12:03:16 <kkeithley> #startmeeting
12:03:16 <zodbot> Meeting started Wed Oct 15 12:03:16 2014 UTC.  The chair is kkeithley. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:03:16 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
12:03:17 <zodbot> davemc: Error: Can't start another meeting, one is in progress.
12:03:22 <kkeithley> oops
12:03:23 <kkeithley> sorry
12:03:34 <davemc> now problem, over to you kkeithley
12:03:36 <kkeithley> #moderator davemc
12:03:44 <ndevos> I think it is #chair?
12:03:51 <kkeithley> #chair davemc
12:03:51 <zodbot> Current chairs: davemc kkeithley
12:03:57 <kkeithley> back to you davemc
12:04:13 <davemc> So, roll call, who's here
12:04:19 * kkeithley is here
12:04:19 * Humble is here
12:04:20 * overclk is here
12:04:21 * davemc is here
12:04:21 * lalatenduM_ is here
12:04:24 * msvbhat is here
12:05:17 * ndevos is here too :)
12:05:41 <davemc> pad is https://public.fsfe.org/p/gluster-community-meetings
12:06:35 <Humble> https://public.pad.fsfe.org/p/gluster-community-meetings davemc this is the correct url
12:06:53 <davemc> my system is fighting me today
12:07:23 <davemc> #topic hagarth on new features in 3.6?
12:07:29 <msvbhat> davemc: Where do I update something which I want to discuss today?
12:07:36 <Humble> msvbhat, https://public.pad.fsfe.org/p/gluster-community-meetings
12:07:54 <lalatenduM_> msvbhat, under "Other agenda items?"
12:08:01 <davemc> msvbhat, other items at the bottom
12:08:18 <davemc> no hagarth?
12:08:36 <kkeithley> hagarth is on paternity leave
12:08:38 <Humble> davemc, he is on pto
12:08:38 <lalatenduM_> hagarth is on PTO
12:08:56 <davemc> ah. anyone else got a pointer to the 3.6 new features?
12:09:20 <ndevos> http://www.gluster.org/community/documentation/index.php/Planning36 maybe?
12:09:34 <davemc> ndevos, way out of date I think
12:09:43 <davemc> and both sub and superset
12:09:45 <lalatenduM_> davemc, Has sent a mail couple of minutes ago to gluster-devel ML
12:10:04 <lalatenduM_> sub is "Documentation of features in 3.6"
12:10:10 <Humble> lalatenduM_, yep
12:10:19 <davemc> k
12:10:22 <Humble> hagarth++
12:10:44 <davemc> #topic BSD beta testing
12:10:48 <davemc> still ongoing
12:11:03 <davemc> #topic BSD port maintainer
12:11:40 <davemc> Justin updated the pad. Looks like we have a potential maintainer. Craig Butler
12:12:05 <ndevos> nice, is Craig here?
12:12:26 <davemc> I'll issue a invite directy
12:12:32 <davemc> s/ty/tly/
12:12:58 <davemc> #topic vm slave for brtfs testing
12:13:35 <davemc> overclk, do you knowif this is working? Pad says problems
12:13:44 <overclk> davemc, well.. Justin had issues getting F20 setup
12:14:15 <davemc> we can leave it at that...
12:14:18 <overclk> davemc, it's still not up AFAIK.
12:14:41 <overclk> davemc, I'll ping Justin and check..
12:14:55 <davemc> #topic ndevos hagarth to discuss remaining 3.5.3 blockers
12:15:05 <davemc> marked done
12:15:15 <ndevos> yeah, should not be on the topics anymore
12:15:33 <davemc> #topic glusterFS consultants
12:16:09 <davemc> still TBD. I've identified enough problems with the overall site that we have multiple people looking into it.
12:16:44 <davemc> Deb and I have exchanged emails, so maybe progress
12:17:02 <davemc> #topic kkeithley on memory leak
12:17:10 <davemc> just updated in pad
12:17:21 <kkeithley> right
12:17:48 <davemc> please note the call out for others to look over the valgrind logs
12:18:07 <kkeithley> I'll post the logs somewhere and send email to gluster-devel
12:18:16 <davemc> #topic 3.6.0 date on website
12:18:24 <pranithk> kkeithley: I would like to look at the valgrind logs as well.
