16:00:14 <vkozhukalov> #startmeeting Fuel
16:00:15 <openstack> Meeting started Thu Apr 17 16:00:14 2014 UTC and is due to finish in 60 minutes.  The chair is vkozhukalov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:16 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:18 <openstack> The meeting name has been set to 'fuel'
16:00:26 <vkozhukalov> #chair vkozhukalov
16:00:27 <openstack> Current chairs: vkozhukalov
16:00:33 <meow-nofer_> hello
16:00:35 <bogdando> hi
16:00:37 <vkozhukalov> Hey everyone
16:00:42 <mihgen> рш
16:01:06 <vkozhukalov> #topic Greeting, announcements
16:01:06 <meow-nofer_> #armchair meow-nofer
16:01:27 <mihgen> hi folks, let's start
16:01:30 <mihgen> agenda is here: https://etherpad.openstack.org/p/fuel-weekly-meeting-agenda
16:01:30 <aglarendil> hi
16:01:45 <mihgen> We are in feature freeze phase
16:01:55 <vkozhukalov> afaiu mihgen is going to tell a bit about Feature freeze & exceptions status
16:01:58 <mihgen> there are a few exceptions we are still working on
16:02:22 <vkozhukalov> mihgen: thx for coping link to agenda
16:02:23 <mihgen> icehouse support was just merged, and pkgs mirrors are being synced now
16:02:28 <mattymo> hi - forgot to greet  saying I'm here
16:02:35 <xarses> here
16:02:48 <dpyzhov> hi
16:02:50 <rvyalov> hi
16:03:00 <AndreyDanin> hi
16:03:02 <mihgen> so everyone who doesn't participate in those exceptions, please jump into bugs
16:03:02 <dshulyak> hi
16:03:17 <ikalnitsky> hi
16:03:22 <mihgen> also, FYI, I've significantly restructured wiki page
16:03:27 <mihgen> wiki.openstack.org/wiki/fuel
16:03:52 <mihgen> so please check it out, and keep watching it, update architecture, etc. when you change it :)
16:03:53 <vkozhukalov> #link https://wiki.openstack.org/wiki/Fuel
16:04:36 <vkozhukalov> ok, mihgen thanks for update
16:04:49 <vkozhukalov> #topic Activities updates & questions
16:05:10 <christopheraedo> hello
16:05:21 <vkozhukalov> #topic icehouse status
16:05:37 <vkozhukalov> so, how about status?
16:05:53 <aglarendil> hi, guys. Icehouse has been merged into master branch along with packages
16:05:59 <aglarendil> we are fighting some issues now
16:06:20 <aglarendil> you can filter them by icehouse tag here: https://bugs.launchpad.net/fuel/+bugs?field.tag=icehouse
16:06:31 <mihgen> in simple mode neutron & nova-network should work
16:06:40 <aglarendil> most of them do already have requests on review
16:06:55 <aglarendil> HA mode is not working for a while. But multinode should
16:07:34 <aglarendil> my estimate that HA should be working as soon as fixes for corresponding bugs are merged
16:07:39 <vkozhukalov> are we going to fix HA mode in the nearest future?
16:07:48 <vkozhukalov> ah, ok
16:07:53 <aglarendil> If the fixes are merged today, then yes
16:08:07 <vkozhukalov> ok, moving on
16:08:08 <aglarendil> the nearest future is 12-24h
16:08:22 <vkozhukalov> #topic  ISO build scripts, packages builds
16:08:25 <dpyzhov> I’m going to remove eggs/gems from mirrors and master node tomorrow
16:08:42 <dpyzhov> Testing started today and it works
16:08:45 <mihgen> so we gonna have RPMs only on master node, right?
16:08:58 <dpyzhov> But it should be re-checked after icehouse merge
16:08:58 <mihgen> or there are gonna be any exceptions?
16:08:59 <mattymo> dpyzhov, if you don't mind me asking - we still have gems in our node installation kickstart in master
16:09:02 <dpyzhov> rpm/deb
16:09:11 <vkozhukalov> and it will be in 5.0, right?
