gluster-meeting
LOGS
15:01:42 <JustinClift> #startmeeting Weekly Gluster Community Meeting
15:01:42 <zodbot> Meeting started Wed May 14 15:01:42 2014 UTC.  The chair is JustinClift. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:42 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:02:25 <JustinClift> Who's here?
15:02:48 * kkeithley esta aqui
15:02:54 * jamesdotcuff waves
15:03:05 * ndevos signs in too
15:03:19 * hchiramm_ in
15:03:22 <api984> hello from croatia
15:03:28 <JustinClift> :)
15:03:45 <JustinClift> k, lets get this show on the road
15:03:56 <api984> had a good laugh today watching cluster storage debate on yt…. :D
15:04:20 <JustinClift> Google Docs link: https://docs.google.com/document/d/1ieA23wQdUY8HMKwl1kVSP6KqT64AueccrQZGjJzVyNU/edit?pli=1
15:04:21 <glusterbot> Title: glusterpad - Google Docs (at docs.google.com)
15:04:30 <JustinClift> #topic Agenda items from last week
15:04:50 <JustinClift> Heh, no agenda items from last week
15:04:56 <JustinClift> #topic 3.5.1
15:05:09 <JustinClift> Anyone know how docs are looking for 3.5.1?
15:05:20 <ndevos> a lot are still missing?
15:05:27 <JustinClift> Damn.
15:05:37 * purpleidea is here
15:05:49 <JustinClift> Do we have anyone responsible for getting them all in 3.5.1, that hassles people as needed?
15:05:55 <ndevos> #link https://bugzilla.redhat.com/showdependencytree.cgi?hide_resolved=0&id=glusterfs-3.5.1
15:05:59 <glusterbot> Title: Dependency tree for Bug 1071800 (at bugzilla.redhat.com)
15:06:19 <ndevos> those bugs in NEW and mentioning "documentation" still need patches
15:06:33 <JustinClift> ndevos: Damn that a lot still
15:06:44 <ndevos> JustinClift: quite :-/
15:06:59 <JustinClift> ndevos: I think we're going to need someone to be a lead for this, to drive it
15:07:21 <ndevos> but, a lot has been written and posted and merged, thanks for that!
15:07:26 <JustinClift> ndevos: You were doing a lot of stuff around this just after 3.5.0 release
15:07:47 <JustinClift> ndevos: Do you have time to drive this through to completion for 3.5.1?
15:07:50 <ndevos> JustinClift: yes, we do, but I do not have time for that this week
15:07:56 <JustinClift> k, np
15:08:17 * ndevos has a $dayjob too, you know
15:08:18 <JustinClift> ndevos: Any thoughts on who might be, or should we push back timeframe?
15:08:27 <JustinClift> ndevos: Yeah, am aware.  Just asking. :)
15:08:49 <ndevos> JustinClift: no idea, I noticed hchiramm_ and lala did a lot of the reviewing
15:08:51 <purpleidea> ndevos: i guess that means JustinClift is nominating himself... right? ;)
15:08:59 <JustinClift> purpleidea: I really don't have time to
15:09:09 <JustinClift> purpleidea: If I had spare time to, I would.  I really don't though.
15:09:34 <ndevos> hchiramm_: is it something you can follow up on? maybe together with lala?
15:09:48 <JustinClift> hchiramm_: You here? :)
15:09:56 <hchiramm_> ndevos, let me give a try :)
15:10:13 <ndevos> hchiramm_: very cool, thanks!
15:10:42 <ndevos> hchiramm_: just find the feature owners (should be assigned to the bugs as much as possible) and push them to write something
15:10:44 <JustinClift> #action hchiramm_ is lead for the 3.5.1 documentation, responsible for getting it ready for 3.5.1
15:11:05 <JustinClift> Moving on...
15:11:12 <ndevos> hchiramm_: maybe some features really do not need any docs, if you think that is the case, jus tclose the bug :)
15:11:35 <hchiramm_> ndevos, yep..
15:11:52 <JustinClift> kkeithley: The alpha/rc... is there more it's waiting on than the docs?
15:12:24 <ndevos> no #topic?
15:13:16 <kkeithley> #topic 3.4.1, no docs, just reviewing the patches. nagging and begging aren't working, so.... we slip
15:13:26 <kkeithley> no docs, just reviewing the patches. nagging and begging aren't working, so.... we slip
15:13:39 <JustinClift> k, lets slip it then
15:13:49 <JustinClift> kkeithley: How much slip are you guessing?  2 weeks?
15:14:29 <kkeithley> as long as it takes for people to review. The patches have been up for review for over a week
15:14:40 * JustinClift notices the message from kkeithley about slips isn't in gluster-devel archives online (wtf?)
