12:03:44 <davemc> #startmeeting
12:03:44 <zodbot> Meeting started Wed Nov 12 12:03:44 2014 UTC.  The chair is davemc. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:03:44 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
12:03:54 <davemc> Roll call
12:04:00 <davemc> who's here
12:04:08 * davemc is here
12:04:20 * JustinClift waves
12:04:27 * hagarth is here
12:04:29 * kkeithley is here
12:04:31 * kshlm is here
12:04:42 <JustinClift> https://public.pad.fsfe.org/p/gluster-community-meetings
12:04:44 <davemc> etherpad is at https://public.pad.fsfe.org/p/gluster-community-meetings
12:04:52 * lalatenduM is here
12:05:07 * Humble is here
12:05:08 * ndevos is here, but also in a confcall
12:05:10 <davemc> if you would, please add yourself to the IRC - IRL translator on the page this week
12:05:22 <davemc> line 27
12:05:31 <hagarth> sounds like a cool translator to implement in glusterfs
12:05:57 <davemc> actions from last meeting, if I scraped correctly
12:06:04 <hagarth> getfattr -n "user.IRL" /mnt/glusterfs/<IRC-nick> could be the interface :)
12:06:22 <davemc> #item hagarth to update maintainers file
12:06:35 <hagarth> davemc: still TBD, intend doing this week
12:06:42 <davemc> k
12:07:06 <davemc> #action  hagarth to update maintainers file
12:07:33 <davemc> #item ndevos to send email / blog on RHEl 6.6
12:07:35 <davemc> this is done
12:07:39 <davemc> I think
12:07:42 <lalatenduM> yes
12:07:44 <hagarth> yes, that is done
12:07:49 <lalatenduM> ndevos++
12:08:00 <davemc> #item hagarth to send an email about release maintainer for 3.6
12:08:11 <hagarth> done, sent out an email few minutes back
12:08:22 <davemc> cool
12:08:35 <davemc> #item Humble create a maintainer wiki page
12:08:42 <Humble> I am collecting information on components and maintainers.. need some more time to finish this task
12:08:53 <davemc> no problem
12:09:02 <davemc> #action Humble create a maintainer wiki page
12:09:17 <davemc> all I have from last meeting
12:09:22 <davemc> ready to move on?
12:09:36 <hagarth> yes
12:09:44 <davemc> #item 3.6
12:10:00 <hagarth> 3.6.1 was released last week to overcome the packaging problems
12:10:14 <davemc> did we announce it somewhere?
12:10:14 <hagarth> we will have a bug tracker for 3.6.2 soon
12:10:27 <hagarth> davemc: yes, announced on users & devel
12:10:28 <Humble> yep ..  indeed a good solution to couple of problems!
12:10:43 <JustinClift> Blog/twitter/etc?
12:10:51 <davemc> announce list?
12:11:07 <hagarth> davemc: nope, we did not hit announce.
12:11:11 <JustinClift> Hmmm, I wonder if we have a release checklist page
12:11:20 <JustinClift> If not, that's probably worth creating
12:11:24 <hagarth> JustinClift: no blog/twitter too
12:11:27 <JustinClift> Help us to not miss things
12:11:33 * overclk is here
12:11:41 <davemc> we _need_ release checklist
12:11:52 <hagarth> JustinClift: +1, can you take a stab at that?
12:12:29 <JustinClift> #action JustinClift to create initial GlusterFS Release Checklist page on the wiki
12:12:37 <davemc> tks
12:12:47 * JustinClift coulda sworn we already have one started somewhere
12:12:53 <JustinClift> Will look later :)
12:13:08 <davemc> probably. one of my major peeves is trying to find what has already been done
12:13:23 <JustinClift> On that note... davemc we need a link to the wiki from the www.gluster.org site
12:13:25 <hagarth> I propose that we have a "release often" strategy for 3.6.x till the branch becomes very stable
12:13:32 <davemc> #item  ndevos,     lalatenduM, kkeithley and Humble to propose a long term solution to     the packaging problem in CentOS/RHEL
12:13:34 <JustinClift> +1
12:13:40 <JustinClift> +1 to release often
12:13:49 <JustinClift> If we can automate large chunks of that, the better
12:13:51 <davemc> JustinClift, meeting f2f with tigert and DV to discuss web site in total
12:14:02 <hagarth> JustinClift: yeah
12:14:34 <davemc> #idea institute an "early and often" till 3.6 line is stable
12:15:00 <davemc> okay
12:15:05 <davemc> #item  ndevos,     lalatenduM, kkeithley and Humble to propose a long term solution to     the packaging problem in CentOS/RHEL
12:15:24 <lalatenduM> davemc, still in works
12:15:29 <JustinClift> davemc: Soumya Deb has input on web stuff too
12:15:30 <kkeithley> there have been some preliminary conversations, but nothing definite yet
12:15:33 <davemc> k, will leave live
12:15:40 <lalatenduM> davemc, we should move this AI for next week
12:15:43 <davemc> #action  #item  ndevos,     lalatenduM, kkeithley and Humble to propose a long term solution to     the packaging problem in CentOS/RHEL
12:15:44 <JustinClift> davemc: See if you can get him to video in while you guys do so
12:16:10 <davemc> jcastillo, will try. time not set yet
12:16:26 <davemc> JustinClift, will try, time note set yet
12:16:27 <hagarth> can anybody provide inputs to Soumya Deb on gluster-infra?
