11:59:43 <kshlm> #startmeeting Weekly community meeting 15-Jun-2016
11:59:43 <zodbot> Meeting started Wed Jun 15 11:59:43 2016 UTC.  The chair is kshlm. Information about MeetBot at http://wiki.debian.org/MeetBot.
11:59:43 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
11:59:43 <zodbot> The meeting name has been set to 'weekly_community_meeting_15-jun-2016'
11:59:48 <kshlm> #topic Rollcall
11:59:53 <kshlm> Hi all!
12:00:15 * kkeithley is here
12:00:16 <post-factum> hello o/
12:00:42 * jiffin is here
12:01:27 * ndevos _o/
12:01:30 <kshlm> Any more?
12:02:59 <kshlm> Where are the others today?
12:03:20 <post-factum> sleeping?
12:03:32 * ira_ wakes up a bit.
12:03:41 <post-factum> good morning!
12:03:45 <ndevos> no idea whats up...
12:04:07 <ira_> ^
12:04:10 <ira_> |
12:04:13 <ira_> That's up.
12:04:40 <kshlm> Anyways, let's continue.
12:04:47 <kshlm> Maybe more people join in later.
12:04:47 <ndevos> did you know that I have one of my 2 monitors vertical?
12:05:04 <post-factum> cat flipped it?
12:05:24 <ira_> Many at RH do ;)
12:05:29 <kshlm> I'll be the next week's host, so the first topic is done.
12:05:36 * rastar_ is here, late
12:05:43 <ndevos> no, it is intentional, but my desktop sometimes fails to pick it up, so "up" may point to the right in that case
12:05:51 <kshlm> Vertical monitors are too tall for me.
12:06:05 <kshlm> #topic GlusterFS-3.9
12:06:30 <kkeithley> who is going to manage 3.9?
12:06:38 <kshlm> We don't have enough devs working on 4.0 so let's get started with 3.9.
12:06:48 <kshlm> kkeithley, That's the question.
12:06:59 <ndevos> spin the bottle and assign someone?
12:07:01 * atinm is here too
12:07:09 * kotreshhr is here
12:07:09 <kkeithley> who wants to manage it?
12:07:20 * aravindavk is here
12:07:21 <ndevos> you?
12:08:01 <ndevos> maybe someone who has spare time? ah, that's YOU gain!
12:08:05 <ndevos> +a
12:09:12 <kshlm> Seems to me that only people who have spare time show up to these meetings.
12:09:19 <kshlm> Everyong else is very busy.
12:09:36 <post-factum> kshlm: including you ;)?
12:09:41 <atinm> then I should log out :D
12:09:41 <ndevos> lol, "spare time"? this is part of every Gluster developers job description
12:09:59 <ira_> kshlm: Huh?  I think you just insulted the whole channel.  gg.
12:10:07 <post-factum> :(
12:10:29 <kshlm> A /s was intended
12:10:31 <ira_> We MAKE time to come to the meeting.
12:10:35 <ira_> :)
12:10:37 <kshlm> :D
12:10:46 <jiffin> ndevos: already created the roadmap for 3.9(using backlog features in 3.8) https://www.gluster.org/community/roadmap/3.9/
12:11:05 <kshlm> I would volunteer to manage 3.9, but I want to streamline stuff I'm doing.
12:11:07 <ndevos> yeah, thats the features that did not make it in 3.8
12:11:10 <kshlm> So I'm passing this time.
12:11:28 <ndevos> kshlm: oh, so we can put you down for 3.9+1 ?
12:11:49 <kshlm> ndevos, Possibly.
12:11:58 <kkeithley> anyway, if not this meeting, then the next one for sure.  We need to identify the release manager and they need to start working on it. Maybe at 10% level of effort.  Three months is very short
12:12:13 <ndevos> I'm happy to assist with 3.9, but I do not have enough time to run it by myself
12:13:07 <kshlm> kkeithley, I agree with you.
12:13:15 <kshlm> Should we take this up next meeting?
12:13:30 <kshlm> Or earlier, on the mailing lists?
12:13:55 <ndevos> I guess the maintainers list should be a good start
12:14:20 <rastar_> yes, mailing list please.
12:14:33 <kshlm> Mailing list it is then.
12:14:33 <ndevos> kshlm: I can send the email there, with some of the tasks that should be handled soon
12:14:43 <kshlm> ndevos, Thank you!