12:18:26 <davemc> so, when does 3.6.0 go out?
12:18:48 <pranithk> davemc: hagarth was saying in the week of 20th Oct
12:19:01 <pranithk> davemc: I went to see their family over the weekend. He told me then
12:19:12 <pranithk> davemc: He will be returning on 20th
12:19:16 <davemc> for the website, can we safely say end of the month?
12:19:32 <pranithk> davemc: Not sure :-)
12:19:32 <Humble> pranithk, he was about to release one more beta before GA..
12:19:34 <davemc> <willing to change it again, of course>
12:19:47 <Humble> so I doubt , GA will be on that week
12:19:54 <pranithk> Humble: oh!
12:20:26 <davemc> okay. Just want to say someting other than Sept 29th on web
12:20:35 <Humble> :(
12:20:52 <ndevos> http://www.gluster.org/community/documentation/index.php/Planning36 sais 22 october
12:20:58 <ndevos> *says even
12:21:14 <davemc> was changed in last 24 hours I think
12:21:39 <ndevos> I dont know, but I'd trust that page more than the static web site
12:21:57 <ndevos> do we know who would be maintaining 3.6?
12:22:02 <davemc> unfortunately, I keep getting pinged from dte on static site
12:22:11 <davemc> and yep, I'd just the wiki first
12:22:25 <davemc> ndevos, I don't
12:22:39 <ndevos> well, someone needs to fix the website, I just would not know how to do it
12:22:46 <davemc> so unless someone does, I'm marking that as an action
12:23:05 <davemc> ndevos, working the static site issues
12:23:30 <davemc> #action anoint/confirm 3.6 maintainer
12:23:36 <ndevos> davemc: that is much appreciated!
12:24:00 <davemc> if you have 30-40 hours, I can tell you my static site woes
12:24:04 <davemc> <grin>
12:24:16 <ndevos> oh, I dont need to know the details ;)
12:24:25 <davemc> #topic small file ideas
12:24:44 <bene> I updated etherpad
12:24:53 <davemc> bene, thanks
12:25:06 <davemc> anyone else have a chance to review
12:25:42 <bene> ....crickets...
12:26:03 <davemc> I'm going to continue this as a working topic. We may make it a focus during 3.7 planning
12:26:26 <davemc> #topic 2.4.6 release with DHT
12:26:32 <davemc> any updates
12:26:42 <kkeithley> _3_.4.6
12:26:57 <ndevos> lol, I was really wondering what that would be about :D
12:26:59 <davemc> okay, tht's not my keyboard.
12:26:59 <kkeithley> updated pad. waiting on reviews to merge and new AFR patches
12:27:22 <davemc> not seeing it on the pad.
12:27:36 <davemc> but trust you
12:27:52 <kkeithley> Oh, further down in 3.4
12:28:01 <davemc> #topic infra for Tiering/bitrot
12:28:22 <kkeithley> where we noted agreed to release 3.4.6 with DHT and 3.4.7 when we get memleak fixed
12:28:33 <davemc> mea culpa
12:28:36 <kkeithley> np
12:28:47 <overclk> davemc, so, we had a meeting a while ago regarding this...
12:29:11 <davemc> this == tiering?
12:29:26 <overclk> davemc, yeh plus bitrot
12:29:31 <davemc> k
12:29:46 <overclk> I guess they still are evaluating performance of DBs (leveldb, etc..)
12:29:58 <bene> I thought tiering was part of data classification idea
12:30:07 <davemc> any summary or pointer to notes ?
12:30:16 <overclk> depending on which we'd need to conclude what to use and the data flow...
12:30:39 <overclk> davemc, http://www.gluster.org/community/documentation/index.php/Features/BitRot
12:30:42 <davemc> tks
12:30:43 <kkeithley> tiering, compliance, data classification; birds of a feather
12:31:03 <kkeithley> bitrot is part of all that
12:31:05 <overclk> davemc, this is just the bitrot feature page...
12:31:17 <bene> also I thought we used xattrs not DBs
12:31:19 <ndevos> overclk: you use the rchecksum FOP for bitrot?
