gluster-meeting
MINUTES
12:11:45 <atinm> #startmeeting
12:11:45 <zodbot> Meeting started Wed Nov  4 12:11:45 2015 UTC.  The chair is atinm. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:11:45 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
12:11:51 <msvbhat> The time is tied to UTC which doesn't have any effect due to DST observation in some parts
12:11:58 <atinm> Its time to have  community meeting for this week
12:12:19 <atinm> #info Agenda is here - https://public.pad.fsfe.org/p/gluster-community-meetings
12:12:24 <rjoseph> Is there any communication on change of time for this meeting?
12:12:29 <jdarcy> msvbhat: Even as a victim of that particular insanity (EDT/EST), I think that's the right choice for an international group.
12:12:39 <atinm> #topic roll call
12:12:45 * partner here
12:12:47 * jdarcy is here (obviously)
12:12:55 * poornimag here
12:12:58 * jiffin is here
12:12:58 * anoopcs is here
12:13:00 * msvbhat is here
12:13:00 * rastar is here
12:13:01 * obnox here
12:13:08 * ndevos is here
12:13:10 * rjoseph is here
12:13:20 * asengupt is here
12:13:28 <atinm> Looks like we have good number of folks around, lets get started
12:13:31 <msvbhat> jdarcy: :) Right
12:13:50 <atinm> Lets scan through the last week action items
12:13:53 * overclk is there
12:13:58 * kshlm is here
12:14:01 <atinm> kshlm to check back with misc on the new jenkins slaves
12:14:09 <atinm> kshlm, could you update on this?
12:14:23 <kshlm> I've not done anything on this.
12:14:43 <ndevos> atinm: if there are any action items for me, consider them as "no progress" and re-check for next week
12:14:56 <atinm> ndevos, sure, thanks for the one liner :)
12:15:15 <atinm> kshlm, shall I consider it for next week's AI?
12:15:20 <kshlm> I'll try to get this started this week.
12:15:43 <atinm> #action kshlm to check back with misc on the new jenkins slaves
12:15:46 <atinm> Moving on
12:15:55 <atinm> krishnan_p and atinmu will remind developers to not work in personal repositories, but request one for github.com/gluster
12:16:33 <atinm> Well, though I've not sent any mails on that but I believe all of 4.0 initiatives are now hosted in github.com/gluster so I consider this action item as closed, any objection?
12:16:44 <jdarcy> None here.
12:16:59 <atinm> jdarcy, thanks
12:17:10 <atinm> Over to next item
12:17:15 <ndevos> if all the repos have been moved, no futher discussion needed :)
12:17:22 <atinm> ndevos send out a reminder to the maintainers about more actively enforcing backports of bugfixes
12:17:26 <atinm> #action ndevos send out a reminder to the maintainers about more actively enforcing backports of bugfixes
12:17:52 <atinm> skoduri, poornimag and obnox_ to post SDC trip report on gluster-devel
12:17:58 <obnox> done (on my side)
12:18:12 <atinm> obnox, thanks
12:18:26 <atinm> Next one is raghu to call for volunteers and help from maintainers for doing backports listed by rwareing to 3.6.7
12:18:37 <atinm> do we have raghu today?
12:19:25 <atinm> From glusterd side I see couple of patches been backported by Anand
12:19:46 <raghu> atinm: I have to yet to call for the volunteers. I will be doing it this week
12:19:56 <atinm> #action raghu to call for volunteers and help from maintainers for doing backports listed by rwareing to 3.6.7
12:20:16 <atinm> Moving on
12:20:22 <atinm> hagarth to post a tracking page on gluster.org for 3.8 by next week's meeting
12:20:46 <atinm> hagarth, I believe this is almost done
12:21:38 <atinm> I saw some conversation between tigert & hagarth on this, there are some issues in it
12:21:50 <atinm> seems like hagarth is not around, moving it for next week
12:21:56 <atinm> #action hagarth to post a tracking page on gluster.org for 3.8 by next week's meeting
12:22:08 <atinm> Next one is on you rafi
12:22:09 <atinm> rafi to setup a doodle poll for bug triage meeting
12:22:27 <atinm> rafi, any updates?
