12:01:05 <kkeithley_> #startmeeting
12:01:05 <zodbot> Meeting started Wed Apr 15 12:01:05 2015 UTC.  The chair is kkeithley_. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:01:05 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
12:01:12 <kkeithley_> #topic roll call
12:01:30 * overclk is here
12:01:34 * raghu is here
12:01:36 * msvbhat o/
12:01:56 * spot is here
12:02:09 <kkeithley_> lots of lurkers
12:02:18 * partner in the dark
12:02:19 * JustinClift prods tigert
12:02:50 <kkeithley_> moving along...
12:03:00 <kkeithley_> #topic action items from last meeting
12:03:09 <kkeithley_> Pranith should report progress on the oVirt Hyperconvergence feature page
12:03:18 <kkeithley_> Is pranith around?
12:03:27 * tigert is in
12:03:32 * JustinClift prods hagarth
12:03:51 * soumya is here
12:03:59 * hagarth partially available
12:03:59 <JustinClift> kkeithley_: Doesn't look like it
12:04:00 * jdarcy is
12:04:03 <tigert> soumya as in Debloper?
12:04:15 <JustinClift> hagarth: Any idea about the oVirt Hyperconvergence page?
12:04:29 <soumya> nope.. Soumya Koduri
12:04:30 <kkeithley_> different soumya
12:04:41 <kkeithley_> okay, no pranith
12:04:52 <kkeithley_> hagarth should pass more details about the overlay xlator on to someone attending the meeting
12:05:12 <kkeithley_> anyone here know about overlay xlator?
12:05:20 <overclk> hagarth, I would be interested in that.
12:05:44 <kkeithley_> hmmm, guess not, next
12:06:01 <kkeithley_> jdarcy will briefly look into writing a script to match the known segfault issue
12:06:13 <kkeithley_> Is this: http://review.gluster.org/#/c/10157/ ?
12:06:28 <jdarcy> No, that's kind of what I did instead.
12:06:39 * kshlm is late
12:06:49 <jdarcy> I should probably still write the segfault identifier.
12:07:19 <JustinClift> Hmmm, not many people on the Etherpad for this meeting.
12:07:25 <JustinClift> This is the Etherpad, if it's helpful: https://public.pad.fsfe.org/p/gluster-community-meetings
12:07:31 <kkeithley_> I keep disconnecting from the e'pad
12:07:36 <JustinClift> :(
12:08:05 <kkeithley_> #action jdarcy   should probably still write the segfault identifier.
12:08:09 <JustinClift> kkeithley_: There was a large outage in PHX2 earler on apparently :/
12:08:48 <kkeithley_> yup
12:08:53 <JustinClift> #action overclk will get details of the overlay xlator from hagarth
12:08:58 <JustinClift> :D
12:09:18 <kkeithley_> #moderator JustinClift
12:09:18 <overclk> Thanks JustinClift :)
12:09:37 <kkeithley_> okay, next item
12:09:46 <kkeithley_> jdarcy will create the glusterfs-4.0.0 tracker bug and sets the Alias field to glusterfs-4.0.0
12:09:54 <kkeithley_> appears partially done
12:09:58 <JustinClift> Do I need to be a mod for #action to work?
12:10:00 <kkeithley_> any other progress to report?
12:10:06 * kkeithley_ doesn't know
12:10:14 <JustinClift> #action overclk will get details of the overlay xlator from hagarth
12:10:20 <JustinClift> Well, it'll either happen once or twixe
12:10:23 <jdarcy> kkeithley_: I think I can manage that one.
12:10:25 <JustinClift> Twice evern. :)
12:11:00 <JustinClift> jdarcy: I looked at the tracker bug for 4.0.  It didn't seem to have an alias for 4.0.0 ?
