16:01:58 <kozhukalov> #startmeeting Fuel
16:01:59 <openstack> Meeting started Thu Oct 16 16:01:58 2014 UTC and is due to finish in 60 minutes.  The chair is kozhukalov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:00 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:02 <openstack> The meeting name has been set to 'fuel'
16:02:04 <kozhukalov> #chair kozhukalov
16:02:04 <openstack> Current chairs: kozhukalov
16:02:13 <kozhukalov> agenda as usual
16:02:23 <kozhukalov> #link https://etherpad.openstack.org/p/fuel-weekly-meeting-agenda
16:02:33 <kozhukalov> who is here?
16:02:43 <christopheraedo> Hello!
16:02:44 <mattymo> hi all
16:03:26 <kozhukalov> ok, let's start our discussion
16:03:28 <aglarendil> hi
16:03:33 <kozhukalov> #topic Summary for usage stats collection (mihgen)
16:03:37 <xarses> o.
16:03:49 <mihgen> hey, I'm just from the webex discussion on this topic
16:04:01 <mihgen> basically we wonna get following stats:
16:04:12 <mihgen> a) static, like configuration of deployment
16:04:27 <mihgen> b) usage patterns of openstack, how users use it
16:04:55 <mihgen> c) support cases - know about failures users experience
16:05:15 <mihgen> for a) it's more or less designed, kozhukalov - please help to find blueprint
16:05:42 <mihgen> for b) it's large discussion actually. Ideally we would use ceilometer if we can configure it for very low amount of info
16:05:58 <mihgen> otherwise it can blow up quickly on even 50 nodes
16:06:18 <kozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/anonymous-statistics-performance-testing
16:06:26 <kozhukalov> am I right?
16:06:26 <mihgen> we can't force user to install ceilometer, so we could do it on Fuel master node
16:06:31 <mihgen> kozhukalov: yep, thx
16:07:07 <mihgen> but looks like for first iteration we should just go ahead with something really trivial as we don't have resources to research other ways any deep
16:07:20 <kozhukalov> I was wrong
16:07:25 <kozhukalov> right link is
16:07:26 <mihgen> kozhukalov: yeah, another bp
16:07:30 <kozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/send-anon-usage
16:07:45 <mihgen> so easiest thing would be to just do REST API call to openstack
16:07:47 <mihgen> it
16:08:15 <mihgen> it's not gonna show dynamic / trends of usage, though we can get periodic snapshots of what is going on the cluster
16:08:31 <mihgen> that's short summary. Any questions on this / comments?
16:09:37 <mihgen> kozhukalov: let's move on
16:09:43 <kozhukalov> ok, looks like no q
16:09:55 <kozhukalov> #topic Juno status (mattymo)
16:10:01 <mattymo> Hi kozhukalov!
16:10:10 <mattymo> Great improvement from last week's turbulent events. Everything is very stable and we're working through minor bugs now.
16:10:11 <kozhukalov> mattymo: hi
16:10:16 <mattymo> We fixed a bug in nova-network this week that was blocking instances from getting internet access, along with HA MongoDB.
16:10:19 <mattymo> All components are working, but we have a floating bug of failing to deploy instances on CentOS in HA that is hard to reproduce.
16:10:23 <mattymo> We've already passed all BVTs and now new features can be added on top of vanilla Juno, including our memcached improvements for keystone.
16:10:34 <mihgen> #link https://wiki.openstack.org/wiki/Fuel#Nightly_builds
16:10:34 <mattymo> Master ISO for 6.0 is now juno and works great, but please report any bugs you see, large or small.
16:10:50 <mihgen> anyone can get ISO from there. Choose 6.0 to get Juno
16:11:13 <mihgen> mattymo: thanks you for driving it through!
16:11:27 <mattymo> I learned a lot, so it was worth it
16:11:30 <mattymo> That's all I have for my update.
16:11:31 <mihgen> :)
16:11:36 <mihgen> thx
16:11:37 <kozhukalov> so, now we can add packages into 6.0 mirror, right?
16:11:46 <mattymo> kozhukalov, yes! Please do
16:11:47 <mihgen> kozhukalov: of course
16:11:47 <kozhukalov> rvyalov: can we?
16:11:58 <rvyalov> yes
16:12:26 <rvyalov> but if it is new packages, need add it 6.0-icehouse too
16:12:48 <kozhukalov> ok, there was PR about adding mdadm to ubuntu
16:13:03 <dpyzhov> Should we get rid of 6.0-icehouse?
16:13:03 <kozhukalov> has it merged?
16:13:25 <mattymo> dpyzhov, we haven't finished cherry-picking non-merged upstream patches on top of juno yet
16:13:30 <mattymo> let's keep it until we're done on that end
16:13:33 <kozhukalov> #link https://review.openstack.org/#/c/127163/
16:13:33 <dpyzhov> ok
16:13:35 <kozhukalov> this one
16:13:38 <mihgen> dpyzhov: nope yet
16:13:42 <mihgen> there were issue
16:13:46 <mihgen> fixed by aglarendil today
16:13:47 <dpyzhov> Does juno passed nightly tests?