12:14:46 <kshlm> ndevos++
12:14:47 <glusterbot> kshlm: ndevos's karma is now 13
12:15:18 <kshlm> #action ndevos will call for 3.9 release-maintainers on the maintainers list
12:15:27 <kshlm> Okay.
12:15:32 <kshlm> Onto the big topic of the day.
12:15:40 <kshlm> #topic GlusterFS-3.8
12:15:57 <post-factum> tagged!
12:16:11 <ndevos> and packages are being built
12:16:35 <kshlm> Hurray!
12:16:36 <ndevos> there will be a blog post and email to the announce list later today (by amye)
12:17:07 * samikshan joins the meeting too
12:17:43 <ndevos> I guess the next update for 3.8.1 will be around the 10th of next month
12:17:55 <ndevos> it'll take the now EOL 3.5 release slot
12:18:56 <ndevos> for 3.8.1 we'll stick to the backport criteria that are getting put together, no risky backports, and definitely no features
12:19:15 <ndevos> is there more to 3.8? or any questions?
12:19:27 <kshlm> ndevos++ jiffin++ Thanks for a well managed release
12:19:48 <kshlm> This is probably the first release which has had a structred release process.
12:20:01 <atinm> ndevos++
12:20:01 <kshlm> More structured than our previous releases.
12:20:01 <glusterbot> atinm: ndevos's karma is now 14
12:20:06 <atinm> jiffin++
12:20:07 <glusterbot> atinm: jiffin's karma is now 3
12:20:18 <ndevos> the plan is to keep that up, and have a similar structure/process for the next releases
12:20:24 <kshlm> Huh! Why didn't you get my karma?
12:20:27 <kshlm> ndevos++
12:20:28 <glusterbot> kshlm: ndevos's karma is now 15
12:20:34 <kshlm> jiffin++
12:20:35 <glusterbot> kshlm: jiffin's karma is now 4
12:20:52 <ndevos> if everyone knows what to expect of a release, everyone should be able to get their features planned accordingly too
12:21:18 <ndevos> (and companies that productize Gluster might benefit from that again)
12:21:38 <post-factum> release early, release often
12:21:41 <kshlm> ndevos, Yup. I'd said on a mail thread that I'd start a document to capture the process.
12:22:09 <kshlm> I've written a basic document that I'll be sharing later to{dat/night}
12:22:19 <ndevos> kshlm: yes, lets start with an etherpad somewhere and push it to the glusterdocs repo shortly after
12:22:56 <kshlm> I don't have anymore questions for 3.8
12:23:04 * ndevos neither
12:23:09 <kshlm> I'd like to congratulate everyone who contributed to 3.8
12:23:17 <post-factum> :)
12:23:37 <kshlm> Thanks again.
12:23:52 <ndevos> I hope the blog post contains some stats too, otherwise we might want to do a post about contributions and such later
12:24:57 <kshlm> We have a script in contrib that does that IIRC.
12:25:08 <ndevos> yes, this: ./extras/who-wrote-glusterfs/who-wrote-glusterfs.sh v3.7.0alpha0~1...v3.8.0
12:26:02 <kkeithley> shall I change the top-level LATEST link on download.g.o? Are we ready to do that?
12:26:15 <kshlm> kkeithley, Right now?
12:26:35 <kshlm> +1 from me.
12:26:45 <ndevos> output with the stats is on http://termbin.com/0s3o for the people following along
12:26:47 <kkeithley> well, now, or in an hour? Or in six hours?
12:27:09 <kshlm> +1 for right now.
12:27:11 <kkeithley> all packages that we're going to have on d.g.o are there now
12:27:19 <kshlm> Should we waiting for anything else?
12:27:35 <ndevos> I dont know when the CentOS Storage SIG packages will be pushed to the mirrors, I guess they are in the testing repository already
12:28:30 <post-factum> ndevos: any link?
12:28:32 <kkeithley> Ubuntu packages for trusty and wily are in the Launchpad PPA. xenial coming soon.
12:28:42 <kkeithley> SuSE later today probably
12:28:57 <kshlm> post-factum, https://wiki.centos.org/SpecialInterestGroup/Storage
12:29:04 <ndevos> post-factum: http://buildlogs.centos.org/centos/7/storage/x86_64/gluster-3.8/ and the like
12:29:09 <kshlm> Should have most of the info you need.
12:29:33 <post-factum> ndevos: thanks. the latter one is useful
12:30:16 <ndevos> post-factum: there is a centos-release-gluster38 package around somewhere too, I need to get that pushed into CentOS Extras when the mirrors contain the 3.8 packages
12:30:35 <post-factum> ndevos: in fact the only thing i need is src.rpm :)
12:31:06 <ndevos> post-factum: I really hope there is no need to rebuild the packages...