12:22:41 * atinm doesn't recollect seeing any mail on this
12:23:06 * ndevos was wondering about that too yesterday, but time didnt change in the etherpad/agenda yet
12:23:36 <atinm> I am moving this for next week then
12:23:44 <atinm> #action rafi to setup a doodle poll for bug triage meeting
12:24:01 <atinm> Next up is rastar and msvbhat to publish a test exit criterion for major/minor releases on gluster.org
12:24:12 <rastar> this is not done yet :(
12:24:18 <msvbhat> Can we move this to next week
12:24:24 <atinm> msvbhat, sure
12:24:25 <msvbhat> We'll discuss this week
12:24:28 <msvbhat> Thanks
12:24:32 <atinm> #action rastar and msvbhat to publish a test exit criterion for major/minor releases on gluster.org
12:24:52 <atinm> hagarth to finish review http://review.gluster.org/#/c/12210/
12:24:58 <atinm> this is already done and merged
12:25:10 <atinm> next one is on me, atinm to send a monthly update for 4.0 initiative
12:25:43 <atinm> GlusterD 2.0 & DHT has sent their updates, we are pending with NSR
12:26:08 <atinm> jdarcy, could you send a progress update on NSR for last two months in devel?
12:26:28 <jdarcy> Sure.
12:26:33 <atinm> jdarcy, thanks
12:26:50 <atinm> #action jdarcy to send monthly update for NSR
12:27:03 <atinm> I'll also ask Samikshan to share the status on Eventing as well
12:27:24 <atinm> #action samikshan to send status on Gluster Eventing
12:27:27 <atinm> Moving on
12:27:39 <atinm> rastar will open a BZ for 3.7.5 upgrade issue with glusterd commands
12:27:49 <atinm> I see the etherpad is updated as done, so moving on
12:27:53 <rastar> that is done..
12:27:59 <atinm> atinm will also put up the GlusterD 2.0 design document for review in a week or two
12:28:29 <atinm> We are working on this and I expect a bit delay on this, we will try to push the first cut of it by next week
12:28:53 <atinm> #action atinm to put up the GlusterD 2.0 design doc by end of next week
12:29:16 <atinm> next is on overclk
12:29:19 <atinm> overclk to review http://review.gluster.org/#/c/12321/
12:29:29 <overclk> atinm: so basically this was a noop for me
12:30:10 <overclk> atinm: Other reviewers are already doing a fantastic job and little left for me
12:30:23 <jdarcy> Shyam and I have gone around a bunch on this.  I think the two of us have converged, but it would be nice for others to weigh in as well.
12:30:49 <atinm> overclk, so I consider this done from your end but need more eyes
12:31:09 <atinm> #action more eyes needed on http://review.gluster.org/#/c/12321/
12:31:16 <ndevos> I'll have a look at that too, I've given input on it earlier and need to check it again
12:31:29 <overclk> jdarcy: yeh, most the crucial points have been covered well. I'll probably go through it again.
12:31:40 <atinm> ndevos, thanks
12:31:58 <atinm> So with that we come to the last AI which I don't have any idea about :)
12:32:00 <atinm> set up faux/pseudo user email for gerrit, bugzilla, github
12:32:07 <atinm> onus is on whom ?
12:32:31 <atinm> ndevos, any idea?
12:32:53 <ndevos> kshlm and csim were discussing that, I think
12:33:13 <rastar> this came up because our gerrit -> github sync is broken
12:33:29 <ndevos> there was an email on the infra mailinglist, but I did not follow it really
12:33:36 <rastar> it was using avati's/hagarth's account
12:33:45 <rastar> and some authentication error happened
12:33:53 <atinm> rastar, so no concrete result out of it
12:34:00 <ndevos> #link http://thread.gmane.org/gmane.comp.file-systems.gluster.infra/531/focus=532
12:34:06 <atinm> rastar, shall I put it on your name for the next week?
12:34:08 <ndevos> no results so far
12:34:26 <rastar> kshlm is looking into it
12:34:28 <rastar> with misc
12:34:50 <atinm> rastar, ok
12:35:16 <atinm> #action kshlm & csim to set up faux/pseudo user email for gerrit, bugzilla, github
12:35:21 <atinm> Moving on
12:35:28 <atinm> #topic GlusterFS 3.7
12:35:38 <atinm> rastar, time to speak :)
12:35:45 <rastar> 3.7.6 is not out yet
12:35:53 <rastar> waiting on few blocker patches
12:36:02 <rastar> without them release does not make sense
12:36:06 <atinm> #info 3.7.6 is delayed because of some blocker patches
12:36:20 <rastar> aiming for this weekend, tentative
12:37:01 <atinm> #info 3.7.6 is rescheduled for end of this week
12:37:10 <ndevos> I was thinking about putting 3.7.5 in the CentOS Storage SIG as a stable release
12:37:27 <ndevos> should I rather wait until 3.7.6 is out so that there are no update issues in the future?