16:14:01 <mihgen> I want to get status first that we don't see random regressions first
16:14:17 <mihgen> once we are stable there and there, we should drop 6.0-icehouse
16:14:23 <mihgen> another question to devops
16:14:29 <mihgen> what do we use for Fuel CI now
16:14:36 <mihgen> for fuel-library checks
16:14:44 <mihgen> icehouse or we switched to juno packages?
16:14:50 <mihgen> bookwar: teran_ nurla ^^^
16:14:54 <kozhukalov> any other q?
16:15:32 <xarses> mihgen: i'd like to know too
16:15:52 <bookwar> mihgen: we tested juno iso and it failed
16:16:14 <bookwar> mihgen: currently it continues to use very old icehouse iso image
16:16:30 <bookwar> qa team works on fixing the tests
16:16:35 <xarses> yuck
16:16:38 <mihgen> bookwar: failed what?
16:16:48 <mihgen> any bug? BVT passes, so what fails?
16:16:50 <bookwar> https://fuel-jenkins.mirantis.com/job/master_fuellib_review_systest_ubuntu/2826/console
16:17:16 <bookwar> the empty test on already merged fuel-library code ^^
16:17:18 <mihgen> bookwar: nice, good test.
16:17:46 <mihgen> bookwar: any bug on that? who is working on fixing it?
16:18:01 <mihgen> it's important to quickly switch to juno packages in Fuel CI
16:18:23 <bookwar> it is supposed to be cuased by this patch https://review.openstack.org/#/c/128680/ ddmitriev will take a look, if it takes more than half an hour, i'll file a bug
16:18:31 <nurla> Dennis are working on fix
16:19:07 <mihgen> nurla: why don't we just revert it for now?
16:19:15 <mihgen> and then provide a better fix?
16:19:38 <mihgen> trying to fix of the fix of the wrong fix doesn't always work as expected :)
16:20:02 <bookwar> we are not sure about the reason yet, we need some time for debugging
16:20:08 <mihgen> ok
16:20:26 <nurla> we've have same passed tests against 6.0 iso, if dennis can't find problem in hour
16:20:26 <kozhukalov> moving on?
16:20:31 <mihgen> all right, ddmitriev wokring on it?
16:20:33 <nurla> i'll revert fix
16:20:40 <nurla> yep
16:20:43 <mihgen> ok, thanks
16:20:49 <mihgen> let's move on then
16:20:55 <kozhukalov> #topic Gerrit move status (rvyalov)
16:21:04 <rvyalov> we deploy new gerrit instance. we move all non openstack packages and openstack specs to gerrit instance . All the repositories in old gerrit will be read-only access to create request/merge and etc
16:21:17 <rvyalov> We planned to move to current 17/10 Friday
16:21:35 <mihgen> rvyalov: excellent
16:21:54 <mihgen> rvyalov: looking forward to see us being able to build ISO with pure vanilla openstack code
16:22:12 <kozhukalov> cool )
16:22:15 <angdraug> me too. is that the next step?
16:22:30 <angdraug> or is there anything else we have to do first?
16:22:36 <mihgen> angdraug: yes. This is pre-requisite for many other things
16:22:47 <mihgen> as I undertand - rvyalov please correct if not
16:23:07 <rvyalov> yes, we can use public gerrit for get openstack specs and build openstack packages when building ISO
16:23:43 <rvyalov> we publish all openstack specs for deb and rpm
16:24:06 <mihgen> ok, let's cross fingers - I'm expecting to see gerrit working on the weekend :)
16:24:32 <kozhukalov> #topic Decreasing the size of upgrade tarball (kozhukalov)
16:24:43 <kozhukalov> the progress here is quite good
16:24:48 <kozhukalov> there are PRs
16:24:59 <kozhukalov> #link https://review.openstack.org/#/c/128323/
16:25:12 <kozhukalov> about upgrade script
16:25:31 <kozhukalov> #link https://review.openstack.org/#/c/124734/
16:25:35 <kozhukalov> about make system
16:25:51 <kozhukalov> I still need to add some code to last one
16:26:01 <mihgen> kozhukalov: what's going to be the decrease in %?
16:26:16 <mihgen> from what would be if we don't do it?
16:26:22 <kozhukalov> I believe on Monday-Tuesday we'll be ready to show a demo
16:26:40 <kozhukalov> I still don't know exactly
16:26:49 <mihgen> kozhukalov: any testing was already done with repos? Just want to make sure we don't face some weirdness with repos / yum / apt-get
16:26:51 <kozhukalov> one of estimations was 50%
16:27:06 <angdraug> spec is still in review
16:27:12 <angdraug> #link https://review.openstack.org/#/c/121591/
16:27:22 <kozhukalov> yes, Serge Kulanov has tested his patch
16:27:38 <kozhukalov> and he was able to create a diff repo
16:27:56 <mihgen> what do we need to complete the spec?
