16:30:51 <sdake> #startmeeting kolla
16:30:52 <openstack> Meeting started Wed Sep 23 16:30:51 2015 UTC and is due to finish in 60 minutes.  The chair is sdake. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:30:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:30:56 <openstack> The meeting name has been set to 'kolla'
16:30:57 <sdake> #topic rollcall
16:30:58 <pbourke_> hi
16:31:00 <rhallisey> hello
16:31:21 <jpeeler> hi
16:31:22 <inc0> good evening
16:31:46 <coolsvap> hello
16:31:48 <akwasnie> hi
16:32:39 <sdake> should be a fast meeting o/
16:32:43 <sdake> #topic rc1 status
16:32:48 <sdake> rc1 releases this friday
16:32:56 <sdake> https://launchpad.net/kolla/+milestone/liberty-rc1
16:33:52 <sdake> inc0 will the logging be done by then?
16:34:06 <inc0> https://review.openstack.org/#/c/215062/ and following patches
16:34:18 <inc0> just few missing, will do tomorrow
16:34:26 <inc0> but I need reviews plz
16:34:36 <sdake> inc0 cool once al lthe patches are in the review queue, plase set to needs review of the launchpad tracker
16:34:51 <sdake> rhallisey ar eall the patches for json work in the queue?
16:34:55 <rhallisey> ya
16:34:59 <inc0> I can do it now, no need for waiting for rest tbh
16:35:10 <rhallisey> swift, horizon murano arn't fully workin in AIO
16:35:18 <rhallisey> but that's because of me
16:35:23 <sdake> ok well folsk if your pathes are in the queue, set to needs review please
16:35:27 <rhallisey> kk
16:35:39 <rhallisey> they all just need one more +2
16:35:44 <rhallisey> sam has +2'd them all
16:36:25 <sdake> #topic open discussion
16:36:52 <rhallisey> nothing
16:36:53 <coolsvap> #link https://etherpad.openstack.org/p/kolla-vagrant-discussion
16:36:55 <inc0> that's really short meeting;)
16:37:02 <sdake> so got an intel 35000 pci-e ssd and it takes 10 minutes to build all of kolla ;-)
16:37:10 <rhallisey> sdake, oh really
16:37:11 <pbourke_> wow
16:37:11 <rhallisey> wow
16:37:14 <pbourke_> :p
16:37:15 <rhallisey> takes me like 3hours
16:37:20 <inc0> sdake, good choice
16:37:20 <sdake> 380$ bucks
16:37:36 <sdake> i am also running btrfs on it
16:37:42 <rhallisey> inc0, you work at intel, unfair enforcement :P
16:37:45 <pbourke_> i got a small request/suggestion for kolla reviews
16:38:00 <rhallisey> inc0, I agree though :P
16:38:12 <inc0> see? It's not empty marketing
16:38:16 <pbourke_> given we do quite a bit of refactoring, it would be nice imo to not do a patchset per image
16:38:20 <inc0> it is marketing, just not empty;)
16:38:39 <pbourke_> big reviews are the other extreme, but 10-20 reviews of essentially the same change is too much clicking
16:39:06 <sdake> i prefer separatepatches
16:39:13 <sdake> that way when i get tired of reviewing a patch set i can sstep away
16:39:17 <sdake> but the work i have done so far is not lost
16:39:24 <pbourke_> both have pros/cons I guess
16:39:25 <pbourke_> its pretty minor
16:39:35 <rhallisey> pbourke_, I know what you mean.  I liked separate for the json stuff because there were a few exceptions
16:39:42 <rhallisey> (keystone mariabd)
16:39:48 <pbourke_> I have two more things to disucss if someone else wants to go first :)
16:39:59 <rhallisey> go ahead
16:40:22 <pbourke_> has anyone else tried Sams ceph patch
16:40:31 <rhallisey> no
16:40:35 <coolsvap> no
16:40:47 <inc0> akwasnie, you got errors right?
16:41:03 <rhallisey> I've only done an eyes only review thus far
16:41:04 <pbourke_> multinode seems to deploy for me but I dont know enough about ceph to smoke test it
16:41:22 <pbourke_> all in one has an ssh issue which Ive posted on the review so would be interested to know if anyone else gets this
16:41:27 <sdake> sam usually takes wed-fri off (itz his weekend)
16:41:44 <pbourke_> I'll hassle people to review on his behalf :)
16:41:46 <sdake> so i suggest you figure out how to smoke test sans tsam :)
16:41:54 <akwasnie> inc0, yes i had, but not sure now which ones
16:42:44 <inc0_> sorry computer died
16:42:49 <pbourke_> final thing for me then is to sort out the vagrant defaults
16:43:36 <sdake> pbourke you want which kind of defaults?
