12:01:10 #startmeeting 12:01:10 Meeting started Wed Dec 17 12:01:10 2014 UTC. The chair is hagarth. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:01:10 Useful Commands: #action #agreed #halp #info #idea #link #topic. 12:01:14 Awesome :) 12:01:39 https://public.pad.fsfe.org/p/gluster-community-meetings 12:01:52 #topic roll call 12:01:57 * raghu is here 12:01:57 who do we have here today? 12:02:00 Moi 12:02:04 * ndevos _o/ 12:02:09 * lalatenduM is here 12:02:22 sounds great, let us move on to AIs from last week. 12:02:33 believe me the list is the longest I've seen in a while :) 12:02:39 hi (long time) 12:02:40 so bear with me while we go through the list 12:02:48 lpabon: welcome back! 12:03:03 hagarth: thanks! :-D 12:03:04 #topic AI review 12:03:12 hchiramm sends a patch that adds the new maintainers/components to the MAINTAINERS file 12:03:15 this has been done 12:03:25 * hchiramm is here 12:03:25 I intend to massage it a bit and merge it later this week 12:03:27 yep. done 12:03:28 not merged I think 12:03:34 yeah 12:03:34 not merged 12:03:41 filed as rfc? 12:03:53 #action hagarth to merge MAINTAINERS file this week 12:04:02 next AI: JustinClift fix the "wiki" link on gluster.org to redirect to the wiki and not to the blog 12:04:24 Didn't do it. Won't do it. 12:04:34 * JustinClift isn't interested in the website stuff, and I'm going on leave 12:04:41 Someone else needs to do it 12:04:49 JustinClift: are you the wrong person to own this AI? :) 12:04:53 Yep :) 12:05:05 JustinClift: can you suggest an alternate for this? 12:05:09 misch: 12:05:11 misc 12:05:22 ok 12:05:31 Actually, tigert for the design, and misc to do the actual roll out 12:05:37 #action misch and tigert to fix the "wiki" link on gluster.org to redirect to the wiki and not to the blog 12:05:48 misc still seems to want other people to do the commits... but hassle him enough and he'll do them himself ;) 12:05:53 maybe we can discuss this on gluster-infra 12:06:01 looks like a mail in gluster-infra :0 12:06:02 Sure 12:06:03 :) 12:06:15 next AI - hchiramm will try to fix the duplicate syndication of posts from blog.nixpanic.net 12:06:21 this is TBD 12:06:26 #action hchiramm will try to fix the duplicate syndication of posts from blog.nixpanic.net 12:06:33 hagarth, yet to do. 12:06:37 next AI - Debloper to write  a blog post about complete revamp 12:06:38 discussing with JustinClift 12:06:42 hchiramm: ok 12:06:57 don't think we have debloper here, will send him a reminder 12:07:08 hagarth, I will check with hum 12:07:11 hum/him 12:07:14 hchiramm: thanks! 12:07:22 next AI - jdarcy schedules a "small file performance" meeting 12:07:34 * jdarcy hangs his head in shame. 12:07:41 lol 12:07:47 :) 12:07:49 an AI in present continuous tense :) 12:08:00 Very continuous. Yes, still pending. 12:08:01 jdarcy: how about monday next? 12:08:23 Week after Xmas won't work. I'll be busy sliding down a mountain. 12:08:43 I could do this coming Monday, though. 12:09:08 Let us aim for it. we would need Ben, shyam and Rudra Siva probably in addition to the usual suspects. 12:09:17 OK. 12:09:32 * ndevos is interested in that too :) 12:09:44 #action jdarcy to schedule small file performance meeting for Dec 22nd 12:09:54 hagarth: Btw, hchiramm has admin on the WordPress site now. So, he's empowered to fix the duplicate syndication stuff 12:10:08 JustinClift: awesome, thanks for that! 12:10:12 :) 12:10:23 next AI - builds in ubuntu 12:10:35 kaleb and partner_ seem to be looking into this one 12:10:48 partner_: if you are around, any update on this one? 12:10:50 JustinClift, thanks 12:11:18 moving on to next AI 12:11:21 atinmu continues looking into the mgmt_v3 regression test issue 12:11:33 we have a patch for this now, needs review & merge 12:11:42 kshlm: are you looking into this patch? 