12:11:12 <JustinClift> jdarcy: This one worked: https://bugzilla.redhat.com/showdependencytree.cgi?id=glusterfs-4.0
12:11:24 <JustinClift> jdarcy: This one didn't: https://bugzilla.redhat.com/showdependencytree.cgi?id=glusterfs-4.0.0
12:11:35 <JustinClift> (if that clarifies things)
12:12:04 <kkeithley_> anything else on this item?
12:12:10 * JustinClift shakes head
12:12:22 <kkeithley_> #topic GlusterFS 3.7
12:12:34 <kkeithley_> hagarth?
12:13:13 <kkeithley_> the branch was created, email sent, a release is imminent....
12:13:53 <kkeithley_> anyone else want to weigh in with anything?
12:14:17 <kkeithley_> going once
12:14:20 <kkeithley_> going twice
12:14:21 <jdarcy> We *really* need to get on top of the regression tests, or this will never finish.
12:14:27 <kkeithley_> agreed
12:14:56 <jdarcy> That means giving regression-test fixes a higher priority (relative to feature patches), actively reviewing, not adding new super-long tests, etc.
12:14:57 <kkeithley_> do we know what's going on with the regression failures?
12:15:55 <jdarcy> I think many, but not all, are timing-related.  We're not waiting long enough for various config actions to complete (because they take longer than they used to) before we try to move on, then we blow up.
12:16:10 <jdarcy> That's the most common pattern, but others are just plain bugs either in the test or the code.
12:17:18 <kkeithley_> Do we know why those config actions are taking longer? Is it because the VMs are just slower? or overloaded?
12:17:34 <kkeithley_> Or slower because they're overloaded?
12:18:20 <jdarcy> Overload (or more likely "bad neighbor" problems within Rackspace) might be a transient issue occasionally.
12:18:29 <kkeithley_> (And is there a reason the netbsd regression queue seems to always be really long?)
12:18:35 <jdarcy> More often I think it's because we really do more.  More translators to initialize, more daemons to start, etc.
12:19:15 <kkeithley_> what should we do to get this resolved? Besides sit around the table and look at each other?
12:19:17 <jdarcy> The brick translator stack has gotten noticeably deeper.
12:19:38 <overclk> and geo-replication regression test(s) suite has recently been merged..
12:19:39 <hagarth> sorry, had to be away for a bit
12:19:55 <jdarcy> Our two main options are to make things faster, e.g. we do an *amazing* amount of work just to validate options, or increase some of the timeouts.
12:20:16 <hagarth> I think the status looks quite apt and agree with Jeff that regression tests need to be accorded the highest priority
12:20:31 <JustinClift> Increasing timouts is fine from my point of view
12:20:44 <JustinClift> They're just values which aren't actually affecting us negatively
12:20:56 <hagarth> I am also wondering if we should run every patchset with a new/modified test unit about X times (X>10) before merging it.
12:21:03 <JustinClift> eg when a timeout thing like this errors out, it's a waste of space for us because it's not showing a real bug in the code
12:21:15 <JustinClift> hagarth: Interesting idea
12:21:26 <jdarcy> Increasing timeouts is a bit of a band-aid, but at this point I think getting the pipeline flowing is important enough to justify it.
12:22:14 <hagarth> jdarcy: agree
12:22:28 <jdarcy> hagarth: A lot of the problems tend to *show up* in old tests, even if they're *caused* by new code.
12:22:42 <kkeithley_> can we have a global manifest timeout that's used, so when we need to lengthen the timeouts the next time we only need to change it in one place
12:22:51 <hagarth> jdarcy: unfortunately that is true.
12:23:29 <hagarth> should we also introduce a non-durable mode in glusterd for running tests? so that we don't spend a lot of time fsync'ing during volfile generation.
12:23:40 <jdarcy> kkeithley_: There's a set in include.rc that covers *most* of what we'd need.  I tried increasing the most important one, but it failed because there are still enough non-timing-related failures to get in the way most of the time.
12:24:18 <jdarcy> Also, I think we really need http://review.gluster.org/#/c/10157/ so we don't waste those resources continuing tests that have already failed.