12:37:33 <atinm> ndevos, I've a question mark on 3.7.5 stability
12:37:36 <rastar> ndevos: no 3.7.5 has a bug
12:37:45 <rastar> atinm can explain more
12:37:49 <ndevos> okay, so 3.7.6 it is :)
12:38:03 * atinm nods his head
12:38:16 <ndevos> I dont really need to know the details, at least not now/here
12:38:26 <atinm> ndevos, thanks for saving the time :)
12:38:38 <atinm> rastar, anything else or I will switch to 3.6
12:39:27 <atinm> I take it as 'No' so over to next topic
12:39:35 <atinm> #topic GlusterFS 3.6
12:39:46 <atinm> raghu, anything we need to discuss?
12:39:58 <raghu> As said, I will be sending a mail to send backports
12:40:10 <raghu> I am planning to make the release on 20th
12:40:25 <raghu> There are only 2 patches for 3.6 which arrived few hours before.
12:40:32 <atinm> #info 3.6.7 to be out by 20th this month
12:40:37 <raghu> atinm: I need your review for those 2 patches
12:41:16 <atinm> raghu, I am actually concerned about a BSD smoke failure specifically in 3.6 branch which has been there for almost more than 3 months now
12:41:45 <atinm> raghu, I remember kshlm looking into it earlier but we couldn't figure it out
12:41:58 <atinm> raghu, I think I'll drop a mail to Manu for his help on this
12:42:18 <raghu> atinm: It was something that was discussed with Emmanuel I think. Many of the netbsd patches were not backported to 3.6 and it was left behind. So I think it was decided to ignore netbsd errors
12:42:21 <atinm> #action atinm to send a mail to Manu asking his help on 3.6 BSD smoke failures
12:42:45 <atinm> raghu, I never saw any mail on this, was it public, or discussed over IRC?
12:43:03 <raghu> atinm: I think over irc
12:43:21 <atinm> raghu, Could you drop a mail to devel about this?
12:43:33 <raghu> atinm: sure.
12:43:40 <atinm> raghu, thanks
12:43:49 <raghu> atinm: if we have to fix netbsd for 3.6 I think it takes sometime
12:43:50 <jdarcy> I was going to bring up NetBSD/FreeBSD build errors in open floor anyway.
12:44:06 <raghu> atinm: are you planning to have them fixed by 3.6.7?
12:44:06 <atinm> #action raghu to send a note in ML to ignore BSD failures
12:44:34 <atinm> raghu, I just want a confirmation before I give +1 to those patches :)
12:45:07 <atinm> raghu, I don't have bandwidth to fix them
12:45:33 <atinm> looks like we are running out of time
12:45:48 <raghu> atinm: I will send out a mail to devel regarding netbsd failures in 3.6 and we can discuss there and probably come up with a solution
12:45:57 <atinm> ndevos, can you host the last part of it since I have to jump over the next meeting?
12:46:17 <atinm> #chair ndevos
12:46:17 <zodbot> Current chairs: atinm ndevos
12:46:22 <ndevos> atinm: sure
12:46:26 <atinm> ndevos, thanks
12:46:48 <ndevos> raghu: was that all on 3.6?
12:47:01 <raghu> ndevos: yes
12:47:30 <ndevos> ok
12:47:39 <ndevos> #topic GlusterFS 3.5
12:47:43 <ndevos> ah, thats mine!
12:48:02 <ndevos> there are no updates needed for now, nobody filed urgent bugs that I am aware about
12:48:08 <ndevos> and nobody sent patches either
12:48:26 <ndevos> any questions on 3.5?
12:48:42 <ndevos> #topic GlusterFS 3.8
12:48:47 <ndevos> #link https://public.pad.fsfe.org/p/gluster-3.8-features
12:48:59 <ndevos> is there someone that would like to talk about 3.8?
12:49:04 <hagarth> ndevos: the feature page on 3.8 should go live soon
12:49:16 <hagarth> ndevos: rather the project tracking page
12:49:23 <ndevos> hagarth: nice!
12:49:39 <poornimag> The upstream Admin guide is in pretty bad shape, and can be made better, was wondering if Doc but for Admin guide be made blocker for 3.8?
12:49:47 <ndevos> oh, and there are currently 16 (!) features listed on that page, I wonder if we can get them all done
12:49:54 <poornimag> s/but/bug
12:49:59 <hagarth> #link https://github.com/gluster/glusterweb/blob/master/source/community/roadmap/3.8.md
12:50:07 <hagarth> ndevos: yes, some of them are stretch goals
12:50:25 <hagarth> at the same time, we do not have features like IPv6 and the ones contributed by facebook listed
12:50:52 <hagarth> poornimag: you mean the one on readthedocs.org?
12:51:02 <poornimag> hagarth, yes
12:51:19 <ndevos> hagarth: where will the roadmap be linked on? got a planned url?