12:12:02 maybe a question for kshlm later 12:12:06 next AI - atinmu will look for someone that can fix the spurious fop-sanity test 12:12:21 don't think we have atinmu today, let us carry this further 12:12:26 #action atinmu will look for someone that can fix the spurious fop-sanity test 12:12:38 next AI - atinmu will contact JustinClift to update the regression.sh script and capture /var/log/messages too 12:12:39 #action kshlm to review into the mgmt_v3 regression test patch 12:12:41 ;) 12:12:45 JustinClift: thanks :) 12:12:57 I think this is complete, JustinClift ? 12:13:21 I haven't updated regression.sh, as I was under the impression it turned out to be unneeded 12:13:30 If that's wrong, I'm happy to update it tho 12:13:32 ??? 12:13:47 JustinClift: I thought we decided to capture messages into the tarball 12:13:48 I think it could help in future? 12:14:00 k, I'll update it then 12:14:15 Probably be live on the slaves in about an hour from now 12:14:25 #action JustinClift to update the regression.sh script and capture /var/log/messages too 12:14:42 next AI - raghu` creates a 3.6.2 beta today (10 december) or tomorrow, release follows later this week or early next week 12:14:58 we have the tarball generated today 12:14:59 I have made the beta1. 12:15:18 hchiramm, lalatenduM: are you folks going to spin the RPMs? 12:15:27 hagarth, yeah 12:15:28 once the RPMs are out we can blog & send out a note 12:15:32 lalatenduM: cool, thanks 12:15:42 next AI - JustinClift and misc should post an update or plan about upgrading Gerrit 12:15:49 will follow up.. 12:16:07 gerrit upgrade is still pending, maybe one of the first things we'll do in 2015 :) 12:16:08 hagarth, upgrading Gerrit IMO is imp 12:16:14 lalatenduM: +1 12:16:32 But there are still some USS (user serviceable snapshots) related patches that have to be backported. I want to do a beta2 before making 3.6.2. Is that ok? 12:16:36 It should be easy to do... pretty much shut down gerrit, backup existing data dirs, and then put the new .jar in place 12:16:38 lalatenduM: check the agenda page, JustinClift seems to have a plan. 12:16:39 it will enable to plug sonar in to gerrit 12:16:47 raghu: let us discsuss this in the 3.6 topic 12:16:48 hagarth, checking 12:17:09 next AI - JustinClift to respond to Soumya Deb on gluster-infra about website 12:17:20 Give that to tigert 12:17:28 we need tigert and debloper to co-ordinate here. 12:17:42 Bother tigert and Soumya are invested in the GlusterFS website 12:17:44 #action tigert to respond to Soumya Deb on gluster-infra about website. 12:17:49 tigert already has next version designs 12:17:55 JustinClift: ok 12:18:13 I'm unsure if they're done in co-ordination with Soumya tho. Hasn't sounded like it. :/ 12:18:53 JustinClift: let us loop all of them into gluster-infra 12:19:04 and carry out our discussions there 12:19:21 next AI - Discuss stabiity issues and directions 12:19:44 I think this was from davemc, since we don't seem to have more context I prefer nuking this. 12:20:00 anybody has objections to drop this from our AI list? 12:20:30 none I guess 12:20:31 No objections atm 12:20:38 will drop that for now 12:21:01 that ends our AI list :) 12:21:04 #topic Gluster 3.6 12:21:19 raghu: can you provide an update on the current status of 3.6? 12:21:28 I have made the beta1. 12:21:38 \o/ 12:21:53 yay! a fair number of fixes already seem to be in 3.6.2beta1 12:22:08 But there are still many changes that have to go in (mainly USS related patches) 12:22:13 raghu: congratulations on your first glusterfs beta release :) 12:22:24 ha ha. Thanks :) 12:22:37 raghu: do we plan a beta2 soon then? 12:22:38 You never forget your first. Mainly because of the scars. 