12:16:28 <JustinClift> davemc: In the meantime, can you please just add a link to the wiki to the website
12:16:43 <JustinClift> That'll help while waiting for the "well though out" solution. ;)
12:16:51 <davemc> JustinClift, will do. I just ht the blofg point and get there
12:16:51 <hagarth> about the website revamp .. he is eager to help us.
12:17:08 <davemc> #action davemc to add pointer to wiki from site
12:17:09 <JustinClift> hagarth: I keep on meaning to, but haven't gotten around to it yet
12:17:31 <hagarth> JustinClift: okay, I think we can rope him in for any further discussions on the website
12:17:38 <JustinClift> #action JustinClift to respond to Soumya Deb on gluster-infra about website
12:18:04 * JustinClift goes for more coffee
12:18:16 <davemc> I think the web site may end up with its own category on meetings for a while
12:18:33 <davemc> any other 3.6 items?
12:18:38 <hagarth> davemc: +1 to that
12:18:51 <hagarth> nothing more from me on 3.6
12:18:53 <pranithk> davemc: yes
12:19:00 <davemc> as noted, approximately 300 3.6.0 downloads through 9-nov
12:19:12 <davemc> your floor pranithk
12:19:14 <pranithk> davemc: I just found that http://review.gluster.com/#/c/8201/ is not merged. SO ext4 is useless with gluster :'-(
12:19:32 <pranithk> davemc: I screwed up :'-(
12:19:49 <davemc> as a former xfs guy, i could make a snide remark
12:19:50 <pranithk> davemc: I was under the impression that it was already merged
12:19:55 <hagarth> pranithk: let us target this for 3.6.2
12:20:09 <davemc> "early nd often" mantra
12:20:41 <hagarth> pranithk: can you own sending a backport of this to release-3.6?
12:21:05 <davemc> that would help
12:21:22 <pranithk> hagarth: I just asked soumya to re-submit it with a bug-id.
12:21:36 <pranithk> hagarth: yes she will do backporting as well. she just agreed to it
12:21:45 <hagarth> pranithk: cool, thanks
12:22:21 <davemc> any more 3.6?
12:22:32 <ndevos> when do we have 3.6.2?
12:22:45 <hagarth> ndevos: before raghavendra bhat comes back ;)
12:23:05 <kkeithley> _not_ before?
12:23:10 <kkeithley> or before?
12:23:16 <hagarth> (in the next week or two )
12:23:21 <ndevos> yes, before, otherwise it won't be early+often
12:24:41 <hagarth> I will send out a note on 3.6.2 tracker later in the day .. feel free to mark dependencies on this bug
12:24:58 <JustinClift> Fibre guy is here doing his thing
12:25:02 <davemc> reminder for the late arrivals.  If you could go fill out the IRC - IRL translator section of the etherpad, it would be nice
12:25:08 <ndevos> hagarth: and please set the Alias for that bug :)
12:25:42 <davemc> okay, more 3.6?
12:25:43 <hagarth> ndevos: sure
12:26:00 <hagarth> davemc: let us move on
12:26:01 <davemc> #item 3.5.3
12:26:22 <ndevos> thats ready, nobody complained about the beta
12:26:23 <davemc> updates. noted some activity in the pad
12:26:38 <ndevos> but, we could include the symbol versioning, if we really want to
12:26:54 <ndevos> that would make material for a 3rd beta though...
12:27:02 <kkeithley> include symbol versions, with or without a beta3?
12:27:21 <lalatenduM> beta3 +1
12:27:36 <hagarth> ndevos: I think we can do without a beta3
12:27:44 <ndevos> in general, additional patches would require a new beta
12:27:53 <hagarth> as the same patch has been tested on release-3.6?