12:32:12 <overclk> ndevos, nope, as of now it's a daemon that periodically (well almost) maintains checksums for files (stored as xattrs)
12:32:41 <ndevos> overclk: ah, okay - then we might be able to remove the rchecksum FOP one day
12:33:23 <davemc> more on tiering/bitrot?
12:33:31 <overclk> bene, well, for tiering and compliance, they'd prefer SQL like queries
12:33:32 <pranithk> ndevos: not as long as self-heal uses it :-)
12:33:40 <ndevos> pranithk: it does?!
12:33:51 <pranithk> ndevos: Yes for healing big files
12:33:51 <ndevos> pranithk: we need to talk then :D
12:33:59 <pranithk> ndevos: sure
12:34:15 * ndevos does not have anything more on bitrot
12:34:30 <overclk> So.. we'd need to check if at all they finalized on anything.
12:34:37 <bene> is support for JBOD dependent on bitrot detection?  The feature page doesn't make clear the relationship between the two.
12:34:42 * davemc finds that last exchange intriguing
12:35:39 <davemc> #topic 3.4
12:35:42 <overclk> bene, having bitrot detection would really help in that case
12:35:55 <davemc> jumped the gun again
12:36:19 <davemc> anything on 3.4
12:36:22 <overclk> davemc, I'm done... all yours ;) [will update the pad]
12:36:30 <davemc> e have the updates on a potential 3.4.6
12:36:59 <davemc> anything else anyone wants to bring up?
12:37:12 <davemc> overclk, tks
12:37:13 <kkeithley> as soon as I get the AFR patches we can release 3.4.6beta
12:37:21 <davemc> cool
12:38:07 <davemc> #topic 3.5
12:38:44 <davemc> anything for 3.5?
12:38:47 <ndevos> 1st beta has been out for a little over a week
12:39:05 <ndevos> there are some fixes that would like to get included in the next release, so a 2nd beta is needed
12:39:20 <davemc> sounds fair to me
12:39:37 <ndevos> I'm not sure when I can do a 2nd beta, maybe later this week
12:40:44 <davemc> k
12:41:00 <davemc> any thing else for 3.5
12:41:19 <ndevos> nope
12:41:25 <davemc> silence indicates moving on
12:41:34 <davemc> #topic 3.6
12:41:52 <lalatenduM_> ndevos, do you think with 3.5.3 is going to be a stable one as compared to 3.4.5?
12:42:30 <ndevos> lalatenduM_: it should be stable, that is the goal of the current 3.5 branch :)
12:42:51 <kkeithley> apart from the memleak that both 3.4 and 3.5 have!
12:42:53 <lalatenduM_> ndevos, :)
12:42:56 <kkeithley> and 3.6
12:42:56 <ndevos> lalatenduM_: when people report bugs, then we'll look into fixing that and doing a next release :)
12:43:13 <ndevos> *minor release
12:43:28 <lalatenduM_> ndevos,  got it
12:43:41 <davemc> okay, 3.6
12:44:05 <davemc> we had a small update on timing earlier
12:44:28 <davemc> I'll try to find and fix the static web release date
12:44:54 <davemc> ahem, get fixed, since the current update process does not work for me
12:45:32 <davemc> any updates other than the potential release plan discussed earlier
12:46:03 <davemc> hearing none
12:46:11 <davemc> and pausing
12:46:25 <davemc> #topic Future
12:46:39 <davemc> I'd like to move the small file planning under this
12:47:24 <davemc> and being The New Guy, how do we usually handle future planning?
12:48:07 <davemc> anyone still here?
12:48:16 <kkeithley> feature pages
12:48:21 <ndevos> we have http://www.gluster.org/community/documentation/index.php/Features
12:48:39 <davemc> k. will review
12:48:42 <ndevos> we should probably start a page for GlusterFS-next?
12:48:50 <davemc> agreed
12:49:14 <kkeithley> different than the GlusterFS-4.0 plans?
12:49:36 <ndevos> I dont know, where are they?
