12:00:16 #startmeeting Weekly GlusterFS Community Meeting 12:00:16 Meeting started Wed Aug 20 12:00:16 2014 UTC. The chair is JustinClift. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:00:16 Useful Commands: #action #agreed #halp #info #idea #link #topic. 12:00:32 #topic roll call 12:00:36 Who've we got? :) 12:00:37 * kkeithley is here 12:01:30 * JustinClift guesses lalatenduM as well 12:01:36 * lalatenduM is here 12:01:51 JustinClift, :) 12:02:29 I'm becoming prety sure that someone did actually cheat with the voting for new meeting times. :/ 12:02:44 #topic Action items from last meeting 12:02:45 JustinClift, yeah I agree :) 12:03:21 Hmmm, pranithk isn't here. Can't ask him about his items. Will need to do that via email. 12:03:31 "lalatenduM send a bug triage email announcing the process, and call for volunteers" 12:03:41 lalatenduM: So, this week to be finished? 12:03:49 JustinClift, I have updated the wiki pages 12:04:04 * ndevos just arrived _o> 12:04:13 lalatenduM: What's the next bit? Announce via email 12:04:14 ? 12:04:18 JustinClift, the mail is in draft, I dont think it is is ready 12:04:23 so plz move it to next week 12:04:28 Sure 12:04:37 ndevos, I need ur help for this 12:04:38 #action lalatenduM send a bug triage email announcing the process, and call for volunteers 12:04:55 "JustinClift to get initial GlusterFS Consultants and Support Companies page online" 12:04:56 ndevos, will send you the draft 12:05:05 lalatenduM: sure! 12:05:07 That's mine dragging out more. Now we have the build instructions from Eco it's closer 12:05:28 "kkeithley to email gluster-devel about the reviews needed for cppcheck stuff on master" 12:05:42 Email done yeah? 12:05:54 yes, sort of 12:05:57 ? 12:06:13 I'll send an email today. 12:06:27 nagging, guilt, threats, bribes, nothing works. ;-( 12:06:45 Hmmm, what kind of bribery have you tried? :) 12:06:58 beer, chocolate, ice cream. ;-) 12:07:04 Heh Heh Heh 12:07:07 :) 12:07:08 LOL 12:07:41 k. hagarth's new review stats thing is meant to help with this, when it's tied in with actual rewards 12:08:04 eg when we're sure it's accurate and good, we can start giving some form of real world rewards to the most active reviewers 12:08:33 I'm not sure what the reward/encouragement system will be. But whatever it is we need to ensure it's effective 12:08:45 kkeithley: I guess until then just keep with the emailing/asking 12:08:55 JustinClift, I think it should just motivation to make Gluster better :) 12:09:02 kkeithley: Hopefully it won't take more than a few weeks to get this reward system to work 12:09:17 lalatenduM: Sure, but people have to choose between several priorities, all of which do that 12:09:28 lalatenduM: This is trying to encourage them to prioritise reviewing :) 12:09:33 yeah, I am sure it will invite more people to the review process.. 12:09:48 Anyway, moving on... 12:09:54 "JustinClift to create blog post of latest leaderboard stats, after receiving from hagarth" 12:09:58 JustinClift, excessive enumeration might be bad too :) 12:10:04 lalatenduM: Agreed 12:10:10 All reviews, no patches ;) 12:10:19 thats my personal opinion :) 12:10:28 JustinClift, apply the same for patches :) 12:10:33 I've received the latest leaderboard stats from hagarth. Need to make a blog post next :) 12:10:46 Yeah, we'll figure things out over time and adjust as needed I guess. 12:10:55 "JustinClift to encourage Tamas Papp to log a bug bout http://supercolony.gluster.org/pipermail/gluster-users/2014-July/018219.html" 12:10:58 Today 12:11:06 "ndevos will verify status of patches proposed for 3.5.3 and request reviews for those patches in master?" 12:11:10 ndevos? 