16:28:02 <kozhukalov> we decided to use md5 sums to figure out which packages we need to add into a diff repo
16:28:43 <kozhukalov> angdraug: yes, I just didn't have enough time to review them
16:29:26 <kozhukalov> mihgen: we need to add another patch to make system so as to be able to insert diff repos into tarball
16:29:46 <mihgen> kozhukalov: ok, thanks. Let's get reviewers..
16:29:59 <kozhukalov> mihgen: ok
16:30:03 <kozhukalov> moving on
16:30:17 <kozhukalov> #topic Image based provisioning (kozhukalov)
16:30:41 <kozhukalov> today around 10 PRs were merged
16:31:41 <kozhukalov> last week we decided to implement choosing which provision scheme to use as a radio button on a setting tab
16:31:43 <mihgen> kozhukalov: hmm where do you get reviewers for them?)
16:31:50 <kozhukalov> it is already merged
16:32:04 <mihgen> kozhukalov: radio button is great. So it's just per env setting?
16:32:23 <kozhukalov> they were D Shulyak, A Kasatkin and A Gordeev
16:32:24 <mihgen> you can choose to provision one env with old anaconda, another - image-based?
16:32:43 <kozhukalov> mihgen: exactly
16:33:01 <mihgen> kozhukalov: what's the overall status? how much work is left still? any serios issues?
16:33:02 <kozhukalov> so we still need to add mdadm package to ubuntu mirror
16:33:14 <kozhukalov> and merge image building scripts
16:33:27 <kozhukalov> which were under work last days
16:34:10 <kozhukalov> i hope it just some bugs are still needed to be fixed and that is it
16:34:51 <kozhukalov> one of the members of QA team has already touched this feature
16:35:03 <kozhukalov> any other q?
16:35:49 <kozhukalov> ok moving on then
16:36:00 <kozhukalov> #topic Upstream neutron manifests (xenolog)
16:36:38 <xenolog> ISO with June and upstream Neutron is ready (custom ISO #65).
16:36:38 <xenolog> I tested it by hands.
16:36:38 <xenolog> System tests (#126, #127, #128) is going now.
16:36:38 <xenolog> My opinion -- changes ready to merge, iso -- for QA team.
16:36:39 <xenolog> Neutron team also tested lastest changes.
16:38:08 <kozhukalov> xenolog thanx
16:38:16 <kozhukalov> any q?
16:38:34 <mihgen> xenolog: thx
16:38:36 <angdraug> xarses: how's review status of ^
16:38:36 <kozhukalov> we still have 20 minutes
16:38:43 <mihgen> what are we waiting for merging this?
16:38:53 <mihgen> aglarendil: angdraug - ^^
16:39:11 <mihgen> when can we merge upstream puppet modules for neutron
16:39:19 <aglarendil> mihgen: we are waiting for custom tests to pass
16:39:29 <aglarendil> mihgen: custom bvts i mean
16:39:36 <aglarendil> mihgen: with latest master
16:39:48 <aglarendil> mihgen: and upstream neutron code
16:39:52 <angdraug> review by xarses from yesterday requested some coding style changes, are all these now addressed?
16:40:33 <aglarendil> angdraug: I am not aware of them
16:40:48 <aglarendil> angdraug: could you send particular links, please?
16:41:03 <angdraug> #link https://review.openstack.org/#/c/103280/
16:41:04 <xarses> angdraug: correct, mostly stile issues, I will check in the AM, there was a note on the card that there where rebase changes I'll probably poke around there
16:41:08 <angdraug> #link https://review.openstack.org/#/c/103279/
16:41:23 <angdraug> #link https://review.openstack.org/#/c/103278/28
16:41:48 <xarses> s/stile/style
16:42:29 <angdraug> #link https://review.openstack.org/#/c/126010/
16:42:36 <angdraug> #link https://review.openstack.org/#/c/126414/
16:43:09 <aglarendil> angdraug: xarses actually, I do not see any -1's from Andrew there
16:43:35 <xarses> thanks angdraug
16:43:56 <angdraug> aglarendil: see patchset 52 in https://review.openstack.org/#/c/103280/
16:44:00 <xarses> xenolog: you fixed the issue you noted in the card with the iso being broken?
16:44:33 <xenolog> broken was previous ISO
16:44:38 <xarses> ok
16:46:00 <kozhukalov> ok, guys
16:46:29 <kozhukalov> looks like we have nothing to discuss any more
16:46:35 <mihgen> thanks fokls
16:46:42 <mihgen> folks*
16:46:47 <kozhukalov> ending?
16:46:54 <mihgen> have a good day!
16:47:05 <kozhukalov> thanx everyone
16:47:15 <kozhukalov> #endmeeting