12:25:10 <hagarth> jdarcy: will be merging 10157 shortly
12:25:16 <jdarcy> Yay!
12:25:35 <jdarcy> Unfortunately it does make the output look a bit different.  :(  Small price to pay, I guess.
12:25:57 <kkeithley_> so 10157 now needs to be backported to the 3.7 branch
12:26:23 * kkeithley_ uses his usual keen eye for the obvious
12:26:52 <hagarth> jdarcy: that should be fine
12:27:09 <jdarcy> kkeithley_: I'll get on that right after the meeting.
12:27:11 <kkeithley_> #chair hagarth
12:27:11 <zodbot> Current chairs: hagarth kkeithley_
12:27:20 <kkeithley_> #chair JustinClift
12:27:20 <zodbot> Current chairs: JustinClift hagarth kkeithley_
12:28:22 * JustinClift wants the musical one :(
12:28:34 * JustinClift is never satisfied ;)
12:28:42 <tigert> everyone wants that when the music stops
12:28:45 * kkeithley_ thinks the music has stopped
12:29:20 <kkeithley_> #action jdarcy to backport  http://review.gluster.org/#/c/10157/ to release-3.7
12:29:41 * jdarcy plays musical chairs with John Cage's 4'33"
12:29:43 <hagarth> 10157 merged now
12:30:48 <kkeithley_> anything else anyone wants to say about 3.7 before we move on?
12:30:48 <jdarcy> W00t.
12:31:10 <kkeithley_> hagarth, anything?
12:31:29 <hagarth> kkeithley_: nothing more, I plan to send a detailed email on the todo for 3.7.0 later
12:32:05 <kkeithley_> #topic GlusterFS 3.6
12:32:54 <kkeithley_> raghu?
12:32:58 <raghu> there were some build failures in 3.6. 2 patches have been sent to fix them. I wanted to include them before making 3.6.3. But unfortunately since they are build failuresI am not able override jenkin's failure
12:33:13 <hagarth> raghu: I can help you there
12:33:19 <raghu> I mean I am not able to merge those patches
12:33:42 <raghu> http://review.gluster.org/#/c/10208/ and http://review.gluster.org/#/c/10243/
12:33:49 <kkeithley_> #action hagarth to help raghu with merging fixes for build failures
12:33:49 <raghu> these are the patches.
12:34:25 <raghu> I will make 3.6.3 once they are merged.
12:34:48 <kkeithley_> beta1, right?
12:35:10 <kkeithley_> or GA?
12:35:34 <raghu> I want to make GA since they are minor fixes. But can make a beta3 if needed
12:35:59 <hagarth> raghu: my vote is to go GA
12:36:35 <JustinClift> If you're sure they won't affect things negatively, then yeah, GA it :)
12:36:41 <raghu> hagarth: Yeah. Even I think so. The fixes just handle compiler warnings. Functionality is not affected by them
12:36:56 <hagarth> right, let's GA tomorrow :)
12:37:19 <raghu> Sure. 3.6.3 tomorrow :)
12:37:27 <kkeithley_> #info 3.6.3 GA tomorrow
12:37:55 <kkeithley_> anytihng else on 3.6?
12:38:04 <raghu> kkeithley_: nope.
12:38:19 <kkeithley_> #topic GlusterFS 3.5
12:38:23 <kkeithley_> ndevos is not here
12:38:50 <kkeithley_> 3.5.4 (beta1 IIRC) is released, and we're waiting for RPMs before announcing
12:39:37 <kkeithley_> 3.5.4 (beta1 IIRC) is released,  _but_ we're waiting for RPMs before announcing
12:40:18 <kkeithley_> #topic GlusterFS 4.0
12:40:32 <kkeithley_> jdarcy?
12:40:55 <jdarcy> Nothing.  Anyone object to a quick meeting *Friday*?
12:41:12 <jdarcy> Too late to send out a notice for Thursday.