12:51:23 <hagarth> poornimag: I am all for it.
12:51:30 <hagarth> www.gluster.org/community/roadmap/
12:51:39 <hagarth> the 3.8 link is broken atm :-/
12:51:49 * kkeithley wonders if daylight saving change broke my calendar
12:52:08 <hagarth> poornimag: we probably should do a divide and conquer approach for fixing admin-guide
12:52:18 <ndevos> kkeithley: yes, it does that 2x a year, and not only yours
12:52:24 <poornimag> Ok, may be i will raise a BZ for it(probably for every component) and mail on gluster-devel,
12:52:31 <obnox> kkeithley: it probably changed it. ... the meeting stuck at UTC time
12:52:32 <atinm> kkeithley, we are sticking with 12:00 UTC
12:52:33 <kkeithley> sigh
12:52:36 <hagarth> poornimag: sounds cool, thanks!
12:52:39 <poornimag> hagarth, yes,
12:53:04 <ndevos> hagarth: got more in 3.8?
12:53:34 * ndevos counts that as a "no"
12:53:35 <ndevos> #topic Gluster 4.0
12:53:36 <hagarth> ndevos: request all feature owners to update links to feature pages in the project tracking page
12:53:48 <ndevos> sorry, too late :D
12:53:50 <hagarth> ndevos: that's about it
12:53:59 <hagarth> ndevos: never mind, will shoot an email soon :D
12:54:12 <ndevos> jdarcy, atinm: any additions to 4.0?
12:54:31 <jdarcy> Um, "work continues"?
12:54:50 <ndevos> ok, well, if that is all, we can move to the Open Floor
12:54:52 <atinm> Yup, that's the shortest and crisp answer what we can come up with
12:54:59 <ndevos> good :)
12:55:04 <ndevos> #topic Open Floor
12:55:15 <jdarcy> FreeBSD/NetBSD regression failures.
12:55:16 <ndevos> #info Weekly reminder to announce Gluster attendance of events: https://public.pad.fsfe.org/p/gluster-events
12:55:27 <ndevos> #info REMINDER to put (even minor) interesting topics on https://public.pad.fsfe.org/p/gluster-weekly-news
12:55:37 <ndevos> jdarcy: yeah, what about it?
12:55:56 <jdarcy> There has been some discussion of whether (or when) these should block merges, but we still need to reach a consensus.
12:56:19 * ndevos is looking into mount-nfs-auth.t failures since yesterday, and has identified at least 2 different racy failures
12:56:32 <jdarcy> On the one hand they're becoming a bottleneck for all development, on the other hand I don't think anyone wants to drop them as supported platforms.
12:56:36 <ndevos> of course they should block merges
12:57:09 <jdarcy> ndevos: So it's OK if all development grinds to a halt because of that?
12:57:14 <ndevos> we need to make our tests more stable, so that they dont have any races anymore
12:57:21 <ndevos> jdarcy: I would say so, yes
12:57:45 <ndevos> I'm in a "drop all, fix the test" mode at the moment
12:58:05 <jdarcy> Then you need to say so in the ML discussion.
12:58:06 <atinm> ndevos, the problem here is the pace at which we fix linux specific issues is not same for BSD platform, we hardly have experts who understands the BSD part of it
12:58:23 <atinm> ndevos, question is how long can you wait ?
12:58:39 <ndevos> atinm: BSD is still part of our community, dvelopers need to know at least the basics
12:59:01 <jdarcy> The general instability of tests is IMO a different issue, which has been biting us continuously for a long time.
12:59:24 <ndevos> atinm: I see very little questions from developers that look into BSD problems, learning without asking questions take really long
12:59:43 <jdarcy> ndevos: How do you propose to make that happen?  Should there be a "BSD for dummies" doc somewhere?  If so, who should create it?
12:59:50 <ndevos> askng a question is relatively simple, and Manu is excellent in answering them
13:00:19 <ndevos> jdarcy: yeah, I had issues with running Gluster on NetBSD too, it is not *that* straight forward
13:00:22 <kkeithley> there are several people who an answer questions about *BSD. E.g. I'm one
13:00:38 <ndevos> documenting that would be good, well, documenting in places we usually check
13:01:38 <ndevos> freebsd is even easier to use than netbsd, imho, so whoever is stuck, should definitely ask questions about it on the -devel list
13:02:12 <ndevos> jdarcy: once we get developers to ask questions, we have an idea what needs to get documented
13:02:13 <jdarcy> Should we repeat that guidance *on the list*?