12:28:21 <kkeithley> either way, the 3.5 and 3.4 patches need to be reviewed
12:28:22 <ndevos> yes, still, I do not like to make exceptions - I'm not merging Praniths patches either
12:28:31 <pranithk> ndevos: yes release it. I also want to do some work for afr backward compatibility that needs to be marged in 3.5.x
12:29:00 <davemc> so release and plan a 3.5.4 upcoming?
12:29:02 <ndevos> we don have symbol versioning in 3.5.2, so I do not see a strong reason to require it now
12:29:12 <ndevos> s/don/dont/
12:29:15 <hagarth> ndevos: ok, let us move it to 3.5.4 then
12:29:39 <ndevos> kkeithley: works for you too?
12:29:40 <davemc> so, are we declaring 3.5.3 ready as of today?
12:29:43 <kkeithley> it does
12:29:48 <ndevos> okay
12:30:04 <ndevos> davemc: not ready, it needs updated release notes, but thats it
12:30:19 <davemc> k
12:30:44 <ndevos> so, later today, with Humble or lalatenduM be able to build RPM packages tomorrow?
12:31:14 <ndevos> JustinClift should hurry up with that checklist ;)
12:31:16 <lalatenduM> ndevos, yeah should be ok
12:31:17 <davemc> we can push it, but let's not "announce" it on Friday.  Much grief came my way
12:31:34 <hagarth> davemc: :)
12:31:38 <Humble> :)
12:31:45 <ndevos> tomorrow is thursday, thats ok?
12:31:53 <davemc> sure...
12:31:53 <kkeithley> grief from?
12:32:00 <ndevos> grief why?
12:32:02 <kkeithley> for announcing in Friday
12:32:10 <davemc> PR, mktg types, journalistic types
12:32:10 <kkeithley> s/in/on
12:32:20 <kkeithley> lol
12:32:58 <ndevos> #action ndevos will push release notes for 3.5.3 today, and prepare for release tomorrow
12:33:14 <davemc> more on 3.5 in general?
12:33:27 <kkeithley> pobrecito journalistos
12:33:37 <ndevos> davemc: can we do multiple releases per day? like 3.4 too?
12:34:07 <davemc> well, it's usually frowned on, but as long as one is a strong lead the other weaker it's okay
12:34:11 * kkeithley isn't sure why not
12:34:22 <lalatenduM> no multiple releases plz
12:34:24 * Humble kkeithley :)
12:34:26 <davemc> you collide your own message
12:34:27 <lalatenduM> on one day
12:34:33 <lalatenduM> davemc, +1
12:34:43 <Humble> I think we will select some days in a week for each release
12:34:44 <Humble> :)
12:34:55 <ndevos> I dont see the "collide", it shows a very active project?
12:34:55 <davemc> Humble, +1
12:34:57 <Humble> monday for 3.6 , wed for 3.5 :)
12:35:22 <ndevos> kkeithley: or one of us needs to release today?
12:35:30 <davemc> ndevos, so you announce 3.5.3 and 3.4.6. which one should get covered
12:35:41 <JustinClift> Btw, best day to announce new stuff is apparently Tuesdays
12:35:43 <davemc> by the press/media
12:35:55 <ndevos> davemc: why not both, they are stable releases, and only include bug fixes
12:36:11 <davemc> well, since everyone uses tuesday, it not necessarily as great as people think
12:36:19 <Humble> :)
12:36:25 <ndevos> JustinClift: lol, with a release early/often, we could do a release every 3 weeks?
12:36:29 <JustinClift> davemc: ;)
12:36:49 <JustinClift> With PostgreSQL, for security updates they do one release announcement that covers all the versions
12:36:54 <kkeithley> meh, whatever. I can wait. I already have release notes, I just have to tag and fire off a release in Jenkins. The real issue AFAIC is Humble and lalatenduM's time to build packages
12:37:00 <JustinClift> Can we do one release announcement that covers several versions?
12:37:05 <davemc> ndevos, then he announce,ment is one announce , updates to the currently stable glusterfs lines
12:37:09 <hagarth> JustinClift: +1
12:37:10 <JustinClift> (even tho it's not security related)
12:37:28 <ndevos> davemc: yes, that sounds good to me
12:37:43 <davemc> not sexy, but works
12:37:52 <JustinClift> "GlusterFS releases updated versions: 3.x.x, 3.y.x, etc"
12:37:59 <ndevos> well, a stable release should not be sexy :)
12:38:01 <Humble> yep.. announcement can be in same day
12:38:17 <davemc> as long as it is a single announce point
12:38:57 <davemc> so are we decided to announce both together, and if so when?