12:41:58 <kkeithley_> I'm not hearing any objections
12:42:13 <hagarth> jdarcy: works for me
12:42:16 <jdarcy> Indeed.  OK, I'll schedule that.
12:42:24 <kkeithley_> #action jdarcy to hold meeting on 17 April
12:42:48 <kkeithley_> next?
12:42:55 <kkeithley_> #topic Gluster Summit
12:42:57 <kkeithley_> spot?
12:43:06 <spot> One moment. :)
12:43:17 <spot> http://www.gluster.org/community/documentation/index.php/GlusterSummit2015
12:43:34 <spot> We've started booking travel for the people who needed it.
12:43:45 <spot> For everyone else, our hotel block is open
12:43:49 <spot> Code is "Gluster"
12:44:12 <spot> I'll make a fancy announcement either today or tomorrow, schedule permitting
12:44:16 <jdarcy> Barcelona, baby!
12:44:20 <hagarth> spot: fantastic!
12:44:33 <spot> and we're working very hard to get the invite letters handled today
12:44:35 <kkeithley_> #action spot to announce summit
12:45:08 <hagarth> spot: great stuff, hopefully all who need a visa get that on time!
12:45:14 * spot hopes so too
12:45:43 <kkeithley_> okay, ready to move on?
12:46:02 * JustinClift nods
12:46:03 <kkeithley_> #topic Gluster Forge v2
12:46:19 <kkeithley_> are we looking for status here?
12:46:35 <JustinClift> Well, I guess
12:46:38 <JustinClift> I'm workign on it atm
12:47:00 * JustinClift created initial code for GitHub stats collection yesterday
12:47:21 <JustinClift> grabs stats like "commits per day", # of watchers on the repo, etc
12:47:43 <JustinClift> Today I'm knocking together code to process that info something that can be displayed
12:48:06 <JustinClift> tigert: Will probably knock together some CSS for it, but I haven't asked when he'd have time
12:48:28 <tigert> yeah
12:48:45 <JustinClift> When we have something basic to show people, we can scope out migrating the current forge projects to GitHub
12:48:56 <JustinClift> I'd like to get it all done asap
12:49:14 <JustinClift> It shouldn't really impact people much though
12:49:17 * JustinClift hopes
12:49:53 <kkeithley_> anything else?
12:49:57 <JustinClift> I also want to get the GitHub ID plugin for our Gerrit-next-version working too
12:50:22 <JustinClift> So, when Google OpenID ceases to work on 20th April (5 days from now!), we can use people's GitHub accounts instead
12:50:29 <JustinClift> ^ may be more controversial ;)
12:50:48 <JustinClift> Um, that's about it for this bit
12:50:54 <kkeithley_> That's for Gerrit-next or gerrit-current?
12:51:00 <JustinClift> Gerrit-next
12:51:18 <JustinClift> This means we'll be upgrading Gerrit in the next few days
12:51:21 <kkeithley_> pardon my ignorance. What will we use on gerrit-current
12:51:23 <kkeithley_> ah
12:51:24 <hagarth> JustinClift: post it on -devel and seek feedback?
12:51:46 <JustinClift> hagarth: Sure.  I'm wanting to have something to show for forge v2 when I do that though
12:51:55 <hagarth> JustinClift: ok
12:52:01 <JustinClift> Which is why I'm trying to get basic forge v2 operational by tonight ;)
12:52:11 <JustinClift> Makes for a stronger argument ;D
12:52:43 <hagarth> ok!
12:53:05 <kkeithley_> #action JustinClift to get basic force v2 operational by 15 April
12:53:34 <JustinClift> Could be a long night
12:53:55 <JustinClift> Moving on... ?
12:54:05 <kkeithley_> #topic Planet Gluster
12:54:16 <JustinClift> Demo: http://209.132.178.43/
12:54:22 <JustinClift> (for Planet Gluster)
12:54:37 <hchiramm_> JustinClift++ awesome!