12:22:50 I want to do beta2 before making 3.6.2 12:23:13 quite a few dependent bugs for 3.6.2 open - https://bugzilla.redhat.com/showdependencytree.cgi?maxdepth=2&hide_resolved=1&id=glusterfs-3.6.2 12:23:30 because I am backporting many of the USS related patches in next few days 12:23:32 raghu: can you propose a date by which you need patches for these bugs? 12:23:34 yeah 12:23:51 if you don't get them, let us move the bugs to 3.6.3 12:23:54 I prefer end of this week. 12:24:02 Ok. Sure. 12:24:13 I suspect that we will have to do quite a few releases in 3.6 12:24:25 Based on the patches I get by this weekend I either do a beta2 or 3.6.3 itself. Is that ok? 12:24:35 oops 3.6.2 12:24:38 raghu: let us do a beta2 by end of this week 12:24:49 hagarth: ok 12:24:57 and probably schedule 3.6.2 as a xmas present to the community :D 12:25:22 raghu: please drop a note on the mailing list about this schedule & deadline for patches 12:25:34 I would have liked it more if all the USS related patches were backported before 3.6.2 12:25:43 hagarth: sure 12:25:51 raghu: fwiw, when I do a beta for 3.5, no patches get merged before a release, more patches mean an other beta 12:25:57 raghu: let us aim to have that. Also do call out if you need help with reviews. 12:26:12 raghu, I think we can wait for all the USS related patches were backported before 3.6.2 GA 12:26:42 ndevos: Good point. Thats why I was waiting till now for making beta1 12:26:45 I think we would need to merge ec and rdma patches too. 12:27:01 I would love to lift the tech preview embargo on rdma in 3.6.2 12:27:19 cool 12:27:20 FYI, the slaves all have regression.sh updated now. It should capture /var/log/messages* from here onwards 12:27:31 JustinClift: awesome, thanks! 12:27:34 :) 12:27:38 anything more on 3.6? 12:27:45 Nope 12:27:55 I will send a mail to mailing list about deadline and patches 12:28:03 #action raghu to send out a mail to communicate 3.6.2 schedule and patch deadlines. 12:28:09 moving on to 3.5 12:28:11 #topic 3.5 12:28:24 ndevos: any updates here? 12:29:05 no, not really... 12:29:23 ndevos: any outstanding patches in release-3.5 that need review attention? 12:29:39 no plans for a release, nothing very urgent that needs fixing that was brought to my attention 12:30:05 several minor changes have been merged 12:30:08 http://review.gluster.org/#/q/branch:+release-3.5,n,z seems to have mostly rdma fixes 12:30:40 yeah, they can use some review 12:30:59 ndevos: ok, let us notify Raghavendra G & Rafi 12:31:21 anything more on 3.5? 12:31:42 Rafi should be aware already, but no harm in pinging him again 12:31:47 no, nothing from me 12:31:53 ndevos: ok 12:31:55 moving on 12:31:57 #topic 3.4 12:32:04 kkeithley: any updates here? 12:32:11 .. 12:32:11 no, nothing of note 12:32:13 .. 12:32:14 . 12:32:22 I am yet to find some time to figure out the memory leak in 3.4 12:32:28 same here 12:32:43 lalatenduM: was that some form of morse code? 12:33:00 don't have a morse decipher handy ;) 12:33:07 hagarth, nope , sorry abt that, not intended :) 12:33:17 lalatenduM: ok :) 12:33:19 moving on 12:33:26 #topic Gluster next 12:33:39 3.7 has had a few new feature pages added 12:33:52 upcalls, CLI for nfs ganesha etc. 12:34:04 bitrot design document is also out for review 12:34:19 please feel free to review and provide your comments 12:34:35 I expect more feature pages to appear in the next few weeks 12:34:52 any thoughts, questions on 3.7? 12:35:11 Is there a "theme" for 3.7? 12:35:28 I mean, something we can state briefly when people ask? 12:35:29 jdarcy: data services - tiering, bitrot etc. 12:35:44 we need a blog :) 12:35:57 nah that can't be a theme 12:36:12 there are two ways of arriving at a theme 12:36:19 In the absence of a real community mangler, I could write a blog entry to drum up some interest. 12:36:31 jdarcy: +1 12:36:39 jdarcy, +1 12:36:47 one is proactively & the other would be a retrofit 12:37:03 for most community projects, we seem to do the latter :) 12:37:26 nevertheless we can try to be more proactive .. 4.0 already has well defined themes 12:37:54 For a theme for 3.7, I'd like us to concentrate on code cleanup and stability 12:38:09 Features are secondary. But that's just me, and unrealistic. :) 12:38:20 performance improvement is definitely I would like to see as a theme for 3.7 12:38:43 By code clean up I mean stuff like "prepare the code, for future big changes" 12:39:16 eg like the stuff kshlm did for getting the name/address stuff in good shape 12:39:23 JustinClift: part of it is in progress, the glusterd folks are doing that. quota & geo-replication will most likely get a cleanup in 3.7. 