12:31:20 <post-factum> ndevos: not all mem-leak related patches were merged
12:31:33 <ndevos> ah, right
12:31:38 <post-factum> ndevos: and not all leaks are hunted. see ML
12:31:40 <ndevos> kshlm: move on?
12:31:59 <kshlm> Sure.
12:32:03 <ndevos> post-factum: yeah, I'm trying to catch up on that too
12:32:15 <kshlm> #topic GlusterFS-3.7
12:32:20 <kkeithley> okay, top-level LATEST link now points to 3.8/3.8.0
12:32:26 <kshlm> kkeithley++
12:32:27 <glusterbot> kshlm: kkeithley's karma is now 7
12:32:28 <post-factum> ndevos: good. let me know if you need any additional info
12:32:39 <kshlm> So, 3.7.12 isn't released yet.
12:32:48 <post-factum> and hagarth isn't here...
12:32:53 <kshlm> Yup.
12:33:13 <kshlm> And not all of the maintainers have given their nods yet.
12:33:14 <atinm> because hagarth is still waiting to hear from the maintainers on the sanity of each components
12:33:41 <kshlm> atinm++ thanks for the heads up for glusterd
12:34:11 <kshlm> I think it's because the request wasn't clear enough.
12:34:35 <ndevos> it surely needs to be spelled out, and included in the release process in some form
12:34:37 <kshlm> The request was done in the thread about changing release-process.
12:34:57 <kshlm> I didn't find it at first.
12:35:28 <kshlm> There should be a seperate thread for this.
12:35:47 <ndevos> I also did not wait for all maintainers for ACK 3.8.0, there was a chance for them to test different RC and nightly builds, no reply counted as "I'm ok with releaseing this"
12:35:49 <kshlm> I'm willling to start this thread, if there are no objections.
12:36:30 <ndevos> yes, please do, and hopefully maintainers will provide automated test cases that we can run nightly in the CentOS CI
12:36:44 <kshlm> Okay. I'll do that after the meeting.
12:37:08 <kshlm> #action kshlm to start seperate thread for maintainer feedback on 3.7.12rc
12:37:22 <kshlm> That should be it for 3.7
12:37:28 <kshlm> #topci GlusterFS-3.6
12:37:34 <kshlm> #topic GlusterFS-3.6
12:37:51 <kshlm> Still no 3.6.10
12:38:04 <post-factum> lets mark 3.6 EOL too :)
12:38:37 <kshlm> Still seem to be a lot of people using 3.6.
12:38:51 <kshlm> post-factum, that would be done with the release of 3.9
12:39:00 <kshlm> ..in sept if all goes to plan.
12:39:18 * kshlm thinks that would be the case
12:39:27 <ndevos> yes, I think so too
12:39:49 <kshlm> I had a month old AI to ask a user if they still required the mem-leak fixes
12:39:59 <kshlm> I've not done that yet.
12:40:10 <jdarcy> There might still be more 3.6 users than 3.7.  We need to figure out how to help them move forward before 3.6 really does go EOL (which it already is with an absentee maintainer and no other devs available/interested in updating it).
12:40:10 <ndevos> shouldnt that be reported in Bugzilla?
12:40:38 <kshlm> Also, in last weeks meeting it was noted that the required changes might not be straight forward backports.
12:40:58 <jdarcy> "*effectively* already is"
12:41:12 <kshlm> jdarcy, Do we have any idea how many users are on which version of GlusterFS?
12:41:34 <jdarcy> kshlm: I don't think we have hard numbers.  The last survey is probably outdated by now.
12:41:56 <jdarcy> kshlm: Just going by what I see elsewhere, and past patterns.
12:42:45 <kshlm> The impression I had from FOSDEM (~4months ago ) was that we had more users on the latest version
12:43:07 <kshlm> When was the last survey held?
12:43:40 <jdarcy> kshlm: A lot of the 3.6 usage is via downstream.
12:44:12 <kshlm> jdarcy, Wasn't it skipped? We went from 3.5 to 3.7 downstream.
12:44:28 <jdarcy> kshlm: Oh, OK, then it's even worse.
12:44:37 <jdarcy> Then again, let downstream worry about downstream.
12:44:40 <kkeithley> I don't know if it's wise to read too much into d.g.o downloads, but 3.7.11 downloads are 4x the 3.6.9 downloads.