12:54:51 <kkeithley_> #info Planet Gluster Demo at http://209.132.178.43/
12:54:52 <JustinClift> hchiramm_: That's all tigert
12:54:55 * JustinClift had nothing to do with it
12:55:02 <hchiramm_> tigert++
12:55:02 <JustinClift> I just had the IP addr handy ;)
12:55:09 <hchiramm_> JustinClift,  that was for the demo url : )
12:55:44 <kkeithley_> anything else
12:55:46 <JustinClift> tigert will be pushing the code for that https://github.com/gluster/planet-gluster shortly
12:55:50 <JustinClift> That's all for that :)
12:55:56 <kkeithley_> four minutes remaining
12:56:01 <kkeithley_> #topic Open Floor
12:56:03 <kkeithley_> anything?
12:56:34 <tigert> Related to gluster.org web stuff, I have a work-in-progress planet.gluster.org in progress here: http://209.132.178.43/
12:56:44 <kkeithley_> Or did we need to say anthing else about gerrit and OpenID
12:56:45 <kkeithley_> ?
12:56:46 <JustinClift> tigert: Read back a few lines ;)
12:56:56 <tigert> ehehe yeah
12:57:13 <tigert> was having my head elsewhere and just saw kkeithley_'s line :)
12:57:29 <hagarth> tigert: good start for p.g.o!
12:57:38 <tigert> and no double posts it seems
12:57:51 <tigert> reinventing the wheel fixed it!
12:57:56 <hagarth> tigert: that is a double treat for us :)
12:57:59 <tigert> :)
12:58:11 <kkeithley_> anything worth writing once is worth writing 3+ times
12:58:18 <JustinClift> tigert has an account on www.gluster.org VM now.
12:58:36 <JustinClift> He's going to see if the Middleman/Ruby needed for the Planet can run there
12:58:42 <tigert> JustinClift: if you can make a vhost and give me write access to the docroot
12:58:45 <JustinClift> If so, we'll chuck it in a virtual host on that box
12:58:46 <tigert> for planet.gluster.org
12:58:51 <JustinClift> tigert: You have sudo root on there
12:58:54 <JustinClift> Do it yourself ;)
12:58:57 <tigert> I can wire it up on cron wtih my account there
12:59:07 <jdarcy> p.g.o. looks nice.  I should write something.
12:59:11 * tigert thinks he did virtualhosts before
12:59:15 <tigert> in 1997
12:59:17 <tigert> ;)
12:59:30 <tigert> jdarcy: it doesnt update yet automatically
12:59:31 <JustinClift> k.  I'll get it happening then. ;)
12:59:45 <jdarcy> tigert: Don't worry, I'm good at nagging people.
12:59:48 <tigert> JustinClift: sorry, I just think you have more routine = less cycles needed for that :)
13:00:27 <JustinClift> I personally use nginx for stuff.  It's been a while since I've touched apache.
13:00:33 <JustinClift> But, I'll be able to figure it out.
13:00:35 <tigert> JustinClift: this was when apache had this new namevirtualhost feature added! it was mindblowing
13:01:22 <tigert> we also had floppy disk drives and wooden keyboards
13:01:25 <tigert> ahem :)
13:02:06 <JustinClift> k, meeting end time? ;)
13:02:29 * jdarcy mumbles something about punch cards.
13:02:52 <JustinClift> http://www.amazon.com/Hacoa-Wooden-Keyboard-H930-W-Walnut/dp/B0046ZSUI4%3FSubscriptionId%3DAKIAILSHYYTFIVPWUY6Q%26tag%3Dduckduckgo-d-20%26linkCode%3Dxm2%26camp%3D2025%26creative%3D165953%26creativeASIN%3DB0046ZSUI4
13:03:36 <tigert> jdarcy: :)=
13:03:50 <kkeithley_> all done?
13:04:02 * JustinClift nods
13:04:46 <kkeithley_> #endmeeting