15:14:52 <JustinClift> kkeithley: k
15:15:07 <JustinClift> Damn, we need someone to hassle people for this too
15:15:26 * kkeithley was tempted to threaten "don't make me come over there" ;-)
15:15:29 <JustinClift> Emails to the mailing list obviously aren't working by themselves.
15:15:41 <JustinClift> kkeithley: Try the "don't make me come over there" approach.
15:15:44 <ndevos> #link http://thread.gmane.org/gmane.comp.file-systems.gluster.devel/6430/focus=6456
15:15:46 <glusterbot> Title: Gmane Loom (at thread.gmane.org)
15:15:50 <JustinClift> ndevos: Cool.
15:16:10 * JustinClift is not winning with gluster servers today/yesterday ;)
15:16:13 * kkeithley is "coming over there" anyway
15:16:37 <JustinClift> k, 3.5.1 will slip until we have the things in place we need.
15:16:42 <kkeithley> I'll just sit in people's cubes and stare at them until they review
15:16:51 <JustinClift> Take a lart with you.
15:17:00 <ndevos> kkeithley: oh, you're very welcome here too!
15:17:01 <JustinClift> Something you can get through customs...
15:17:09 <JustinClift> k, moving on
15:17:23 <JustinClift> #topic 3.4.4
15:18:04 <JustinClift> Again, slipping due to lack of reviews?
15:18:05 <kkeithley> fwiw, I need a layover on the way back, 20+ hours in a plane is too long
15:18:19 <kkeithley> anyway, yes, I was really talking about 3.4.4 slipping
15:18:23 <kkeithley> due to lack of reviews
15:18:24 <JustinClift> k
15:18:32 <JustinClift> kkeithley: No worries.
15:19:06 <JustinClift> We really have a persistent problem in not getting traction for these reviews
15:20:00 <JustinClift> #action JustinClift to discuss with hagarth + Mgmt if a new approach to prioritisation is needed
15:20:21 <kkeithley> within teams we have better results getting team members to do the reviews. Problem is, I'm not part of any team
15:20:30 <JustinClift> Yeah. I'll see if there's something RH organisational that can be done there.
15:21:08 <JustinClift> It's kind of seeming like it might be too much focus on immediate deliverables
15:21:12 * kkeithley remembers to be careful what he wishes for
15:21:24 <JustinClift> Yeah, multi-edged sword ;)
15:21:41 <JustinClift> So, I'll have a chat with people smarter + wiser than me, to see what they say
15:21:58 <JustinClift> k, moving on
15:22:06 <JustinClift> #topic 3.6
15:22:14 <JustinClift> Anyone have anything for this topic?
15:22:22 <kkeithley> well, I will in in BLR in about two weeks. If they haven't been reviewed by then, then I will camp out in people's cubes and stare at them
15:22:43 <JustinClift> ;)
15:23:11 <ndevos> kkeithley: I might have some reviews that need some staring too :D
15:23:12 <JustinClift> k, nothing for 3.6?
15:23:23 <JustinClift> ...
15:23:26 <JustinClift> Moving on...
15:23:43 <JustinClift> #topic Regression tests
15:23:53 <JustinClift> So, current setup is sucking
15:24:03 <JustinClift> Much delay. So sad.
15:24:18 <ndevos> yes, it is quite horrible
15:24:29 <kkeithley> what's not working?
15:24:41 <JustinClift> multi-hour delays
15:24:53 <ndevos> well, its working, but you request a regression test, it may fail and you only have the result the next day
15:25:00 <JustinClift> Two potential approaches for fix this:
15:25:21 <JustinClift> a) Integrate my regression testing scripting with Jenkins properly, for on demand scalable testing
15:25:31 <JustinClift> or b) Add some Jenkins slaves in Rackspace
15:25:44 <JustinClift> I'm actually thinking b) might be faster and easier
15:26:00 * ndevos doesnt have a preference, and suggests the easiest
15:26:41 <JustinClift> I've setting up Jenkins in VM locally yesterday, manually copying across settings and keys from build.gluster.org
15:27:10 <JustinClift> There are a _few_ differences in the newer versions of jenkins compared to what we're running, but they don't seem huge
15:27:21 <kkeithley> I am working on getting a beefy machine that we can host VMs on into our PHX co-lo DMZ in case we run up against $ limits on rackspace.  IT can't be bothered to respond to my inquiries though.