12:11:30 mostly done, where there were patches - which is few 12:11:56 ndevos: When do you reckon it'll be done? 12:12:04 eg the verify status of patches for 3.5.3 12:12:11 after kkeithley's nagging :) 12:12:27 Due today then. :) 12:12:34 well, status check is easy, getting the reviews done is harder 12:12:57 k. When kkeithley sends nag email, can you follow that up with extra encouragement? :) 12:13:23 "JustinClift to test out Gitorious 3.x, to see if it's better than GitLab" 12:13:29 sure, I'll point out *again* that 3.5.3 needs the reviews in master first 12:13:33 Haven't yet done it. Got caught up with other stuff. 12:13:54 ndevos: Repeating things is a good idea. Reinforcement style. :) 12:14:10 "kkeithley will nag pranithk about fixes for 3.4.6" 12:14:33 not done 12:14:39 Next week? 12:14:44 yes 12:14:57 dht backport too 12:14:59 JustinClift: COnsidering I am here, not needed I guess 12:15:12 kkeithley: I will send the patch this week 12:15:15 So, consider it done. :) 12:15:40 "pranithk to send a list with components and their assignees - soliciting others to contribute too" 12:15:45 excellent 12:15:58 pranithk: Did that get done? 12:15:59 JustinClift: huh? I don't remember it 12:16:12 It's from a few weeks ago meeting 12:16:16 pranithk: possibly the design documentation? 12:16:22 JustinClift: ah! okay will do 12:16:31 JustinClift: or is it the QA list thing? 12:16:37 Heh Heh Heh 12:16:55 JustinClift: I am confused. I don't remember :-( 12:16:59 pranithk: probably worth a check in the meeting log archive 12:17:06 pranithk: and please file bugs for the single documentation parts :) 12:17:07 Yeah 12:17:15 hagarth: yeah 12:17:18 ndevos: yeah 12:17:33 pranithk: Go look through the meeting log archives now? We'll move onto the other items meanwhile and come back to this 12:17:43 JustinClift: yes 12:18:01 #action kkeithley DHT backport 3.5.3->3.4.6 https://bugzilla.redhat.com:443/show_bug.cgi?id=1116150, https://bugzilla.redhat.com:443/show_bug.cgi?id=1117851 12:18:03 Bug 1116150: high, unspecified, ---, nsathyan, POST , [DHT:REBALANCE]: Rebalance failures are seen with error message " remote operation failed: File exists" 12:18:04 Bug 1117851: high, unspecified, ---, jdarcy, POST , DHT :- data loss - file is missing on renaming same file from multiple client at same time 12:18:15 "ndevos to add branch checking to rh-bugs" 12:18:24 So, still being worked on then yeah? 12:18:28 yes 12:18:32 eg making sure it's ok before enforcing 12:18:33 Cool 12:18:53 just checked http://build.gluster.org/view/Bugzilla%20Checks/job/compare-bug-version-and-git-branch/ and there are some failures 12:18:57 #action ndevos to add branch checking to rh-bugs 12:19:04 ndevos: I'm sure you'll get then sorted :) 12:19:06 I'll have to check what those failures are, and if they are valid 12:19:35 Is "ndevos to update the rh-bugs script in Jenkins to check git/branch and bug/version" a different thing? 12:20:03 Or is that a subsequent task, after you've got it working well? 12:20:03 JustinClift: I think they are the same 12:20:10 * JustinClift nukes that item 12:20:16 JustinClift: yeah, should be the same 12:20:30 "hagarth to evolve plan for the test weekend and update schedule for 3.6" 12:20:43 We'll do this as part of 3.6 topic 12:20:53 "kkeithley to discuss -dev and -geo-rep subpackages for Debian with pmatthai" 12:20:56 Done 12:21:15 Sounds like we're just going to have everything in one package 12:21:37 kkeithley: Would you be ok to email gluster-devel to let everyone know how that went? 12:21:55 eg so ppl know everything will be in one package (or whatever) 12:22:17 there are four .debs in the Debian/Ubuntu packaging. semiosis put the "source" on github. Maybe we should put in in our source too, a la .../extras/Linux{RPM,dpkg)? 