16:09:15 <dpyzhov> Yes
16:09:19 <vkozhukalov> great
16:09:29 <mihgen> it's actually dependent for fuel upgrades feature
16:09:38 <vkozhukalov> any other updates on this topic?
16:09:56 <mihgen> dpyzhov: we will need docs for it for sure :)
16:10:00 <mattymo> vkozhukalov, that's me
16:10:09 <vkozhukalov> moving on
16:10:24 <mattymo> Puppet code has been more or less finalized for all services. Separation of each component of Fuel node is ready and each service functions
16:10:35 <vkozhukalov> #topic fuel master node upgrade status
16:10:40 <mattymo> the remaining loose ends are related to a few configuration parameters for mcollective and cobbler
16:10:57 <evgeniyl> Hi
16:11:02 <evgeniyl> I’m working on upgrade scripts, akasatkin and warpc are working on backend part (nailgun/orchestrator), so right now I’m trying to build all containers for all services, to test my scripts, we don’t have ready Dockerfiles and I try to make it work, it takes some time, I builded several services, and I hope tomorrow I will have all services.
16:11:04 <mattymo> CentOS deploys, but has an mcollective bug. Ubuntu fails due to processing preseed
16:11:15 <meow-nofer_> hi
16:11:24 <evgeniyl> akasatkin: are you here?
16:11:28 <akasatkin> hi
16:11:38 <evgeniyl> Could you please provide status of backend part?
16:12:11 <mattymo> Sergii Golovatiuk and Alexander Didenko have been working on updating our ISO build scripts to build docker containers and bundle them in a new deployment script. It should be finished in the next 24 hrs
16:12:36 <akasatkin> we're agreed on entities  and their placement for upgrade and update.  I'm fixing upgrade part now.
16:12:45 <mihgen> mattymo: cool. are there any blockers or risky items still?
16:12:56 <evgeniyl> Does Alexander have ready to use Dockerfile's?
16:13:27 <mattymo> ubuntu, but I'm 99% sure it's not the fault of any of the services, but instead I didn't rebase my server to the new ubuntu release
16:13:36 <mattymo> evgeniyl, yes
16:13:45 <evgeniyl> Looks like I have a bit outdated Dockerfiles and I have to fix them.
16:13:45 <akasatkin> I'm going to finish it today.
16:13:56 <mattymo> evgeniyl, just write me and we'll set you up
16:14:01 <mihgen> mattymo: what about ubuntu?
16:14:06 <mihgen> why is it in the picture?
16:14:13 <mattymo> mihgen, because we need to deploy openstack still :)
16:14:27 <mihgen> well how is it related to the work you do?
16:14:35 <mihgen> cobbler misconfig / repos?
16:14:56 <mattymo> it's related because I can't deploy an openstack environment yet because of broken items above: mcollective, cobbler parameters, and preseed bug in ubuntu
16:15:05 <mihgen> mattymo: evgeniyl what's gonna be docker version finally?
16:15:09 <mattymo> 0.10.0
16:15:44 <mihgen> ok cool. mattymo - pls share bug to preseed when you get time, I want to take a look
16:15:45 <mattymo> Alexander Didenko found a bug in my commit for docker puppet class I wrote that led to docker breaking. He's proposed a fix and it's waiting on merge
16:16:06 <mihgen> ok. is it all on fuel upgrades?
16:16:23 <vkozhukalov> ok, moving on
16:16:38 <mattymo> yes, move on
16:16:57 <vkozhukalov> #topic Fuel CI status
16:17:06 <bookwar> hi
16:17:16 <vkozhukalov> bookwar: hi
16:17:25 <vkozhukalov> what about ci?
16:17:27 <mihgen> Fuel CI works ;)
16:17:34 <bookwar> it grows
16:17:39 <vkozhukalov> mihgen: ))
16:17:49 <mihgen> http://fuel-jenkins.vm.mirantis.net:8080/ I see :)
16:17:55 <bookwar> this week we've got new server, and we plan to get one more till the end of the week
16:17:58 <vkozhukalov> bookwar: any details?
16:18:10 <vkozhukalov> ok
16:18:11 <mihgen> sorry wrong link, right one: https://fuel-jenkins.mirantis.com/
16:18:22 <mihgen> what about ISO we use for library testing?