12:46:00 <kkeithley> Ubuntu Launchpad PPA downloads are about evenly divided between 3.7.11 and 3.6.9
12:46:26 <jdarcy> The main thing is that there are still significant numbers of users on a release that's practically but not officially EOL.
12:46:27 <kshlm> ndevos, Any stats from the storage-sig?
12:46:31 <post-factum> shame for ppa. those feel like fresh installs of 3.6 :(
12:47:25 <kkeithley> http://download.gluster.org/pub/gluster/glusterfs/download-stats.html
12:47:49 <ndevos> kshlm: no, I have no idea if the stats are available... Pretty sure there are none for Fedora either
12:48:16 <kkeithley> <cough>mirrors</cough>
12:48:31 <kshlm> Huh, we have more downloads for 3.5.9 than 3.6.9
12:48:55 <jdarcy> We also don't know how many installs correspond to each download.  Some orgs might download once, install literally thousands.
12:49:04 <post-factum> users should be stimulated somehow
12:49:37 <jdarcy> Or build from source, because they have thousands of lines of their own patches.
12:49:42 <kshlm> we could add a version-checker that phones home and let's the user know a newer version is available.
12:49:54 <kshlm> That would be controversial.
12:49:56 <kkeithley> well, f23 only has 3.7.x, and f22 only has 3.6.x.  If anyone really using Fedora in production they have to go to d.g.o for something other than what's available from yum/dnf.
12:49:59 <jdarcy> kshlm: Very.
12:50:20 <kshlm> Ah okay.
12:50:41 <kshlm> We've not arrived at any conclusion about this, and we're running out of time.
12:50:43 <ndevos> kshlm: yeah, I doubt many users appreciate such a "feature"
12:50:49 <jdarcy> kshlm: Indeed.
12:51:08 <kshlm> We'll take this up on the mailing lists as well.
12:51:12 <kkeithley> we could turn on ABRT reporting in Fedora. Then we'd know how many had crashes
12:51:30 <kkeithley> (we get ABRT reports for nfs-ganesha)
12:51:33 <kshlm> #action Start a mailing list discussion on EOLing 3.6
12:51:43 <kshlm> kkeithley, What would that take?
12:52:00 <kshlm> Changes in our source?
12:52:03 <kkeithley> probably just a setting in Bodhi
12:52:05 <ndevos> kkeithley: that should be turned on by default? if it is not, we need to fix it
12:52:49 <ndevos> but, abrt only reports segmentation faults and stack-traces, maybe Gluster is just better than Ganesha ;-)
12:53:12 <kkeithley> haha
12:53:19 <post-factum> we could also implement google analytics xlator
12:53:40 <ndevos> nice idea!
12:53:52 <kshlm> That's enough side-track
12:53:57 <kshlm> #topic GlusterFS-3.5
12:54:03 <kshlm> This is now EOLed.
12:54:17 <ndevos> I'll commit a change to the repo, mentioning that
12:54:18 <kshlm> Should an announcement be made seperately for this?
12:54:24 <post-factum> yes
12:54:26 <ndevos> and close all the left over 3.5 bugs with a note
12:54:35 <kshlm> Or would amye's annoucement cover this?
12:54:50 <ndevos> I expect it to be mentioned there as well
12:55:13 <kshlm> Okay. Good then.
12:55:20 <ndevos> in any case, if we do an announcement, it should be done after the 3.8 communication
12:55:24 <jdarcy> https://en.wikipedia.org/wiki/File:Taps_on_bugle.ogg
12:55:37 <kshlm> Thanks for managing the 3.5 releases ndevos!
12:55:49 <ndevos> np, time to move on!
12:56:23 <kshlm> #topic NFS-Ganesha
12:56:46 <ndevos> kkeithley: you're up!
12:57:02 <kkeithley> nothing to report
12:57:05 <ndevos> or skoduri or jif<tab>
12:57:44 <skoduri> yes nothing new.. stabilizing md-cache layer
12:57:57 <kkeithley> next big features are queued behind things like upcall, and compound fops
12:58:04 <kkeithley> in gluster
12:58:20 <kshlm> Did any of those land in 3.8?
12:58:33 <kshlm> There was a list last week from poornimag.
12:59:13 <kshlm> #topic Samba
12:59:20 <kshlm> We're nearly outta time.
12:59:34 <kshlm> rastar_, obnox, around to provide any updates?
12:59:36 <rastar_> Nothing to add for this week.