12:22:18 an email with the results of the discussion would be good, yes 12:22:39 +1 12:22:43 fwiw it was all discussed on #gluster-dev IIRC 12:22:48 but I'll send an email 12:22:51 Thx 12:23:02 kkeithley++ :) 12:23:14 We need a bot here 12:23:29 thoughts about .../extras/LinuxDPKG/ ? 12:23:32 #action kkeithley to email gluster-devel with info on how the Debian packaging will be done 12:23:40 kkeithley: No objections here 12:23:58 kkeithley: whatever makes it easy for people that need to build .debs 12:24:00 kkeithley: Will it be useful in some way, or is it a way to just not lose it? 12:24:05 glusterbot: kkeithley's karma is now 1 12:24:06 I'll bring it up on #gluster-dev 12:24:12 pranithk: Heh 12:24:23 kkeithley: Sure 12:24:29 JustinClift: do we need an action item to get a bot here? ;) 12:24:38 a way not to lose it. Certainly I'll use it (and others) to build for wheezy and jessie 12:24:42 kkeithley, no objections, should be fine 12:25:04 #action kkeithley to figure out if the .deb packaging info should be carried in main GlusterFS git repo 12:25:17 I believe semiosis is still the preferred packager for Ubuntu Launchpad PPAs 12:25:36 Anyone remember who set up the Karmba bots for the other channels? 12:25:41 kkeithley: I think he started a launchpad group for packaging? 12:25:42 * JustinClift can feel an action item coming on 12:25:49 JustinClift, JoeJulian ? 12:25:55 I think it is JoeJulian 12:26:25 #action hchiramm_ to ask JoeJulian to set up a karma bot for #gluster-meeting channel :) 12:26:36 sure :) JustinClift 12:26:41 :) 12:26:42 sounds good! 12:26:46 "hagarth, JustinClift and misc to evolve a proposal for web site maintenance" 12:26:52 In progress 12:27:03 "ndevos when we know how to update the webpage, add a pointer to proposed features for new contriburos" 12:27:13 In progress 12:27:15 :) 12:27:24 JustinClift, I propose "Who is who" page as well :) 12:27:25 Eco's email yesterday should get us going 12:27:32 Good idea 12:27:43 #action hchiramm_ to create a "Who is Who" page for GlusterFS 12:27:48 thanks :) 12:27:52 :) 12:27:56 haha :D 12:28:04 ndevos, :( 12:28:16 "kkeithley will talk to Ira and Jose about samba packaging." 12:28:29 "TBD" 12:28:31 I just sent them an email 12:28:53 "In progress. Initial email sent." 12:28:58 #topic 3.4 12:29:10 kkeithley: Go for it 12:29:32 nothing to report. Just waiting on critical mass of fixes for 3.4.6 12:29:37 #action pranithk to send backport for self-heald memory leak 12:30:22 :) 12:30:38 How's this item go? "Should the patch/fix backport request page just point to the respective tracker BZs?" 12:30:45 yes 12:30:53 Did we get any written instructions updated? 12:31:18 kkeithley: Manu found a NFS bug that can cause hangs on certain (NetBSD) NFS clients, want a backport for that? 12:31:33 sure 12:31:42 AI it 12:31:54 okay, I'll clone, "backport" and post it later 12:32:00 thanks 12:32:10 I think we need to update the backport request page 12:32:17 #action ndevos to clone, backport, and post the NFS bug that can cause hangs on certain (NetBSD) NFS clients 12:32:43 We need sections for 3.4.6 and also possibly for 3.6.0 12:32:49 hagarth: Who should do that? 12:32:55 lalatenduM: Do you have time to? 12:33:04 hagarth: I prefer to get backports requests as bug against the tracker... 12:33:18 ndevos: yes, we can add a link to the tracker bz from that page 12:33:20 Yeah. We can update the backport request page to instruct people to do that 12:33:38 eg link to each of the tracking BZ's and such 12:33:55 We just need a volunteer.... 