16:18:24 <vkozhukalov> #link https://fuel-jenkins.mirantis.com/
16:18:26 <bookwar> unfortunately the number of jobs grows as well
16:18:45 <bookwar> currently we use icehouse iso 409
16:18:46 <mihgen> bookwar: it's actually good, it means we have more contributions )
16:19:10 <mihgen> ok. so all new patches are gonna be verified by new icehouse iso, right?
16:19:29 <bookwar> yes, we test them against icehouse now
16:19:30 <mihgen> as far as I know we had to disable HA check there, as it doesn't work anyway due to bug
16:19:31 <angdraug> any plans about https://bugs.launchpad.net/fuel/+bug/1308010 ?
16:19:46 <mihgen> so it's gonna be only multi-node pass now, right?
16:19:53 <bookwar> but we plan to have new merged iso tonight so we won't need the separate test for that
16:20:19 <vkozhukalov> #link https://bugs.launchpad.net/fuel/+bug/1308010
16:20:19 <angdraug> mihgen: last I checked it was a single-controller ha lately
16:20:19 <bookwar> right now we test icehouse on centos with simple multinode setup
16:21:15 <mihgen> so Fuel CI should correctly +1 good patches which work on multimode, right?
16:21:22 <bookwar> mihgen: yes
16:21:40 <mihgen> cool
16:21:41 <mihgen> thanks
16:21:42 <vkozhukalov> ok
16:21:57 <vkozhukalov> #topic  Ironic + Fuel status
16:22:01 <mihgen> angdraug: what single-controller HA? in Fuel CI? yep, I think so..
16:22:40 <vkozhukalov> the majority of work about ironic integration is around ironic agent
16:23:02 <vkozhukalov> flow extension patch from agordeev is merged https://review.openstack.org/#/c/85412/
16:23:11 <mihgen> as far as I heard, Ironic doesn't want LVM partitioning to be part of the agent, right?
16:23:20 <vkozhukalov> after some discussion with others it was decided not to support lvm in agent (abandoned request)
16:23:32 <vkozhukalov> mihgen: yes, it is true
16:23:47 <mihgen> ok. we still need it in Fuel, it should be implemented somewhere
16:23:47 <vkozhukalov> actually our use cases can be covered with metadisk (MD) feature
16:23:51 <mihgen> any ideas on this?
16:24:06 <mihgen> hmm what about volume group for cinder, for instacne
16:24:09 <vkozhukalov> it is not very big problem as far as we can implement managing lvm in puppet wherever we really need it
16:24:42 <vkozhukalov> we can create empty space for allocating vg and then create it during puppet stage
16:24:51 <mihgen> vkozhukalov: ok
16:25:04 <vkozhukalov> right now working on https://blueprints.launchpad.net/ironic/+spec/ironic-python-agent-partition
16:25:10 <mihgen> vkozhukalov: have you looked at actual integration with Fuel
16:25:20 <vkozhukalov> disk utils parts is still under review https://review.openstack.org/#/c/86163/
16:25:26 <mihgen> I know you had POC working with keystone & ironic deployed on fuel master node
16:25:34 <vkozhukalov> mihgen: no, agent is not still working
16:25:58 <vkozhukalov> rackspace guys have made several requests about ironic driver for agent
16:26:09 <vkozhukalov> it is still under review
16:26:46 <mihgen> vkozhukalov: ok. so for now there is 0 work done in terms of integration?
16:26:47 <vkozhukalov> mihgen: yes, POC works, but it is not very functional w/o agent
16:27:30 <vkozhukalov> we have puppet module to deploy ironic, we'll need to modify it a bit
16:27:46 <vkozhukalov> we have ruby binding for ironic
16:27:55 <vkozhukalov> so we need working agent
16:28:17 <vkozhukalov> and as far as i understand our discovering flow will change
16:28:39 <vkozhukalov> because it is supposed to use agent for discovering
16:28:58 <vkozhukalov> and nailgun will need to ask ironic about available nodes
16:29:47 <mihgen> vkozhukalov: ok. we need to start design doc about it
16:29:47 <vkozhukalov> any other questions?
