12:01:28 #startmeeting 12:01:28 Meeting started Wed Feb 11 12:01:28 2015 UTC. The chair is ndevos. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:01:28 Useful Commands: #action #agreed #halp #info #idea #link #topic. 12:01:34 Hey all! 12:01:38 hey all 12:01:38 :) 12:01:40 hi 12:01:41 ndevos: o/ 12:01:47 hi 12:01:51 new member here <-- 12:01:55 Agenda for today: https://public.pad.fsfe.org/p/gluster-community-meetings 12:01:59 itpings: welcome aboard! 12:02:02 itpings: Welcome :D 12:02:07 #chair hagarth 12:02:07 Current chairs: hagarth ndevos 12:02:12 ty guys 12:02:21 #topic Roll Call 12:02:41 hagarth: you want to run the meeting, or do you need to leave early? 12:02:59 ndevos: I am bound to be distracted in between today, so please go ahead 12:03:10 hagarth: ok, no problem 12:03:36 heeep 12:04:22 hmm, we seem to be missing some of the regular attendees... 12:04:30 * kkeithley is here 12:04:39 * hchiramm is here 12:04:41 * jimjag is here 12:05:17 #topic Action Items from last week 12:05:31 #info Subtopic: ndevos should publish an article on his blog 12:05:43 yeah, I still need to do that :-/ 12:05:58 #info Subtopic: hchiramm will try to fix the duplicate syndication of posts from ndevos 12:06:12 and that is waiting on my post... 12:06:19 thx :) 12:06:27 #info Subtopic: hchiramm will start a discussion on the mailinglist about the RHEL/Community packaging issues and solutions 12:06:40 Yep, its progressing 12:06:41 :) 12:06:50 I did notice a discussion, but I do not think is went to the mailinglist yet? 12:07:07 I think we need to take into gluster MLs at this tage 12:07:10 tage/stage 12:07:13 ndevos: Maybe just to a "test post (we're trying to fix an issue)" instead of something important? People probably won't be too bothered... 12:07:48 ah.. typo.. I think we dont need to take into gluster MLs at this tage 12:07:49 JustinClift: I'm sure I can come up with something better, and it is a non-urgent thing anyway 12:07:56 :) 12:08:03 as it involves the discussions about the downstream as well 12:08:17 hchiramm: okay, when something solid/final comes out, will you share it with the lists? 12:08:25 yeah, that can be done 12:08:52 #action hchiramm will share the outcome of the non-mailinglist packagng discussions on the mailinglist 12:09:16 #info Subtopic: hagarth to open a feature page for (k)vm hyperconvergence 12:09:22 Should we ask the gluster-board for their sign-off on the solution? 12:09:46 hagarth mentioned by email that he will have that "By next week for certain" 12:10:06 Hmmmm... 12:10:09 k 12:10:44 JustinClift: I do not know how/if the board should be involved, it is something many community projects need to handle, its not Gluster Community specific (the packaging) 12:11:01 ndevos: yes, will do that by next week. sorry for missing out on this one for a few weeks now! 12:11:14 hagarth: np, we'll keep reminding you :) 12:11:25 ndevos: The board has technical users, and had a lot of strong feeling about the problem itself 12:11:26 * kshlm is here 12:11:27 * jimjag suggests just making sure the board knows is likely a Good Idea 12:11:47 An engaged board is a good board 12:11:55 ndevos: Not at least keeping them in the loop is probably not a great idea 12:11:58 jimjag: +1 12:12:19 JustinClift: I think they will be kept in the loop through hchiramm? 12:12:58 Sure. I'm just suggesting hchiramm makes a point of it :D 12:13:02 As a board member myself, I can report that the board, as an entity, feels out of the loop 12:13:14 #action hchiramm should keep the Gluster Community Board in the loop on the outcome of the packaging discussions 12:13:26 ndevos: Good idea :) 12:13:44 JustinClift, ndevos sure 12:14:03 JustinClift, jimjag: how would the board like to be kept in the loop? do they follow these meetings and the minutes? 