13:02:40 <ndevos> possibly, but I also hope I gave an example about it yesterday
13:03:34 <jdarcy> My impression is that when most developers get one of those failures they just set the patch aside, where it gets further out of date until it needs a non-trivial rebase.  Increases total work.
13:04:37 <ndevos> well, thats an issue with the developers workflow then... not sure how to address that
13:05:28 <kkeithley> So, with NetBSD, we get help from Emmanuel. For FreeBSD we get no help from the community. As much as I'm a fan of FreeBSD, if we're not getting support from the community, maybe we should drop it as a merge blocker.
13:05:53 <asengupt> the other issue is when we expect developers to understand their way around glusterd in BSD and fix patches, we kind of limit outside contribution to certain extent
13:06:06 <jdarcy> kkeithley: That seems fair.
13:06:07 <asengupt> i saw that with a couple of facebook patches
13:06:25 <ndevos> jdarcy: also, I am much in favour of having distinct Verified+1 for NetBSD and CentOS
13:06:52 <ndevos> kkeithley: FreeBSD only runs smoke tests, I think that is reasonable for a blocker, it is mainly compile only
13:07:00 <jdarcy> ndevos: OK, let's try and figure out how to make that happen.  I think it's key to making this problem tractable.
13:07:29 <kkeithley> smoke is a bit more than just compile.
13:07:36 <ndevos> jdarcy: kshlm is one of the Gerrit admins and I think he should be able to add new labels - there have been email on the infra list before about it
13:07:51 <ndevos> "a bit", but only very basic functionality
13:08:21 <kshlm> ndevos, jdarcy, I can help with that.
13:08:28 <kkeithley> And why Verified+1 for CentOS? I could argue that if we're being egalitarian we ought to have distinct Ubuntu, Debian, CentOS, Fedora, etc.
13:08:40 <ndevos> jdarcy: I'll find the email to the infra list, and reply to the other email thread
13:09:18 <ndevos> kkeithley: well, I would not say Linux, because we only test on CentOS, maybe we want to test on other distributions at one point?
13:09:18 <jdarcy> Sounds good.
13:09:51 <ndevos> I'm pretty sure we broke Gluster on Debian several times without noticing it...
13:10:29 <ndevos> right, the agenda contains one more topic
13:10:35 <ndevos> Should we promote/watch http://serverfault.com/questions/tagged/glusterfs
13:10:54 <ndevos> there seem to be some questions on serverfault about Gluster, and there even is a tag for them
13:11:40 <ndevos> amye: this is mainly a topic for you, I think, not sure if you're lurking?
13:12:10 <ndevos> anyway, I wanted to make you aware that it exists, and some users post their questions there
13:12:17 <amye> Oh, excellent
13:12:28 <amye> I will add that to my radar, I was not aware that it existed.
13:12:55 <kkeithley> maybe when we can keep up with the questions on #gluster and gluster-users email. What's the status of our "office hours"? Are we even managing to do those? Do them well?
13:13:11 <ndevos> amye: we just need to come up with a plan where at least some people get notifications and can poke others to answer those questions
13:13:52 <amye> ndevos: A plan without volunteers is going to make people sad, so let's start with kkeithley's suggestion first
13:14:06 <ndevos> oh, that gives me an idea, maybe JoeJulian can have glusterbot watch the rss feed for the serverfault tag and post those questions on irc
13:14:16 <amye> #chatops ++
13:14:19 <kkeithley> +1
13:14:54 <ndevos> #link https://public.pad.fsfe.org/p/gluster-office-hours
13:15:24 <ndevos> not sure how well we're doing with them, but the hours I am listed are blocked for meetings in my agenda
13:16:06 <amye> Heh, I am not sure how effective these are, given as I was not aware that we had them (which is fine) and how many blank spots there are.
13:16:15 <ndevos> well, I think thats it for today...
13:17:02 <msvbhat> FWIW there are gluster related questions even in Quora
13:17:04 <ndevos> amye: well, at least we have some people answering questions on irc now, its a start :)
13:17:13 <msvbhat> Not many recently thought
13:17:16 <msvbhat> *though
13:17:16 <ndevos> some *more* people
13:17:40 <ndevos> msvbhat: I never heard of Quora before?
13:17:52 <msvbhat> ndevos: W00t
13:18:00 * ndevos lives under a rock
13:18:02 <msvbhat> https://www.quora.com/topic/Gluster/
13:18:48 <msvbhat> jdarcy used to answer a lot in there. But now the questions about gluster are very less there
13:19:16 <ndevos> oh, okay, that sounds as something amye can think about too then :)
13:19:51 <ndevos> we really have to stop for the day, thanks for joining all!
13:19:55 <ndevos> #endmeeting