16:30:03 <mihgen> I don't have any more high level question on this topic
16:30:08 <bogdando> Non backward compatible changes should be avoided, may be we should add new discovery flow, not change?..
16:30:36 <vkozhukalov> #action vkozhukalov will create design doc for integration ironic with fuel
16:30:40 <mihgen> bogdando: well these are details already…. we need to look into details, to get it
16:30:46 <bogdando> consider it at design docs please
16:30:50 <mihgen> and discuss this in the design doc I believe
16:30:59 <vkozhukalov> moving on
16:31:17 <vkozhukalov> #topic  VCenter integration status
16:31:36 <AndreyDanin> There are four pull requests on review.
16:32:08 <AndreyDanin> And modified cirros images have merged to our repo.
16:32:39 <angdraug> mihgen: here's an example of single-controller ha in fuel-ci: https://fuel-jenkins.mirantis.com/job/master_fuellib_review_systest_ubuntu/457/
16:32:40 <mihgen> already merged?
16:32:42 <AndreyDanin> We have a lot of restrictions which force us to make ugly hacks in UI.
16:33:14 <mihgen> AndreyDanin: hacks do not sound any good
16:33:27 <AndreyDanin> mihgen, I mean cirros.vmdk images packed into rpm and deb.
16:33:29 <mihgen> I hope we can avoid many by simplifying use case somehow
16:33:44 <mihgen> AndreyDanin: so it's different package, right?
16:34:21 <mihgen> vk: are you around? I suggest to use simplest and fast solution, instead of making beautiful but very hacky
16:34:36 <mihgen> it's A part of the integration, let's make it better in B…
16:34:40 <AndreyDanin> We can move all vCenter stuff to separate Releases. It will allow us to implement some of these restrictions by Release metadata.
16:34:57 <vk> mihgen, we're trying, but we are still getting more and more additional requirement
16:35:03 <vk> the last one i got 20 min ago
16:35:23 <AndreyDanin> mihgen, vCenter uses vmdk images, not qcow or raw images. We have to repack cirros to vmdk format.
16:35:45 <mihgen> vk: please involve me in the discussion. Let's make it simple but working at least for A part
16:35:47 <xarses> AndreyDanin: mihgen thats what I proposed, but David E. and Roman A made it sound like using a seperate release was off the table. Let's re-raise it if it makes it much cleaner
16:36:15 <xarses> AndreyDanin: cant we just use the image tools to convert it on the fly?
16:36:15 <mihgen> it's gonna be 10x more complex then and buggy
16:36:24 <mihgen> I bet it's something they don't consider
16:36:47 <mihgen> xarses: if we can explain how much pain it is, then it would go easier I believe
16:37:05 <vkozhukalov> ok guy, moving on
16:37:34 <vkozhukalov> #topic openstack update
16:37:42 <akasatkin> Patching of OpenStack. I keep related PRs up-to-date. There are some changes to be introduced after today's discussions. But I don't know for sure do we plan to include update API in 5.0. API and UI are to be tested after these changes. Library part seems to be in good state.
16:37:43 <AndreyDanin> xarses, maybe we can use such tools. I didn't test it. We just repacked cirros and put it to a separate package.
16:38:26 <mihgen> patching of openstack is out, we missed FF date, and it was not any complete
16:38:44 <mihgen> akasatkin: so I don't consider to land web part of patching into master
16:38:45 <akasatkin> ok
16:39:00 <vkozhukalov> moving on
16:39:05 <mihgen> however the part required for fuel upgrades, such as repos versioning, should definitely go
16:39:16 <vkozhukalov> #topic Blueprints which need to be updated
16:39:30 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/relocate-haproxy-to-its-own-network-namespace
16:39:45 <mihgen> xenolog is not around
16:39:54 <mihgen> aglarendil: what's the status ^^
16:40:03 <mihgen> it's actually merged as far as I know
16:40:10 <mihgen> so I'm wondering if we can close it
16:40:15 <vkozhukalov> I believe it is done
16:40:16 <aglarendil> why are we discussing this blueprint?