12:39:06 <ndevos> davemc: what is for you an announcement? email to the list, a blog post with release notes, or something else?
12:39:14 <JustinClift> If the text is ready to go today, do it today
12:39:28 <JustinClift> ndevos: All of the above
12:39:32 <davemc> ndevos, all
12:39:46 <JustinClift> Mailing lists + Twitter + Blog + other social bits
12:39:49 <davemc> followed by social media channels
12:39:58 <JustinClift> All together = "a release" :)
12:40:20 <davemc> and again, when?
12:40:27 <JustinClift> Today
12:40:27 <davemc> just not friday, please
12:40:36 <kkeithley> today?
12:40:40 <kkeithley> +1 for today
12:40:40 <JustinClift> If it's ready today, do it today
12:40:43 <davemc> works for me
12:40:43 <ndevos> hmm, okay - but I'd like to have the release notes included in those announcements, which makes it tricky for multiple version releases
12:40:55 <hagarth> ndevos: URLs to both would do
12:41:25 <ndevos> hagarth: I got some positive reactions on the full-fledged emails, I think those are appreciated
12:41:26 <JustinClift> Yeah, maybe the highlight summaries in the announce text bit, and URLS to the full release notes for each
12:41:38 <JustinClift> ?
12:42:01 <ndevos> maybe, we can try :)
12:42:01 <kkeithley> release notes for 3.4.6 are already done. Blog post can be copy-paste of the 3.4.6beta2 blog post, minus "beta"
12:42:13 <JustinClift> :)
12:42:21 <hagarth> ndevos: yeah, no harm in trying ;)
12:42:36 <davemc> so the structure can be multifold.
12:42:54 <davemc> a 'release announce' email/blog with highlights.
12:43:08 <ndevos> davemc: will you gather the notes and do the announcement?
12:43:15 <JustinClift> ndevos: Once you try this approach, would you be ok to directly email the ppl you got positive reactions from last time and ask if this highlights+URLS approach is still good for them?
12:43:18 <davemc> a supplemental email/blog with detail as deep as possible
12:43:38 <davemc> ndevos, not today, I'm locked in Storage BU planning all day
12:44:07 <JustinClift> kkeithley has his text ready
12:44:08 <ndevos> davemc: no, tomorrow, 3.5.3 needs a little time and Humble and lalatenduM would probably appreciate it too
12:44:21 <lalatenduM> ndevos, yup :)
12:44:32 <ndevos> well, I got the text, it just needs to get included in the repo, and tag the release - http://blog.nixpanic.net/2014/11/glusterfs-353beta2-release-notes.html
12:44:36 <JustinClift> Ahh
12:44:40 <JustinClift> In this case
12:44:46 <JustinClift> How about we do 3.4.6 today
12:44:49 <JustinClift> And 3.5.3 tomorrow
12:44:57 <JustinClift> Since they're not going to be ready at the same time
12:44:59 <davemc> that would work
12:45:10 <davemc> and the order makes sense as well
12:45:22 <JustinClift> And ndevos can do the full emails again ;)
12:45:27 <ndevos> well, it does depend on lalatenduM and Humble too - otherwise users ask "where are the packages"?
12:45:29 <davemc> and can be the full blown detailed message
12:45:39 <Humble> JustinClift, need some time to prepare packages
12:45:40 * hagarth senses that today is going to be a day of announcements ;)
12:45:58 <kkeithley> we have done "packages will land soon on d.g.o" in the past
12:46:03 <davemc> humble for both 3.4.6 and 3.5.3?
12:46:12 <Humble> yep .
12:46:15 <davemc> k
12:46:19 <JustinClift> kkeithley: We should never do release announcements without packages
12:46:29 <JustinClift> If that means we delay the release announcements, so be it
12:46:43 <davemc> not having packages was the loudest grumble I heard from 3.6.0
12:46:43 <ndevos> yes, and people always ask when the packages are ready...
12:46:44 <kkeithley> Isn't that kinda like the Pirate Code?
12:46:53 <Humble> now a days the tar ball is mailed to package maintiners and once we are ready with packages in d.g.o we go for announcement
12:47:06 <Humble> I think thats the right process to follow
12:47:10 <kkeithley> well, there never were going to be packages for 3.6.0, so I can live with that kinda grumbling
12:47:11 <JustinClift> Yeah
12:47:15 <JustinClift> Lets take this offline
12:47:23 <JustinClift> Well, to the mailing list
12:47:42 <davemc> agreed here
12:47:52 <davemc> 13 minutes left
12:47:55 <JustinClift> Lets get the packages ready to go asap, and co-ordinate the actual release day when we have packages ready to go
12:48:05 <ndevos> subject: tarballs of 3.4.6 and 3.5.3 will be released today, announcement follows tomorrow
12:48:06 <Humble> make sense
12:48:32 <hagarth> awesome!