12:14:37 jimjag: Yours :D 12:14:38 #info Subtopic: keep the Gluster Community Board more informed and involved 12:14:39 ndevos: I follow the meetings, for sure, but mailing the minutes specifically to the board list makes sense 12:15:18 jimjag: the board email list is marked as private, can anyone post things there without getting stuck in the moderation queue? 12:15:31 I will mod it thru 12:15:39 I approve stuff that seems appropriate 12:15:56 So, moderation queue shouldn't be a blocker 12:16:09 hmm, okay, if you guys say so 12:16:40 whats the email address of the board list again? 12:17:04 board@gluster.org 12:17:15 * msvbhat arrives late to the party 12:17:17 #action *all*moderators* should post meeting minutes to the Gluster Community Board 12:17:26 http://www.gluster.org/mailman/listinfo/board 12:17:54 #info Subtopic: spot to investigate repetitive posts on social networking sites 12:18:12 hchiramm, hagarth, JustinClift: do you know about that? 12:18:48 is it still an issue, the topic has been around for months already? 12:18:57 ndevos, unfortunately no 12:19:09 Repetative posts? 12:19:10 examples? 12:19:11 partner: I dont know, and nobody seems to know the progress :-/ 12:19:26 i suggest we trash the topic and open it up again if the issue persists 12:19:39 yeah, I'm thinking about that as well 12:19:50 its 1/4th of the meeting around that topic each week 12:20:13 btw, was topic #1 about getting people to package qemu with glusterfs support? 12:20:21 (sorry for being late, just had a phone call) 12:20:36 #agreed drop the topic from the list, if the issue happens again, file a bug against the project-infrastructure component in Bugzilla 12:20:52 telmich, no.. it was about rpm versions of glusterfs-* 12:20:57 hm, ok 12:21:20 telmich: maybe, I think that is a little related, but you can add that topic to the "Open Floor" on the agenda if you like 12:21:35 * ndevos hands https://public.pad.fsfe.org/p/gluster-community-meetings to telmich 12:21:39 ndevos: open floor is at the end? 12:21:42 ndevos: gracias! 12:21:59 #info Subtopic: spot to reach out to community about website messaging 12:22:14 I do not think that has been done? 12:22:31 * ndevos adds that as a point to yet an other discussion 12:23:17 #info Subtopic: hagarth to carry forward discussion on automated builds for various platforms in gluster-infra ML 12:23:24 ndevos: again for next week 12:23:32 hah, sure! 12:23:46 do I sound like a broken record? ;) 12:24:26 who wants to make a bet if hagarth manages to get the two action items done by next week? 12:24:47 pint for each completed item! 12:24:47 * JustinClift puts money on it for "no" 12:24:57 * hagarth puts his money on "yes" 12:25:05 ndevos: hagarth: There was some minor issue while using autoo-build repo for installing. 12:25:09 How much moolah? :D 12:25:15 we can discuss details about that after the meeting! 12:25:44 ndevos: sure. I don't have a fix suggestion anyway 12:25:46 msvbhat: yeah... maybe get some ideas hashed out on the -devel list about that? 12:26:04 #topic Gluster 3.6 12:26:09 raghu`: you're up! 12:26:14 ndevos: Yeah. That would be better 12:26:35 ndevos: I have made glusterfs-3.6.3beta1. 12:26:52 Tarball is ready. RPMS have to be created 12:26:53 raghu++ thanks.. :) 12:27:06 hchiramm: will you run the rpm builds? 12:27:10 yep 12:27:10 and kkeithley the debs? 12:27:44 raghu`: how long before you think it should become a non-beta release? 12:27:57 raghu`: Does it have the fixes we've found for Corvid Tech issues so far? 12:28:07 * JustinClift is wondering whether we encourage them to try it 12:28:44 I want to make 3.6.3 before Feb 20. 12:28:45 There are still some patches which I have to take in (either they have to pass the regression tests or need an ack) 12:29:38 raghu`: Are you ok to highlight them on the -devel mailing list, and encourage people to test them? 12:29:58 eg "These are bug fixes for 3.5.3, and we need reviews/ack's on them" 12:29:59 raghu`: okay, so when more patches get merged, you'll do a beta2? 