12:39:24 But for other things too 12:39:29 np :) 12:39:32 I do need to think some more about what "4.0 prep" might belong in 3.7 12:40:18 jdarcy: right, there is some ongoing work on sharding too which might land in 3.7 .. so there are bound to be overlaps between 3.next releases & 4.0 12:40:23 It'd be good to get IPv6 support into 3.7 too I suppose. 12:40:41 No idea of how important/priority/effort-requred tho 12:40:51 JustinClift: yes, recently I came across a IPv6 only cloud 12:41:02 * JustinClift is just spitballing, and this isn't the place for that ;) 12:41:19 That's what the wiki is for. 12:41:35 and somebody else mentioned that due to the address crunch in v4, IPv4 addresses/network was being charged while IPv6 was being given out for free. 12:41:49 JustinClift: please feel free to open a few feature pages :) 12:42:06 BTW, some of us are looking at freedcamp for 4.0 PM. People might want to think about it for 3.7 as well. 12:42:11 having feature pages will be useful in getting new contributors over 12:42:17 jdarcy: sounds like a good idea 12:42:41 jdarcy: any updates from 4.0 that you want to share here? 12:42:59 Only one thing, really... 12:43:28 We've had some discussions about sharing infrastructure for databases and logs, since several features depend on one or the other. 12:43:57 right.. 12:43:57 I invite anyone working on (or interested in) such a feature to make sure their needs in that area are represented on the 4.0 feature pages. 12:44:22 Otherwise we'll have three databases and four logs in our I/O stack, and we'll be slower than molasses in January. 12:44:54 jdarcy: ok, what are the database and log generators that you foresee? 12:44:59 changelog, NSR, tiering? 12:45:45 There has been talk of it in AFR. Some of what AFR/georep do is also related in purpose (even if it uses links or xattrs currently). 12:45:53 Also talk in EC, I meant to say. 12:45:54 (is meeting now?) 12:45:58 right 12:46:24 we can also throw quota into the mix 12:46:24 purpleidea: Yeah, we're at the end of it nearly 12:46:29 purpleidea: Last few minutes 12:46:36 hagarth: Yes, I knew there was at least one I was forgetting. Thanks. 12:46:55 Even DHT might have a use for such a thing. 12:47:03 jdarcy: should we start a discussion on gluster-devel? 12:47:10 right 12:47:12 hagarth: Yes. 12:47:32 jdarcy: let us do this.. else we'll end up fsyncing our way to performance hell. 12:47:33 JustinClift: better late than never? ;) 12:47:39 purpleidea: :) 12:48:10 BTW, information from those same logs might be useful for performance monitoring/tuning. 12:48:39 jdarcy: agree, we could probably look at making it more consumable/presentable 12:49:24 ok, anything more on Gluster.next? 12:49:38 * jdarcy shakes his head. 12:49:44 OK, let's move on 12:49:46 #topic Other Agenda items 12:50:03 we discussed Gerrit & Ubuntu already 12:50:38 random question - do any of us happen to notice duplicate posts from gluster.org on Facebook/twitter? 12:51:01 Not so much dups as *really old* stuff reappearing. 12:51:03 Hmmm, I think that used to be a problem with Twitter. Haven't looked recently tho 12:51:08 jdarcy: Yeah 12:51:39 is that @glusterfs or something else? 12:51:47 Don't recall one for the last coupla weeks though. 12:51:55 Hmm, we need to figure out what's happening here. Twitter, blog & Facebook are linked in some fashion. An update somewhere probably triggers duplication/stale reposts. 