16:40:21 <aglarendil> it was discussed 2 weeks ago
16:40:28 <mihgen> aglarendil: we need to update the status
16:40:30 <vkozhukalov> because it is in agenda
16:40:30 <xarses> well if we cant upgrade openstack, we can upgrade fuel to upgrade openstack =)
16:40:31 <aglarendil> and it was merged more than 2 weeks ago
16:40:32 <mihgen> I believe it's done
16:40:47 <mihgen> it's merged, but status is not updated
16:40:48 <aglarendil> ok. update the status - blueprint is merged
16:41:00 <mihgen> updated
16:41:04 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/linux-bonding
16:41:04 <mihgen> let's move on
16:41:25 <mihgen> aglarendil: you are again the most aware on this when we don't see xenolog
16:41:35 <aglarendil> linux-bonding is also merged
16:41:43 <vkozhukalov> aglarendil: ok
16:41:46 <mihgen> should we close this one, and create separate for 5.1 for web part?
16:42:06 <aglarendil> I think so
16:42:12 <mihgen> ok
16:42:13 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/refactor-logging-puppet-openstack-services
16:42:13 <aglarendil> there may be some bugs as we did not run all the acceptance testing, but they should be opened as bugs
16:42:16 <bogdando> I've updated the status of BP. The progress now is good (no more blockers with IceHouse ISO to finish it and test results)
16:42:35 <bogdando> I've as well fixed the Py26 compatibility issue in Oslo
16:42:49 <bogdando> which was discovered in Centos
16:43:01 <bogdando> now I'm gonna finish to implement stage 2 of BP
16:43:38 <mihgen> so we need to postpone it to 5.1 then
16:43:43 <vkozhukalov> bogdando: great
16:43:45 <mihgen> we are in FF phase
16:44:06 <mihgen> logs are important, I was hitting a few strange things last days...
16:44:14 <bogdando> mihgen, we could but this is a fix, not a feature to be freezed...
16:44:41 <mihgen> do we have bug or should we then just wait a little fix from you?
16:45:02 <bogdando> I will submit the final patch for fuel-library today, I hope
16:45:14 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/murano-update-services-json
16:45:16 <bogdando> or tomorrow
16:45:27 <mihgen> bogdando: if it's gonna be big, then just forget about it ;)
16:45:39 <mihgen> we need time to stabilize already unstable master...
16:46:08 <mihgen> anyone is aware of murano-update bp?
16:46:16 <mihgen> I see one patch into FL: https://review.openstack.org/#/c/86022/
16:46:33 <mihgen> sorry to OSTF
16:47:28 <mihgen> ok, so let's move on
16:47:30 <vkozhukalov> ok, as far as we don't have people here who knows something about
16:47:35 <vkozhukalov> moving on
16:47:36 <mihgen> I'll follow up with them
16:47:53 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/doc-change-fueladmin-net-on-the-fly
16:48:22 <mihgen> rmoe: do you plan to work on that ^^
16:48:29 <vkozhukalov> rmoe assigned this bp
16:48:30 <mihgen> or should we move it?
16:49:26 <vkozhukalov> moving on
16:49:44 <vkozhukalov> #topic Bugs
16:49:53 <vkozhukalov> any updates?
16:50:03 <mihgen> ok, so as we merged icehouse, we gonna see a lot of them
16:50:10 <xarses> mihgen: it's was going to part of the multiple-cluster-network changes it should be for 5.1
16:50:11 <mihgen> now we have about 8 criticals
16:50:19 <mihgen> we need to urgently take a look into: http://fuel-launchpad.mirantis.com/project/fuel/bug_table_for_status/Open/5.0
16:50:31 <aglarendil> ?
16:50:45 <vkozhukalov> #link http://fuel-launchpad.mirantis.com/project/fuel/bug_table_for_status/Open/5.0
16:50:51 <mihgen> and even more important, keep watching for new bugs and confirm/triage them shortly http://fuel-launchpad.mirantis.com/project/fuel/bug_table_for_status/New/None
16:51:39 <vkozhukalov> #link  http://fuel-launchpad.mirantis.com/project/fuel/bug_table_for_status/New/None
16:51:39 <mihgen> xarses: are you about https://blueprints.launchpad.net/fuel/+spec/doc-change-fueladmin-net-on-the-fly ?