12:30:12 raghu`: ah, and include some release notes too? 12:30:27 JustinClift: I am not sure whether beta1 contains Corvid Tech issues (I am not complete sure about those issues though). Beta1 mostly contains ec fixes, some afr fixes andsome posix fixes 12:31:07 Hmmmm, there's a dht issue that was being looked into. No idea if it's become a patch tho. 12:31:19 Maybe for betat2 (or .4), depending... 12:31:19 :) 12:31:20 JustinClift: Yeah. I can inform the -devel list about beta1 (should I do it after rpms are ready or right now)? 12:31:32 Right away? 12:31:37 ndevos: yeah. Want to do a beta2 once those pathces are in 12:31:45 Might get more reviews in quicker 12:31:53 yes, one of us, semiosis or me, will make them. Probably me 12:31:57 ndevos: yeah. sure. Will include release notes as well. 12:32:09 raghu`: okay, thanks, sounds good 12:32:23 raghu`: anything else to add? 12:32:30 JustinClift: Ok. Will send a mail to -devel list about the availability of 3.6.3beta1 today 12:32:54 ndevos: nope. But thanks a lot to you for the script 12:32:57 :) 12:33:01 it helped a lot 12:33:12 #action raghu` will send a notification about 3.6.3beta1 to the mailinglists 12:33:15 for moving the status of the bugs 12:33:16 sure, np! 12:33:38 many of us got mail-bombed by all the closure messages :) 12:33:46 I think its better to announce after rpms are available 12:33:47 ;) 12:34:23 Ahhh. What I said before about "right away" vs now 12:34:41 If its for people to download a new release, then yeah, after we have rpm's ready 12:34:47 hchiramm, raghu`, kkeithley, semiosis: you guys should discuss when to announce the beta release - but not here/now 12:34:58 * JustinClift shuts up :D 12:35:15 #topic Gluster 3.5 12:35:35 there has been some progress there, more patches merged 12:36:01 a beta1 for 3.5.4 will probably be done somewhere next week 12:36:23 I dont remember of any very urgent issues, so there is no need to rush it 12:36:54 #action ndevos to prepare for the next 3.5 beta/update somewhere next week 12:37:02 questions on 3.5? 12:37:31 did I hear a "no"? 12:37:38 #topic Gluster 3.4 12:37:50 kkeithley: please tell us all about it! 12:37:56 nothing to report. I still consider the memory leak the major show stopper 12:38:37 okay, and nobody is questioning you on that either, it seems 12:38:38 and I don't think there are fixes for the other 3.4 bugs in the trackker, so... 12:38:55 no, I'm working to try to narrow down the leak 12:39:04 well, I'm sure people reported bugs against 3.4 that are not on the tracker? 12:39:24 yes, probably 12:39:48 but, if there is nothing urgent, well, you're old enough to decide on them yourseld :P 12:40:04 ;-) 12:40:16 #topic Gluster Next 12:40:17 I haven't heard anyone clamoring for 3.4.7, so 12:40:26 #info Subtopic: Gluster 3.7 12:40:41 hagarth: anything you wanted to say about 3.7? 12:41:42 ndevos: sure 12:41:44 anyone else wants to tell us about 3.7? 12:41:51 ah, there you are :) 12:41:57 we've got some feature pages updated 12:42:40 looks like we'll have bitrot, tiering, sharding, some improvements in small file performance, self-healing improvements and more in 3.7 12:43:19 request all feature owners to submit patches sooner and not await the last week for getting all reviews done 12:43:34 that's my short update on 3.7 12:43:50 Um... for 3.7... 12:44:00 hagarth: how do you keep track of the progress of the patches for these features? 12:44:31 * JustinClift suggests we ask the board (directly) for their input. eg Ask JimJag to engage them 12:44:47 ndevos: it is rudimentary - looking up the feature page and content out for review ;) 12:45:16 hagarth: lol! 12:45:23 ndevos hagarth: Does it make sense to ask feature owners to add the CR #'s to the feature page? 12:45:42 JustinClift: I think that is what most do 12:45:50 We also have 2 big changes for GlusterD, but they are not exactly features. (Daemon code refactoring and RCU protection). Would these be thought of as features and would they be bound to the feature freeze date? 12:45:58 * kshlm just wants to know 12:46:25 kshlm: If they're not mentioned on the feature page, it'll be hard to remember to add info about them into press releases, general announcements, etc 12:46:29 Where it the bitrot code right now? I'd like to get a head start on reviews. 