12:50:40 <kkeithley> not sure
12:50:55 <ndevos> I guess thats the whole point :)
12:51:07 <lalatenduM_> AFAIR it was a google docs page
12:51:26 <lalatenduM_> but dont have the page handy
12:51:37 <davemc> it might be a good starting point for a -next wiki page
12:51:38 <lalatenduM_> pranithk, ^^
12:52:03 <pranithk> lalatenduM_: ?
12:52:03 <ndevos> that might work for a idea gather session, like an etherpad, but it should land in the wiki, just like the other planning pages
12:52:38 <kdhananjay> http://goo.gl/qLw3Vz
12:52:55 <davemc> I'll take an action to get the 4.0 up to a wiki
12:53:28 <davemc> #action davemc to create a GlusterFS-next planning page using 4.0 google page as seed
12:53:34 <lalatenduM_> thanks kdhananjay
12:54:03 <lalatenduM_> pranithk, kdhananjay answered :)
12:54:09 <davemc> anything before "other"
12:54:09 <pranithk> lalatenduM_: Ah! you were asking for google docs :-)
12:54:16 <pranithk> davemc: why do we not have experimental/stable releases like lot of other open source projects?
12:54:44 <davemc> pranithk, unknown to me.
12:54:52 <ndevos> pranithk: we have nightly builds?
12:54:59 <Humble> even beta releases
12:55:22 <pranithk> ndevos: for example 3.5.x didn;t become reasonably stable till 3.5.2
12:55:28 <lalatenduM_> Also why we are not giving beta builds as debug builds? it would have to catch all asserts
12:55:41 <lalatenduM_> I am one or two bedug builds
12:55:49 <lalatenduM_> s/am/mean/
12:56:10 <ndevos> pranithk: only because people keep finding bugs, we should advertise our beta builds better and get more testers
12:56:35 <pranithk> ndevos: Hmm... How do we get better at this?
12:56:51 <kkeithley> we can certainly turn on --enable-debug for the beta builds
12:56:56 <ndevos> pranithk: no idea... maybe davemc can help with that?
12:57:27 <davemc> some ideas around status and swag, but nothing definitive yet
12:57:36 <ndevos> kkeithley, lalatenduM_: and/or --enable-debug in the nightly ones
12:57:41 <davemc> will think about it
12:57:54 <pranithk> ndevos: For example, for 3.6 Most of the bugs reported in afr are by either Vijay/afr-team.
12:58:06 <pranithk> ndevos: We are not doing a good job of involving the community?
12:58:12 <pranithk> ndevos: For testing I mean
12:58:37 <lalatenduM_> ndevos, may be , because of some reason we never test debug builds , which we should
12:59:11 <ndevos> pranithk: yeah, I think most community users start to use releases and only very few test the beta builds
12:59:32 <pranithk> ndevos: And then we fix those production bugs and then it becomes stable.
12:59:40 <davemc> sounds like a whole topic for deeper discussion, community involvement with beta
13:00:02 <pranithk> ndevos: The only reason we got a stable release fast for 3.5 is because of corvidtech IMO. We need to engage such companies?
13:00:05 <pranithk> davemc: makes sense
13:00:11 <ndevos> lalatenduM_: you should think about it, and send a proposal with --enable-debug to the list, I'm sure we can work something out
13:00:16 <davemc> could be risk aversion to risking data
13:00:21 <kkeithley> what happened to our Test Days?
13:00:41 <pranithk> kkeithley: nothing :-(. Like I said, only vijay and afr-team found bugs for afrv2
13:00:43 <davemc> #action set up discussion on community involvement and beta testing
13:00:44 <ndevos> pranithk: yes, companies like that, but also the smaller uesrs
13:00:53 <lalatenduM_> ndevos, will do that
13:00:56 <pranithk> ndevos: agreed.
13:01:01 <pranithk> davemc: cool
13:01:19 <davemc> one more topic
13:01:27 <davemc> #other items
13:02:05 <davemc> anything to discuss here?
13:02:16 <kkeithley> CentOS's problem!
13:02:22 <davemc> go for it
13:02:39 <ndevos> argh, is that the same as the RHEL-6.6 problem?
13:02:52 <kkeithley> They're running their infra with gluster+gfapi+libvirt, and performance is terrible
13:02:57 * lalatenduM_ leaving now, need to go to another meeting
13:03:05 <ndevos> oh, no, thats different
13:03:08 <kkeithley> I've been begging for someone to take a look. I don't see anything wrong
13:03:36 <pranithk> kkeithley: I am interested.