15:27:30 <JustinClift> Ugh
15:28:06 <JustinClift> kkeithley: Can you forward your email inquiry to me?  We have a guy in my team in RH that used to work in IT and knows all the right people there to chat with directly
15:28:18 <JustinClift> kkeithley: He gets stuff done. ;)
15:28:33 <kkeithley> I'll forward a link to the eng-ops ticket
15:28:38 <JustinClift> Perfect
15:28:57 <ndevos> JustinClift: do you have time to work on the Jenkins slave(s)
15:28:59 <ndevos> ?
15:29:08 <JustinClift> Yeah
15:29:55 <JustinClift> So yeah, I should have things working in my local setup "fairly well" for a single machine setup.  I'll then get it working with a Jenkins slave, which according to initial reports isn't hard.
15:30:07 <ndevos> cool
15:30:15 <JustinClift> If that goes ok, then I'll setup some Jenkins slaves in Rackspace + hook them into build.gluster.org
15:30:55 <JustinClift> It's probably going to be a lot easier than using my regression testing stuff for on demand, which looks like strange + unusual integration work ;)
15:31:17 <JustinClift> #action JustinClift to hook some Jenkins slaves into build.gluster.org
15:31:31 <JustinClift> Hopefully have it working by Friday (famous last words)
15:31:36 <JustinClift> k, moving on...
15:31:49 <JustinClift> #topic Other agenda items
15:31:52 <JustinClift> Heh
15:31:59 <JustinClift> "Justin broke build.gluster.org overnight.  Sorry. :("
15:32:03 <JustinClift> Yeah, sorry about that.
15:32:22 <JustinClift> Next..
15:32:24 <JustinClift> "Google Docs is a lot harder to work in than Etherpad"
15:32:31 <JustinClift> I put that one on there.
15:32:47 <JustinClift> Personally, I'm finding the editor in this Google Docs to be complete suckage compared Etherpad
15:32:50 * ndevos prefers etherpad too, but a stable one
15:32:50 <kkeithley> did etherpad get fixed?
15:33:06 <purpleidea> -1 to goggle
15:33:20 <JustinClift> kkeithley: Emailed the people running that Etherpad, and they said they wouldn't have a fix for it for at least a few weeks
15:33:37 <JustinClift> kkeithley: So, I'm kind of thinking we go back to an etherpad, but pick a different one
15:33:47 <JustinClift> kkeithley: There were a few other choices people had mentioned.
15:33:54 <JustinClift> (prior to trying Google Docs)
15:34:04 <JustinClift> So I'm thinking we should investigate one of them
15:34:27 <JustinClift> Anyone care strongly, or should I just pick one of the other previous suggestions in time for next week?
15:35:07 <kkeithley> my only pref is for one that works, reliably
15:35:22 <JustinClift> Yeah, that's baseline requirement :)
15:35:24 <ndevos> and without a need for creating an account
15:35:32 <JustinClift> Agreed
15:35:54 <JustinClift> k, I'll pick something for next week after looking through the previous suggestions
15:36:32 <JustinClift> #action JustinClift to locate next Etherpad to try. Must be reliable and not require accounts for ppl
15:36:44 <JustinClift> #topic Any other items?
15:37:52 <hchiramm_> JustinClift,  may be etherpad.in ?
15:38:15 <JustinClift> hchiramm_: Thanks.  I'll add that to the list to try. :)
15:38:24 <hchiramm_> yeah.. do poc :)
15:38:28 <JustinClift> Yeah
15:38:29 <kkeithley> If I (ever) get the machine in the co-lo DMZ, would we want to host our own?
15:38:39 <kkeithley> Or is that too much work?
15:38:52 <JustinClift> kkeithley: JohnMark shot down that idea.  We don't want to host it ourselves.
15:39:04 * JustinClift thinks it would take minimal work, but whatever.
15:39:17 <kkeithley> to ask permission is to seek denial
15:39:23 <JustinClift> Heh
15:39:36 <purpleidea> JustinClift: is the machine properly and fully automated? if so, otherwise it's not minimal work
15:39:39 <JustinClift> Lets try using something external.  If it's all suckage, we'll just spin up our own.
15:39:53 <JustinClift> purpleidea: Yeah, I was thinking a fully automated solution.
15:39:55 <ndevos> maybe OpenShift makes it easy to host our own?
15:40:22 <JustinClift> ndevos: Tried that about a year ago.  Kept falling over and needing cron job restart
15:40:29 <ndevos> ew.
15:40:32 <JustinClift> It might be better now
15:40:42 <JustinClift> (no idea, not really inclined to test it)
15:41:03 <JustinClift> Anyway, lets try other people's solutions :)
15:41:17 <JustinClift> If they all suck, we can self host (fully automated ;>)
15:41:20 <JustinClift> #endmeeting