12:46:38 kshlm: the refactoring is less of a feature, I think (unless it is user visible) 12:47:09 kshlm: but, the liburcu seems to be a good feature to promote, other areas should probably adopt that too one day 12:47:33 jdarcy: bit-rot code should be out for review soon. atleast by early next week 12:47:53 raghu`: It's in a private/downstream repo then? 12:48:01 jdarcy: I'm not sure where bitrot lives, but data-classification is on the forge 12:48:43 nope. Currently the changes are being added to our github repos. 12:48:43 raghu`: getting it on the forge, and include design documents and user documentation would be good 12:48:45 ndevos, that sounds reasonable. 12:49:27 #info Subtopic: Gluster 4.0 12:49:50 jdarcy: please educate us about the latest news 12:50:05 We had a very productive IRC meeting on Friday. 12:50:10 I'll add a new section "Other big changes" to Planning37 and these under that. 12:50:29 Participation/interest seemed high, several action items came out of it. 12:50:29 kshlm: sounds good to me 12:51:14 http://meetbot.fedoraproject.org/gluster-meeting/2015-02-06/glusterfs_4.0.2015-02-06-12.05.html 12:51:33 Looks like we have a couple of new volunteers for composite operations. 12:51:54 Other than that, nothing to add to what's in that log. 12:52:35 okay, thanks 12:52:41 Oh, Xavier and I have had some subsequent discussions about DFC (coordination framework for EC and others). 12:53:02 I understand that you plan to do bi-weekly meetings for 4.0? 12:53:13 Yes. 12:53:33 nice :) 12:53:40 #topic Open Floor 12:53:57 #info Subtopic: Maintainer responsibilities 12:53:59 Next one is likely to be February 26 or 27, right in the middle of my vacation. 12:54:20 oh, okay, you should find someone to take charge of that one then 12:54:29 Oh, I'll be there. That's how life is. 12:54:44 ah, ok 12:54:57 It's either that or spend time with my family. ;) 12:55:09 anyway, I'll probably send out a reminder to the -devel list about the maintainer responsibilities 12:55:23 we are your family too! 12:55:34 as a newbie here, can I quickly ask who is responsible for what in here? 12:55:59 #action ndevos should send out a reminder about Maintainer responsibilities to the -devel list 12:56:06 telmich: we can do that after the meeting, ok? 12:56:10 ndevos: ok 12:56:24 #info Subtopic: Upcoming talks at conference 12:56:39 reminder to update https://public.pad.fsfe.org/p/gluster-events 12:56:55 add conferences to it. add your name and details! 12:57:09 #info Subtopic: GSOC 2015 12:57:22 kshlm: a short status update? 12:57:57 JustinClift: maybe you know more about it too? 12:58:07 I sent out a mail asking for contributions to build an ideas list. 12:58:24 * JustinClift reads back 12:58:39 Nope, haven't been keeping an eye on this 12:58:43 There was response, and there were some contributions. 12:58:57 kshlm: okay, next steps? 12:58:57 We've still not started preparing a proposal. 12:59:11 We need help with this. 12:59:12 Hmmmm, someone asked me something about GSOC a few days ago, but I haven't read the email properly yet so no idea what it's about 12:59:34 I'd mailed JustinClift and spot for help, but didn't hear back. 12:59:47 I was busy with other stuff and couldn't follow up on that. 12:59:57 So JustinClift, can you help with this? 13:00:07 #action JustinClift should follow up on the GSOC email 13:00:39 kshlm: I think you guys should be able to figure something out by email (or on IRC but outside of this meeting) 13:00:53 Yup. 13:01:00 * ndevos notes we're running out of time... 13:01:06 Sure 13:01:10 #info Subtopic: Qemu / Clusterfs integration 13:01:23 alright 13:01:24 telmich: I think you added that to the list? 13:01:38 I was wondering what the status is regarding pushing qemu packages in the distros? 