12:33:59 hagarth: right, thats what I've done for 3.5.x 12:34:16 hchiramm_: can you help here? 12:34:21 hagarth, sure 12:34:33 hi, sorry to be late 12:34:53 hchiramm_: thanks, JustinClift - one more AI for hchiramm_ ;) 12:34:58 * kshlm says pranithk hi! 12:35:01 #action hchiramm_ to update the backport request page, adding the tracker BZ's for each release and instruction people to add proposed patches to the appropriate tracker BZ's 12:35:08 hagarth, thanks :) 12:35:40 #topic 3.5 12:35:47 Is there anything for this? 12:36:11 JustinClift, missed ur question , are u waiting for me 12:36:15 on me* 12:36:22 lalatenduM: Nah, we got it sorted :) 12:36:32 sorry abt it 12:36:34 k, sounds like nothing new for 3.5 this week. 12:36:36 3.5 -> dht fixes coming this or next week 12:36:37 lalatenduM: Not a killer 12:36:37 we still need dht fixes for 3.5.3 12:36:57 when they land, I'll get a beta out, so maybe next week friday 12:37:03 Cool 12:37:18 okies 12:37:40 k, updated etherpad with that eta. 29th Aug 12:37:47 #topic 3.6 12:37:56 hagarth: Yours? 12:38:01 we had planned a test day this week 12:38:21 however there are issues in dht and afrv2 that need to be addressed 12:38:32 k, so we'll get it organised after they're done 12:38:40 I don't think our test day will be effective without those fixes 12:38:44 "put our best foot forward" 12:38:51 It's all good. 12:38:51 we could test one-brick configs? but well, uhm... 12:38:56 Heh 12:39:09 We have the regression test stuff for that ;) 12:39:12 (kinda) 12:39:23 hagarth: Sounds workable 12:39:33 I will push out the test day by a week 12:39:42 hagarth: Update the schedule page? 12:39:43 pranithk: do you want to provide an update on afrv2 fixes? 12:39:47 JustinClift: will do 12:39:50 :) 12:39:57 did you announce the test day already? I might have missed the emails 12:40:11 ndevos: no, I had just updated the Planning36 page 12:40:23 was waiting for this decision to send out a note on the MLs. 12:40:27 hagarth: ah, okay 12:40:52 Anything else for 3.6? 12:41:04 nothing else from me 12:41:07 I am started working on the specfile 12:41:08 hagarth: I found 2 issues in data-self-heal. 3 in entry-self-heal and testing metadata self-heal today. 12:41:10 for 3.6 12:41:18 one more thing around 3.6 planning 12:41:27 pranithk: Applicable to 3.6? 12:41:44 JustinClift: yes they are afrv2 problems the ones hagarth mentioned a while earlier... 12:41:46 would it be good to start community hangouts that talk about the features? 12:41:50 Cool 12:42:02 lalatenduM: hmm, what needs to be done for 3.6? the nightly builds at least build :) 12:42:03 s/features/new features in 3.6/ 12:42:24 oh, I'd love to see a hangout with pranithk! 12:42:25 ndevos, the fedora spec file is diff than source 12:42:31 hagarth: Hangouts is one way. We could do that. There's also web instructional video type stuff. 12:42:34 lalatenduM: ah, yes, sure 12:42:37 I'm all for whatever works. :) 12:43:00 right, let us have this planned and start these sessions as a build up to 3.6 GA 12:43:05 ndevos: sure why not ;-) 12:43:16 hagarth: We should probably try it out, and see how well it goes. 12:43:24 JustinClift: yeah 12:43:27 ndevos: English could be terrible, don't say I didn't warn you folks :-) 12:43:34 Might take a few to knock off any "rough edges" our people have 12:43:43 pranithk: That's what subtitles are for. :D 12:43:43 pranithk++ 12:43:50 pranithk++ 12:43:56 pranithk: most of us could use that disclaimer before we start :D 12:44:06 JustinClift: spoke like a hollywood producer :-D 12:44:14 Heh Heh Heh 12:44:30 cool, I will start a thread for doing this planning 12:44:42 Cool, sounds like pranithk has been volunteered for the first one too 12:44:49 hagarth: AI it :) 12:44:54 we some VFX studios using Gluster :) 12:45:06 #action hagarth to evolve a plan for 3.6 feature sessions 12:45:18 #action pranithk to write a script for his afrv2 hangout 12:45:27 that's it from me around 3.6 12:45:29 ndevos: script? 