, setting cluster.op-version doesn't regenerate volfiles.
13:06:48 <kshlm> jdarcy, Nothing else left actually
13:06:50 <ndevos> kshlm: no, but that was what the afr error message suggested
13:07:02 <jdarcy> So, open floor?
13:07:08 <kshlm> Yeah
13:07:13 <kshlm> #topic Open floor
13:07:20 <jdarcy> I'd like to propose that we add Ganesha and Samba reports to the agenda.
13:07:34 <kkeithley> +1
13:07:40 <kshlm> +1
13:07:55 <ndevos> reports, like the ones we have for the versions?
13:08:00 <jdarcy> Exactly.
13:08:09 <ndevos> sounds good to me
13:08:19 <jdarcy> Those teams are as large and active as any working on core releases.
13:08:32 <jdarcy> With as much impact to the codebase.
13:09:19 <jdarcy> Also, I delivered a talk on Gluster at a Google Compute Engine meetup in Cambridge last week.  Went reasonably well.
13:09:19 <kshlm> They've driven a lot of 3.8 features, haven't they?
13:09:24 <jdarcy> kshlm: Indeed.
13:09:26 <ndevos> maybe also an item 'Other sister projects'? when qemu or some other project needed something from us Gluster community?
13:09:41 <amye> jdarcy, that wasn't.. recorded, was it?
13:09:43 <jdarcy> ndevos: Works for me.
13:09:50 <jdarcy> amye: No, I don't think so.
13:10:02 <kshlm> I'll add the projects to next weeks agenda.
13:10:06 * amye inserts frowny-face here. All good.
13:10:22 <jdarcy> amye: It's one of the bigger audiences I've had, actually.
13:10:43 * jdarcy had to blast through ~45 minutes' worth of slides in under 20.  Fun.
13:11:00 <ndevos> hah, not bad :)
13:11:59 <kshlm> Do we need to inform Ganesha and Samba devs of the plan?
13:12:14 <kkeithley> I think the Ganesha team lead knows
13:12:18 <jdarcy> Heh.
13:12:24 <jdarcy> Anybody here from the Samba side?
13:12:29 <kshlm> So just the samba devs then?
13:12:29 <kkeithley> lol
13:12:36 <post-factum> ira?
13:12:42 <kkeithley> madam
13:12:43 <ndevos> obnox is listening in here too
13:12:51 <kkeithley> aka obnox
13:13:20 <jdarcy> I just figured out why Michael doesn't use first name plus last initial for his IRC handle.
13:13:23 * ndevos always feels awkward when emails from madam@... come in
13:13:35 * jdarcy is a bit slow sometimes.
13:13:35 <kkeithley> I'm Adam
13:13:38 <post-factum> :D
13:13:51 <amye> kkeithley, you're adam now too?
13:14:01 <post-factum> we are all adams here, amye
13:14:09 <jdarcy> How Biblical.
13:14:44 <kkeithley> Doc, note. I dissent. A fast never prevents a fatness. I diet on Cod
13:14:53 <kshlm> So who's gonna notify either ira or obnox ?
13:15:07 <jdarcy> I'll volunteer for that.
13:15:08 <post-factum> glusterbot on #gluster? :)
13:15:30 <kshlm> Thanks jdarcy
13:15:35 <amye> jdarcy, are you thinking every week, every other week for agenda items, or?
13:15:37 <ndevos> I'm pretty sure obnox will read this later today :)
13:15:48 <jdarcy> amye: I was thinking every week, just like the releases.
13:15:53 <amye> ack
13:15:53 <kshlm> post-factum, that would work if they actually logged out.
13:16:22 <post-factum> no logged in ira there now
13:17:05 <kshlm> post-factum, next time we'll try glusterbot
13:17:13 <kshlm> :)
13:17:23 <kshlm> I think this is the end of today's meeting.
13:17:37 <kshlm> Thank you all who attended
13:17:41 <kshlm> #endmeeting