13:02:14 so far I can see that Debian doesn't pickup glusterfs support at all 13:02:23 well, for RHEL/CentOS, qemu contains glusterfs support out of the box 13:02:30 and ubuntu *does* have a ppa, but the default packages are not supporting glusterfs 13:02:34 neither in archlinux 13:02:49 I was talking to the package maintainers 13:03:00 but it seems there is no/not much interest from the community at the moment 13:03:24 partner, semiosis: you guys know more about Debian/Ubuntu - got any idea if/what qemu packages support glusterfs? 13:03:26 as I have seen that the patches are pretty old, I was wondering if anyone is in charge or looks for distro integration besides redhat? 13:03:56 ndevos: debian doesn't have it at all, neither in backports nor in testing nor in sid 13:04:00 * kkeithley is trying to find the packager list. Isn't semiosis working with a Debian person to get newer versions of gluster into Debian 13:04:11 there is one ppa, I guess my someone maintained in here, for ubunt 13:04:12 once we have that then qemu can be built with gluster support 13:04:47 found it. Patrick Matthaei 13:04:49 so, yeah, essentially my question is, is anyone driving non-redhat qemu+glusterfs support? 13:05:02 kkeithley: that is the-me in #gluster 13:05:11 ah, okay 13:05:42 telmich: I have no idea... maybe you should ask about that on the gluster-users list? 13:05:51 ndevos: I can do that 13:06:04 telmich: I think partner would be interested in that too - he seems to be a Debian guy 13:06:17 another short related question: is there *any* recommendation for which glusterfs version to use in production for VM hosting at the moment? 13:06:25 I have asked on the ML, but didn't get any response 13:06:44 #action telmich will send an email to the gluster-users list about Gluster support in QEMU on Debian/Ubuntu 13:06:46 partner: do you want to have a follow up after the meeting? 13:07:27 (background: we are a swiss linux infrastructure company and are about to change to gluster from ceph, but are missing some information towards what to pick) 13:07:49 telmich: I think the latest 3.6 should be stable enough for VM hosting - raghu`? 13:09:02 k, I'd quickly like to suggest that jimjag engage the board, and ask them directly if they have stuff they'd like in 3.7 and 4.0 13:09:13 I wonder who our VM on Gluster specialist is, we should get him/her involved on the mailinglist 13:09:18 jimjag: Workable? 13:09:30 JustinClift: +1 13:09:58 ndevos: I think yeah. 3.6.2 seems to be in a better state 13:10:00 #action jimjag to engage the board, asking for their direction and input for both 3.7, and 4.0 releases 13:10:07 jimjag: Tx :) 13:11:05 JustinClift, jimjag: there have been several requests for feedback on 3.7 and 4.0 features, I would expect the board to at least follow some mailinglist threads? 13:11:20 ndevos: Wrong approach 13:11:44 ndevos: That's the "we expect them to engage us" idea. This is more of the "lets actively engage them" approach 13:11:50 JustinClift: right, then we need someone that acts as a mediator between the community and the board? 13:11:58 jimjag is the man ;) 13:12:07 \o/ 13:12:18 okay, sounds good 13:12:40 telmich: raghu` is the maintainer for 3.6, so if there are any urgent issues, you need to bug him ;) 13:13:21 telmich: Btw, from memory the VM hosting software you use (eg oVirt, etc), can affect things too. It's probably worth having an in-depth discussion with the right people about that 13:13:27 ndevos: I understand 13:13:36 we're already 10+ minutes over the scheduled time, so lets finish it for today 13:13:42 :) 13:13:43 JustinClift: we are pretty much on opennebula for the moment, because it is easy to hack 13:13:51 ndevos: thanks for the pointer 13:13:58 telmich: Sure. Here isn't the right place to discuss tho 13:14:07 JustinClift: np 13:14:24 telmich: you can stay in the channel for a while, and ask more questions :) 13:14:27 #endmeeting