12:45:31 Scripts even :) 12:45:44 pranithk: hehe, yes, or at least something :D 12:45:50 ndevos: I thought it is a video of me explaining afrv2 12:45:58 pranithk: script -> what happens in the scene / act / hangout 12:46:06 JustinClift: ah! 12:46:07 eg words, actions, etc 12:46:28 Gah, ksmlm got away without an AI 12:46:33 * JustinClift will do better next time 12:46:33 pranithk: be prepared, as in the scout motto! 12:46:36 Moving on... :) 12:46:41 Aha 12:46:43 JustinClift, lol :) 12:46:48 #action kshlm to do a hangout on glusterd 12:46:53 haha 12:46:55 Awesome :) 12:46:58 hagarth++ 12:47:07 What! 12:47:15 You been volunteered :) 12:47:17 ++ 12:47:25 kshlm: yes, you have been volunteered for better peer identification 12:47:25 :) 12:47:31 Moving on 12:47:36 "Other agenda items" 12:47:51 #topic cppcheck fixes nag 12:47:56 JustinClift: 6th August meeting minutes didn't contain any action item on me related to this 'list' 13th contains it, but kkeithley says he is picking it up from previous meeting. So not sure. Ah! I think it is from bug-triage meeting? 12:48:27 Is it present in the weeks before meeting summary action items? 12:48:37 JustinClift: not in the mails you sent 12:48:55 pranithk: Yeah, maybe from the bug-triage meeting 12:48:56 pranithk, https://public.pad.fsfe.org/p/gluster-community-meetings :) grep there 12:49:05 If you can't find it, we can drop it I guess 12:49:10 hchiramm_: Good idea 12:49:21 :) 12:49:29 We need to get this cppcheck stuff done 12:49:41 kkeithley's nag emails don't seem to work effectively for it 12:49:58 I just sent email to #gluster-dev 12:50:06 JustinClift: I will get to it this week 12:50:10 kkeithley: Do you have known people for each branch that can be assigned to them, instead of requesting? 12:50:25 hagarth: AI it :) 12:50:32 JustinClift: not branch, sub-system 12:50:47 Good piont 12:50:49 point 12:50:50 #action hagarth to review cppcheck fixes this week 12:50:58 hchiramm_: I think we should drop it. It doesn't have what we are talking about either :-( 12:51:02 QUESTION from ndevos: is there a deadline/planning/date we need to meet to get 3.5.x or 3.6.x in Ubuntu? 12:51:10 I guess that smaller patches, one per subsystem is quicker to review/merge 12:51:39 ndevos: Yeah. That way has an easier defined list of ppl to review stuff 12:51:53 But, hagarth said he'd get this done, so it's probably moot point for now 12:51:59 ndevos: +1 12:52:10 JustinClift: it would help for next patches :) 12:52:15 Yep 12:52:16 JustinClift: future 12:52:23 pranithk: +1 12:52:26 Oh, we should remove me from the Glupy maintainers list 12:52:30 * JustinClift will sent patch 12:52:33 send 12:52:40 * ndevos wont review that one :P 12:52:50 #action JustinClift to remove himself/Glupy from maintainers list 12:53:03 * pranithk -2 Do not submit :-) 12:53:06 I haven't touched the code in ages, so useless to help ppl with it :( 12:53:25 #topic Backups for GlusterFS Infrastructure 12:53:30 JustinClift: it will become orphan Justing 12:53:36 Yeah, I know 12:53:54 No-one else seems to have any interest in it at all. Non-optimal, but it is what it is. 12:54:02 JustinClift: that backup stuff, some interns did it I thought... hagarth? 12:54:14 pranithk: this is different from NDMP 12:54:19 With the backup stuff, it's being actively worked out on #gluster-infra 12:54:33 hagarth: ah! got t 12:54:38 pranithk: There's currently no backups being done of any of the GlusterFS infrastructure 12:54:39 JustinClift: we need to collate all our infra details as hchiramm_ mentions 12:54:44 Which is scaery in all kinds of ways :) 12:54:49 JustinClift: oh, there is a #gluster-infra channel? 