12:52:05 I don't know the mechanism that's used to push the posts to thigs 12:52:06 things 12:52:12 kkeithley: @gluster.. unfortunately we don't own @glusterfs 12:52:28 oh, haha 12:52:43 hagarth: Are you sure? 12:52:50 JustinClift: absolutely 12:52:56 lol, https://twitter.com/gluster has my last blog post 4x 12:52:59 hagarth: Didn't Twitter give that to us recently, after we asked them for it? Or was that a different one? 12:53:10 It's the Twitter equivalent of a domain squatter. Wants $$$ IIRC. 12:53:22 JustinClift: haven't heard that 12:53:25 * JustinClift could have sworn we got some version of that rixed not long ago 12:53:36 ndevos: yeah, we need to sort this over. 12:53:50 ndevos: could be linked to the duplicate syndication problem 12:53:50 #action hagarth to contact Rikki Endsley about fixing the twitter glusterfs domain squatter problem 12:53:59 JustinClift: will do 12:54:35 another topic - we need an interim replacement for JustinClift to manage/moderate mailing lists when he is away 12:54:40 any volunteers? 12:55:02 It just needs someone to go through the queue once per day, and approve the messages that are valid 12:55:08 It's 90% spam 12:55:08 We do have a trademark on "GlusterFS" don't we? That should be sufficient as far as Twitter is concerned. 12:55:21 jdarcy: Yeah, I don't think it'll take us much effort 12:55:26 jdarcy: yes the gluster.org board owns the trademark IIUC. 12:55:37 hchiramm: can you help with moderation? 12:56:02 yep 12:56:08 hchiramm: awesome, thanks! 12:56:10 Unfortunately, everyone from Red Hat US tends to be on forced leave over the holidays. 12:56:22 JustinClift: you have your answer here for list-admin :) 12:56:33 hchiramm: it is. :) 12:56:45 :) 12:56:50 shall we cancel the community meetings for next 2 weeks? 12:56:58 +! 12:57:04 +1 12:57:13 +1! 12:57:17 hchiramm: Ahhh, you already get the mailing list notifications yeah? 12:57:28 yep 12:57:38 k. I'll leave them to you to do then. 12:57:39 OK, next meeting will be on Jan 7th. 12:57:44 another listed topic from purpleidea 12:57:48 purpleidea: go ahead please 12:57:56 have a look at: https://github.com/purpleidea/puppet-gluster/commit/9b5f690da9f6c000325682f5b8e0b8003eeb5219#diff-7cb3f60a533975d869ffd4a772d66cfeR1 12:58:06 puppet-gluster needs to know about the operating version table... 12:58:17 i made it easy to add new version mappings by dropping in a yaml file... 12:58:36 so if anyone is bumping the versions, please send me a patch, or poke me if you remember :) 12:58:40 purpleidea: cool, can you please also drop a note on gluster-devel? 12:58:46 hagarth: sure! 12:59:03 purpleidea: we normally bump the op-versions at least once per major release 12:59:19 kshlm and others would be interested in this. 12:59:30 we have one more minute left .. any other topics? 12:59:54 any plans for FOSDEM? 12:59:58 Happy holidays everyone that's doing them? 13:00:07 i'll be at config mgmt camp and maybe fosdem too 13:00:10 or, the CentOS Dojo happening the day before? 13:00:15 RH OSAS team will almost all be at FOSDEM is my understanding, if that's useful 13:00:20 ndevos: If my talk gets accepted, I will be there 13:00:40 ah, cool, I'll be there - and I think kkeithley too 13:00:54 ndevos: maybe we could have a Gluster BoF session there 13:00:59 I registered, yes 13:01:25 for Fosdem check http://lists.centos.org/pipermail/centos-promo/2014-December/001492.html 13:01:31 jdarcy: Freedcamp != open source :( 13:01:52 I might go too, if my talks gets accepted 13:01:58 JustinClift: True enough, but it's hard to find a service in this particular space that is. 13:02:01 we probably need a community events calendar for Gluster .. let's have one in 2015 13:02:02 +1 for a BoF 13:02:02 hagarth: or a gluster meetup/drinks/dinner with people who are interested to talk glusterey things 13:02:23 ndevos: sure, anything works 13:02:28 jdarcy: Damn. Once of those non-optimal situations. Oh well. ;) 13:02:45 I was tempted to go to the Gluster meetup in Finland, but it's nine days before Fosdem 13:03:15 kkeithley: yeah, that's unfortunate. 13:03:21 ah, right, I wanted to ask you about that 13:03:25 alright folks, we seem to be done for this year :) 13:03:37 happy holidays to everyone and see you all in 2015! 13:03:39 :) 13:03:46 Thanks everyone. Have fun, etc. :D 13:03:48 wohoo, I can go home now! 13:03:51 same to you :) 13:03:56 #endmeeting