16:43:39 <sdake> and whatare they today
16:43:55 <pbourke_> sdake: tl;dr it was virtualbox but now defaults to libvirt
16:44:11 <coolsvap> I think this is kind of stepping stone for the discussion when we have multiple providers
16:44:19 <sdake> can it be changed with one config option?
16:44:30 <pbourke_> yes and no
16:44:35 <coolsvap> we should consider having it in config
16:44:55 <pbourke_> in theory it can but there's other choices made in the vagrantfile that are dont gel well with both
16:44:58 <jpeeler> i think this is relevant https://etherpad.openstack.org/p/kolla-vagrant-discussion
16:45:28 <pbourke_> my feeling on vagrant is that it's not a production setup, so it should work out of the box as easily as possible at the cost of some performance
16:45:36 <Hui> in my ubuntu host, vagrant failed to create virbr0
16:46:05 <pbourke_> it seems some cores dont use vagrant at all but others such as myself its a daily driver
16:46:10 <pbourke_> so when it stops working I cant review
16:46:39 <sdake> pbourke i think mandre's work was to bring that work to a wider audience so everyone could use it
16:46:46 <sdake> but if he ended up blocking you that is no beuno
16:46:53 <jpeeler> so are we talking about defaults or supporting both providers
16:47:10 <sdake> both providers need to be impliemented
16:47:16 <sdake> i was thiking it would b dsomething simple like a config option
16:47:21 <sdake> "libvirt" or "vbox"
16:47:23 <jpeeler> because honestly who cares what the default is if they both work - i think the problem is that they don't both work currently
16:47:39 <sdake> jpeeler ++
16:47:58 <sdake> pbourke_ is the problem mor ethat the vbox got busted in the process of adding libvirt?
16:48:31 <jpeeler> apparently the other image uses vbox extensions not in the default centos image
16:48:43 <jpeeler> that etherpad captures the story i think
16:49:18 <pbourke_> sdake: correct
16:49:22 <pbourke_> let me check the etherpad
16:49:45 <jpeeler> virtualbox is available for ubuntu and debian, but it's packaged in rpmfusion for fedora. so "out of box" differs based on platform
16:50:05 <jpeeler> mandre has an outstanding patch to bring in alignment the network configuration
16:50:11 <jpeeler> so that'll help
16:51:05 <pbourke_> etherpad look to capture things fine
16:51:12 <pbourke_> we can construct a list of tasks from this
16:51:31 <coolsvap> pbourke_, that was the intention behind creating the etherpad
16:52:14 <coolsvap> i think we want vagrant to work at our end and lets do it for what we need
16:53:56 <sdake> ok so sound slike the resolution to this problem is to finish the job on vagrant
16:53:59 <sdake> which isn't done yet
16:54:18 <sdake> pbourke_ is that accurate?
16:54:49 <pbourke_> sdake: yes that plus us officially deciding to support both in a better way
16:55:06 <sdake> i think we have already made that decision just not executed on it
16:55:16 <pbourke_> sdake: before there was ambiguity I think on which would be most used
16:55:26 <sdake> no there was no libvirt support previously
16:55:32 <sdake> so it was unambiguolsy onl ybox :)
16:55:39 <pbourke_> sdake: right but the libvirt support crippled vbox
16:55:39 <sdake> only vbox that is
16:55:43 <pbourke_> which was frustrating
16:56:01 <pbourke_> but whats done is done, we can focus on fixing
16:56:17 <coolsvap> yes
16:56:22 <sdake> ya there is two solutions revert mandre's patch set or fix mandres implmeentaiton
16:56:27 <sdake> i'd prefer go forward with #2
16:56:39 <Hui> vbox has terrible performance which will affects the deployment time
16:57:32 <pbourke_> #2 is fine
16:57:34 <coolsvap> i think we should have both separate implementations and a configuration option to use whichever you want
16:57:47 <coolsvap> so that one does not conflict other
16:57:54 <sdake> cool i'll send an email to the ml and ask mandre to sort out an implementaiton
16:57:56 <sdake> soudn good/
16:57:57 <coolsvap> with changes
16:58:17 <jpeeler> coolsvap: there's a lot of overlap, they don't need to be completely separate
16:58:40 <jpeeler> especially once the libvirt dhcp fix gets in
16:59:12 <jpeeler> pbourke_: do you install vagrant via the oracle addons repo?
16:59:15 <coolsvap> jpeeler, i agree, i provided an option which we can always merge later
17:00:40 <pbourke_> jpeeler: no I tend to get it from the official site for the latest version
17:00:58 <jpeeler> was just curious
17:01:19 <jpeeler> that's probably enough vagrant talk
17:01:24 <pbourke_> agree
17:01:27 <pbourke_> :)
17:01:36 <sdake> anything else, or hsould I end the meeting?
17:02:10 <coolsvap> i think yes
17:18:06 <rhallisey> #endmeeting kolla
17:18:11 <sdake> #endmeeting