12:54:52 So, we're getting that fixed 12:55:01 JustinClift: got it 12:55:08 ndevos: Oops, I meant gluster-infra mailing list 12:55:14 ah, okay 12:55:26 #topic Website maintenance strategy - Eco's post on gluster-infra 12:55:31 http://www.gluster.org/pipermail/gluster-infra/2014-August/000299.html 12:55:37 * JustinClift is going to try the steps out today 12:55:44 please let us know your comments on this workflow 12:55:50 Definitely 12:55:53 "only 23 steps" 12:56:08 From memory there's some kind of warning/error that happens with Middleman 12:56:17 But it still builds the site ok I'm told 12:56:31 we still need to evovle more details there 12:56:33 When we get that error fixed, then we should (hopefully) be able to automate the whole thing 12:56:36 Yeah 12:56:40 and reduce the overall number of steps 12:56:45 That's what I'm going through it today for. 12:56:49 JustinClift: cool 12:56:54 * JustinClift *really* wants it fully automated 12:56:57 and remove the needed ssh access to servers :) 12:57:04 ;) 12:57:42 #topic owncloud/GlusterFS talk on Friday in Linuxcon - http://lccona14.sched.org/event/d74123964f922f5fe84a28066ca11633#.U_SMIGeeHQZ 12:57:59 We should probably send details of that out via mailing lists / blog 12:58:00 nvm, there's a talk on glusterfs & owncloud in LinuxCon NA this friday 12:58:12 ... or not then? 12:58:35 please spread the word around this in all your favorite social networking sites :) 12:58:39 :) 12:59:05 a mail on gluster-users could help too 12:59:07 #action JustinClift to blog about upcoming owncloud/GlusterFS talk happening on Friday. And email gluster-* abt it too 12:59:20 #topic Bug triage meetings - do we need bi-weekly bug triage meetings in #gluster-meeting ? 12:59:37 Who's item is that? 12:59:50 mine 13:00:08 we started off with a bug triage meeting earlier to define the process 13:00:14 I guess it would be good to do that bi-weekly 13:00:39 right 13:00:43 Note - We've just hit meeting end time. But we're nearly through the items so lets get them finished 13:01:01 plz review http://www.gluster.org/community/documentation/index.php/Bug_triage 13:01:03 ndevos: would you want to setup those meetings? 13:01:05 hagarth: Discuss on gluster-devel and +1/-1 there? 13:01:20 if you have not reviewed it before 13:01:22 hagarth: yeah, I think I can - any idea who would want to join? 13:01:41 ndevos: we need the sub-maintainers and anybody else who's interested 13:01:58 ndevos: we can just drop a note to gluster-devel about the meeting series 13:01:58 hagarth: okay, sounds good to me 13:02:19 AI it 13:02:25 #action ndevos to work on setting up periodic bug triage meetings 13:02:26 done 13:02:34 #topic Gluster related talk at the File and Storage Systems track / Linux Plumbers Conference? 13:02:49 yes, I just added that 13:03:02 fs-cache for fuse sounds like a good proposal 13:03:19 any other ideas? 13:03:27 This is a "we need ideas" for the talk? 13:03:32 yes 13:03:50 or, rather, "what talk would you like to see?" 13:03:53 Something to do with clustering and striping? 13:03:55 LPC is for slightly futuristic stuff - we don't need a lot of involved implementation details 13:04:32 Our striping implementation isn't great speed wise. Maybe discussion of ways to address that, if appropriate to LPC 13:04:45 what abt replication vs EC in gluster? 13:04:51 s/striping/sharding/ is going to be our answer :) 13:05:20 * JustinClift would argue with that 13:05:38 lalatenduM: possible, if we can have a novel idea around replication or EC 13:05:44 Yeah 13:05:59 ndevos: Also ask on gluster-users? 13:06:02 hagarth, also the performance implications 13:06:19 ndevos: you can probably reply to Ric's email on gluster-users and see if we get more ideas 13:06:29 JustinClift: yes, I plan to send an email about it after Vimal confirmed he'll be working on it :) 13:06:29 k, that's the last item for today's agenda 13:06:35 Re: latest samba, Jose is building samba-4.1.11 for download.gluster.org now. you can mark that as done 13:06:38 ndevos: Cool :) 13:06:49 hagarth: thats the thought, yes 13:07:05