13:03:47 <kkeithley> link in the etherpad
13:03:55 <davemc> we probably should jump on this
13:04:48 <msvbhat> I don't have anything to discuss but want to mention.
13:05:10 <pranithk> msvbhat: where is the repo for automation hosted?
13:05:11 * ndevos has added an other issue that popped up yesterday...
13:05:34 <msvbhat> I have written a new automation framework and hosted in github for now
13:05:44 <pranithk> msvbhat: link?
13:05:47 <msvbhat> pranithk: https://github.com/msvbhat/glusterfs-automation-framework
13:05:57 <msvbhat> pranithk: But it's not 100% complete yet
13:06:17 <davemc> added pointer to pad
13:06:23 <ndevos> sounds cool
13:06:46 <msvbhat> It is multi node test framework and I have used it to automated more that 100 test cases so far.
13:06:49 <pranithk> msvbhat: we will get it there, don't worry.
13:06:56 <msvbhat> pranithk: :)
13:06:57 <pranithk> msvbhat: nice!
13:07:06 <davemc> ndevos, your addition?
13:07:15 <ndevos> RHEL-6.6 now provides glusterfs-3.6 client packages
13:07:16 <kkeithley> yeah, releasing 3.6 downstream ahead of 3.6 upstream is confusing people
13:07:30 <kkeithley> and breaking things
13:07:34 <msvbhat> And it can be integrated with docker/dockit for automation multi node tests in a single system
13:07:36 <ndevos> and, of course those 3.6 packages have a conflicting version
13:07:41 <davemc> ouch
13:07:50 <davemc> who do I go kick
13:07:54 <pranithk> msvbhat: sweet!
13:08:15 <ndevos> it's not about kicking, it's probably a common issue for community projects
13:08:17 <davemc> (I'll figure it out)
13:08:27 <ndevos> but, we did not need to deal with that before
13:09:05 <ndevos> anyone wanting to install glusterfs from our community repo (on RHEL/CentOS) will likely get troubles :-/
13:09:14 <kkeithley> Well the faux branching of 3.6 for downstream is one part of the problem. The continuous slipping of 3.6 upstream hasn't helped things either.
13:09:28 <davemc> kkeithley, +1
13:10:04 <ndevos> we can 'fix' it by providing a higher version in our community repository
13:10:18 <ndevos> but that is not possible for users running 3.4 or 3.5 :-/
13:10:20 <davemc> Let me go poke a bit on this and see if its something we have dealt with in other community projects
13:10:30 <Humble> msvbhat, pranithk yeah , its possible to integrate with dockit (https://github.com/humblec/dockit)
13:10:43 <ndevos> it's something that we need to fix soon (but not in this meeting)
13:10:53 <kkeithley> beatings continue until morale improves or 3.6 ships
13:11:01 <davemc> I think this is worthy of an action
13:11:31 <ndevos> yes, it needs action for sure - users updating to rhel-6.6 will get hurt
13:11:51 <kkeithley> s/will get/already are getting/
13:12:10 <davemc> #action figure out how to fix the downstream release of 3.6 in RHEL-6.6
13:12:27 <davemc> anything else to discuss today?
13:12:40 <davemc> going, going,
13:12:46 <kkeithley> apparently it's not possible to add an exclude to /etc/yum.repos.d/redhat.repo
13:12:47 <davemc> gone
13:12:55 <kkeithley> that will fix it
13:13:26 <ndevos> kkeithley: yum priorities is an option, it can prefer repositories over others
13:13:45 <ndevos> we need to modify our .repo files, and users need to pull those changes too
13:13:46 <kkeithley> ndevos: is there an example of that somewhere?
13:13:57 <ndevos> kkeithley: sure, but I dont have it here
13:14:09 <kkeithley> okay, we'll take care of it off-line
13:14:37 <davemc> okay, closing the meeting
13:14:49 <kkeithley> #action: ndevos, kkeithley to fix rhel6 yum repos
13:15:13 <davemc> anything else?
13:15:35 <davemc> #endmeeting