12:48:35 <JustinClift> "announcement follows when packages are ready" ;)
12:48:37 <davemc> I think we may have cover 3.4.6 here as well
12:48:41 * lalatenduM will leave for a meeting now
12:48:47 <Humble> JustinClift, yep :)
12:48:50 <davemc> bye lalatenduM
12:48:57 <hagarth> lalatenduM: ttyl
12:48:58 <kkeithley> yes, 3.4.6 is ready to go. Just tell me when I should push the button(s)
12:49:07 <davemc> k, kkeithley
12:49:10 <JustinClift> kkeithley: For building packages, now :)
12:49:23 <JustinClift> kkeithley: For release announcement, after the packages are ready
12:49:50 <kkeithley> acknowleged
12:49:57 <davemc> moving to next item
12:50:11 <davemc> #item gluster.next
12:50:37 <hagarth> lot of activities on this front
12:50:42 <davemc> #item  jdarcy to convene a meeting to discuss improvements in small file performance
12:50:49 <davemc> hagarth, yes
12:50:52 <hagarth> davemc: that is TBD
12:51:00 <hagarth> let us carry forward this action item
12:51:03 <davemc> #action  jdarcy to convene a meeting to discuss improvements in small file performance
12:51:25 <hagarth> I am interested in addressing a lot of our complaints about small file perf in 3.7
12:51:26 <davemc> I think most of these are still in the working category
12:51:44 <davemc> I haven't may feature pages
12:52:06 <davemc> we do have messaging for the two items
12:52:15 <davemc> are we going to send that email out today?
12:52:21 <hagarth> yes, I want to provide an update on 3.7 and 4.0 planning
12:52:44 <davemc> The suggested text you sent was fine with me
12:52:50 <hagarth> basically we will announce planning for 3.7 and 4.0 today to the broader community
12:53:01 <hagarth> in essence, we are kickstarting both releases today :)
12:53:14 <hagarth> davemc: cool
12:53:30 <hagarth> 3.7 will follow the 6 month release cycle
12:53:30 <kkeithley> sorry to interrupt the new thread, just want to check that I should not have the 3.4.6 release build send email?
12:53:48 <hagarth> kkeithley: send that email to packagers
12:53:59 <ndevos> kkeithley: and not to the lists
12:54:09 <kkeithley> acknowledged
12:54:10 <hagarth> 4.0 will have a longer release cycle
12:54:18 * ndevos wonders if that could be changed by default?
12:54:52 <hagarth> more details on both releases will be available in the announcement email to be sent later
12:54:55 <ndevos> hagarth: "longer" - any definition thereof?
12:55:03 <ndevos> ok :)
12:55:19 <hagarth> ndevos: looking at a 12 - 15 month cycle (ballpark estimate) :)
12:55:53 * ndevos throws a squash ball and sees where it lands
12:56:26 <hagarth> any questions, thoughts on gluster.next releases?
12:56:43 <ndevos> maybe after the announcement :)
12:56:44 <davemc> do we have landing pages yet?
12:57:00 <hagarth> davemc: yes
12:57:04 <davemc> k
12:57:18 <hagarth> #link www.gluster.org/community/documentation/index.php/Planning37
12:57:45 <hagarth> #link http://www.gluster.org/community/documentation/index.php/Planning40 (more details will appear here soon)
12:58:18 <davemc> hagarth. tks.  I need to figure out how to ge tthose more visible on th esite
12:59:03 <davemc> I'm ready to see the announce sent out for 3.7 & 4.0
12:59:15 <hagarth> davemc: can we start overview sessions on planned features in 3.7 from next week?
12:59:37 <hagarth> davemc: overclk is happy to do an overview/review hangout on bitrot detection
12:59:51 <davemc> works for me.
13:00:09 <hagarth> davemc: let us work with overclk on this
13:00:18 <davemc> hagarth, k
13:00:33 <davemc> almost over
13:00:43 <davemc> #item other items
13:00:58 <davemc> anything other than my plea for video volunteers?
13:01:28 <davemc> going once
13:01:34 <davemc> going twice
13:01:38 <hagarth> davemc: let us follow up on that offline
13:01:41 <davemc> and gone.
13:01:56 <davemc> if nothing else, let's end the meeting
13:02:16 <davemc> thanks all
13:02:19 <hagarth> davemc: thanks!
13:02:20 <davemc> #endmeeting