16:51:49 <xarses> mihgen: yes
16:52:06 <mihgen> ok I changed milestone
16:52:12 <mihgen> can you add dependency please?
16:52:27 <mihgen> I ddn't know it was dependent
16:52:40 <mihgen> #link https://www.google.com/url?q=https%3A%2F%2Fbugs.launchpad.net%2Ffuel%2F%2Bbugs%3Forderby%3Ddate_last_updated%26field.status%253Alist%3DINPROGRESS%26assignee_option%3Dany&sa=D&sntz=1&usg=AFQjCNFcQnnjCGFyutlovHYM0Ks8t9tx-A
16:52:45 <xarses> mihgen: I don't have permissions to because i didn't create it
16:53:00 <mihgen> here you can find all staled In progress bugs we need to sort it out
16:54:00 <mihgen> ok I'll create a dependency graph for https://blueprints.launchpad.net/fuel/+spec/doc-change-fueladmin-net-on-the-fly
16:54:12 <mihgen> to link it with multiple-cluster-network
16:54:15 <angdraug> I've added dependency
16:54:23 <vkozhukalov> #action mihgen creates a dependency graph for https://blueprints.launchpad.net/fuel/+spec/doc-change-fueladmin-net-on-the-fly
16:54:46 <vkozhukalov> 5 mins for open discussion
16:54:54 <vkozhukalov> #topic Open discussion
16:55:20 <evgeniyl> #link http://www.mail-archive.com/fuel-dev@lists.launchpad.net/msg00697.html
16:55:28 <evgeniyl> We had a discussion about migration files.
16:55:46 <evgeniyl> My suggestion was to have a single migration file per release
16:56:09 <vkozhukalov> evgeniyl: but everyone else is against, right?
16:56:21 <evgeniyl> Not everyone.
16:56:22 <mihgen> if it simplifies life, why everyone is against?
16:56:44 <mihgen> I agree with evgeniyl that we don't need migrations files before 5.0
16:56:50 <evgeniyl> I had arguments against single file per patch.
16:57:06 <evgeniyl> There will be a lot of organizational problems
16:57:10 <mihgen> at all. And to me it looks like we need for every patch we release, no often
16:57:16 <salmon_> it's easier to have many files. If you want to have one file, someone needs to prepare it
16:57:21 <xarses> I think if we are only going to maintain one file per release, then I don't know if we need to maintain current file since we can just make one at the end of the release
16:57:24 <meow-nofer_> I'm not against having one file per release
16:57:45 <meow-nofer_> but I'm against keeping it all in one file dusing development process
16:57:51 <meow-nofer_> *during
16:58:07 <evgeniyl> then we need a person who will merge all of this files at the end of release
16:58:19 <meow-nofer_> we need just a simplest script
16:58:30 <xarses> evgeniyl: no, jut drop the intermediate files, and generate a new migration
16:58:33 <evgeniyl> I don't think that simple script can solve this problem
16:58:45 <meow-nofer_> xarses, no, this won't work so easy
16:58:57 <evgeniyl> xarses: alembic can't generate migartion file automatically
16:58:58 <meow-nofer_> evgeniyl, actually, it can :)
16:59:09 <salmon_> sure it can
16:59:21 <evgeniyl> So, my suggestion at least squash all migration file in a single fuel_5.0.py file
16:59:32 <vkozhukalov> evgeniyl: xarses: it can but it is not 100% working
16:59:40 <mihgen> folks let's move the discussion to #fuel-dev…
16:59:46 <meow-nofer_> right now I'm agree with evgeniyl about merging it all into one 5.0 file
16:59:50 <mihgen> I assume this channel is gonna be used for next meeting
16:59:51 <evgeniyl> salmon_: it cannot add new type of valume for enum for example
17:00:02 <vkozhukalov> yes, lets go on this discussion in mail list
17:00:20 <vkozhukalov> ending meeting, thanks everyone
17:00:20 <mihgen> welcome to continue it in #fuel-dev right now
17:00:32 <vkozhukalov> #endmeeting Fuel