12:59:43 <kshlm> Thanks.
13:00:04 <kshlm> #topic Last weeks AIs
13:00:31 <kshlm> I've got 2 of the 4 AIs from last week, and I know they're not done.
13:00:39 <kshlm> rastar_, Has one of the other 2.
13:00:49 <kshlm> #topic rastar to look at 3.6 builds failures on BSD
13:00:53 <rastar_> I have prepared the skip code for that
13:01:02 <rastar_> Yet to update Jenkins
13:01:11 <kshlm> rastar_, Was this the 3.6 build failures in glusterd?
13:01:22 <rastar_> kshlm: I think yes
13:01:30 <kshlm> Should I carry this AI over to next week?
13:01:45 <rastar_> but it was decided to skip build checks for 3.6 on *BSD
13:01:48 <rastar_> yes
13:02:03 <kshlm> Cool.
13:02:55 <kshlm> I have a question about my AI to check if mem-leak fixes are required? Should I still do it?
13:03:11 <kshlm> mem-leak fixes for 3.6 that is.
13:03:31 <ndevos> kshlm: if nobody complains or files bugs, I do not think we need to be very active with 3.6
13:03:59 <kshlm> ndevos, that AI came up because of a user list thread.
13:04:11 <kshlm> But no bug has been filed officially yet.
13:04:30 <kshlm> So I guess I'll just scratch this AI off for now.
13:04:42 <ndevos> kshlm: ask them to file a bug if they really care and can not move to 3.7? Keep it simple and short ;-)
13:04:56 <kshlm> ndevos, Okay. I'm gonna do that.
13:05:03 <ndevos> or, get them to move to 3.8, of course!
13:05:03 <kshlm> #topic Open floor
13:05:33 <kshlm> We have a couple of topics, that seems to have been added after the bug-triage yesterday.
13:05:45 <kshlm> I guess kkeithley added them.
13:05:59 <kshlm> Any context on these?
13:06:17 <kkeithley> I think they're self explanatory
13:06:36 <kshlm> Seems like they're announcements/guidelines, more than a discussion topic.
13:06:52 <kkeithley> a reminder.
13:06:52 <kshlm> Did you intend to discuss them now?
13:06:56 <kkeithley> not per se
13:07:03 <kshlm> Okay.
13:07:44 <ndevos> ... oh the tension for the readers that do not have the etherpad open
13:07:48 <kshlm> #info Bug self triage. When you open a bug for yourself, assign it (to yourself) and add the keyword "Triaged"
13:07:48 <kshlm> #info If it's not for yourself, but you know who it does belong to, assign it to them and add the keyword "Triaged"
13:07:48 <kshlm> #info If you submit a patch for a bug, set the bug state to POST.
13:07:48 <kshlm> #info If your patch gets commited/merged, and the commiter forgets, set the bug state to MODIFIED
13:08:14 <kshlm> This would be better in someplace in glusterdocs.
13:08:38 <ndevos> I think we have that documented :)
13:08:50 <kshlm> The second topic, kkeithley is that a discussion as well?
13:08:59 <kshlm> 'kkeithley Saravanakmr will set up Coverity, clang, etc on public  facing machine and run it regularly'
13:09:01 <kkeithley> no, not really
13:09:02 <ndevos> #link http://gluster.readthedocs.io/en/latest/Contributors-Guide/Bug-Triage/
13:09:13 <kkeithley> just an AI for misc and nigelb
13:09:25 <ndevos> #link http://gluster.readthedocs.io/en/latest/Contributors-Guide/Bug-report-Life-Cycle/
13:09:49 <kshlm> For misc/nigelb or for Saravanakmr?
13:10:21 <kshlm> I'll add an AI in any case.
13:10:24 <nigelb> just me
13:10:26 <nigelb> for that one.
13:10:37 <kshlm> #action kkeithley Saravanakmr with nigelb will set up Coverity, clang, etc on public  facing machine and run it regularly
13:10:55 <kshlm> nigelb, You have an AI now.
13:11:07 <kshlm> That's all of the topics we had today.
13:11:18 <kshlm> If we have nothing more to discuss I'll end the meeting.
13:11:24 <kshlm> ..in 1 minute.
13:11:31 <post-factum> ok thx bye :)
13:11:41 <kshlm> Bye post-factum
13:12:38 <kshlm> Okay.
13:12:46 <kshlm> Thanks for attending the meeting everyone
13:12:49 <kshlm> #endmeeting