16:01:35 <vkozhukalov> #startmeeting Fuel
16:01:36 <openstack> Meeting started Thu Mar 27 16:01:35 2014 UTC and is due to finish in 60 minutes.  The chair is vkozhukalov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:37 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:39 <openstack> The meeting name has been set to 'fuel'
16:02:36 <vkozhukalov> #chair vkozhukalov
16:02:37 <openstack> Current chairs: vkozhukalov
16:02:54 <vkozhukalov> agenda https://etherpad.openstack.org/p/fuel-weekly-meeting-agenda
16:03:17 <vkozhukalov> #topic Greeting, announcements
16:04:10 <vkozhukalov> who is here, guys?
16:04:26 <ikalnitsky> i am
16:04:28 <evgeniyl`> Hi, I'm here.
16:04:28 <e0ne> i am
16:04:29 <xarses> i
16:04:32 <agordeev> o/
16:04:40 <alex_didenko> I'm
16:04:49 <mattymo> Am I late for roll call?
16:04:50 <dpyzhov> Hi. We have moved from centos 6.4 to 6.5. And we’ve moved to 3.10 kernel on our bootstrap nodes
16:05:04 <vk> me
16:05:07 <vkozhukalov> ok, looks like 16:00 UTC is much more appropriate time
16:05:25 <akasatkin> hi!
16:05:29 <dpyzhov> We have negative feedback about 3.10 kernel, it needs to be investigated
16:05:30 <meow-nofer_> me, too
16:05:46 <vkozhukalov> dpyzhov: what is wrong?
16:06:06 <vkozhukalov> dpyzhov: i mean with 3.10
16:06:10 <dpyzhov> Looks like there is no bug report
16:06:32 <dpyzhov> So I’ll communicate with reporter
16:06:35 <xarses> there was some missing firmware
16:06:41 <xarses> from one of the irc users
16:06:51 <xarses> couldn't use 3.10 on Cent
16:07:10 <dpyzhov> Is it fixable?
16:07:20 <dpyzhov> xarses: ?
16:07:43 <vkozhukalov> let's create a bug and assign it on someone
16:07:55 <vkozhukalov> dpyzhov: will you create?
16:07:57 <xarses> I didn't look very far into it, he'd rather have ubuntu working so i helped with that
16:08:26 <dpyzhov> vkozhukalov: sure
16:08:43 <vkozhukalov> if there is no any other announcements, let's move on then
16:09:12 <vkozhukalov> #topic Blueprints
16:09:26 <evgeniyl`> https://blueprints.launchpad.net/fuel/+spec/nailgun-bootable-checkbox-for-disks this bp assigned to 5.1 milestione but linked bug assigned to 5.0 milestione, should we move it to 5.0?
16:09:35 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/fuel-upgrade
16:10:33 <evgeniyl`> this bp assigned to 5.1 milestione but linked bug assigned to 5.0 milestione, should we move it to 5.0?
16:11:06 <angdraug> I think we should move the linked bug to 5.1 instead
16:11:20 <vkozhukalov> evgeniyl`: we are discussing fuel master node upgrades
16:11:32 <evgeniyl`> For fuel-upgrades we've merged patches with initial implementation of upgrade system.
16:11:38 <akasatkin> https://blueprints.launchpad.net/fuel/+spec/linux-bonding
16:11:38 <akasatkin> It's in our 5.0 release discussion doc but no BP was created until today. Who is aware of this stuff? I moved it to 5.1 as it seems that nobody was dealing with it.
16:12:52 <dpyzhov> akasatkin: Vasilenko is working on it
16:13:07 <akasatkin> Ok
16:13:20 <evgeniyl`> #link https://blueprints.launchpad.net/fuel/+spec/nailgun-bootable-checkbox-for-disks
16:13:22 <meow-nofer_> I finally uploaded working version for Node object in Nailgun, but some tests still failed. After I fix it it will be much easier to implement other blueprints regarding API stabilizing before upgrades
16:13:45 <dpyzhov> akasatkin: assign him and move to 5.0
16:13:48 <dpyzhov> please
16:13:54 <akasatkin> yep
16:14:47 <evgeniyl`> what about nailgun-bootable-checkbox-for-disks bp? Do we want to implement it in 5.0 release?
16:15:20 <dpyzhov> evgeniyl`: moving to 5.1
16:15:30 <evgeniyl`> #link https://blueprints.launchpad.net/fuel/+spec/fuel-stop-provision
16:15:41 <evgeniyl`> What is the status of stop provisioning bp? I see several patches, are they ready to merge? Or do we wait result of testing from QA?
16:15:45 <meow-nofer_> my PR on stopping provisioning is on review
16:15:51 <meow-nofer_> and have 5 plus ones
16:15:58 <akasatkin> dpyzhov: please set priority on linux-bonding.
16:16:36 <meow-nofer_> as far as I know it’s working, let’s ask Vova Sharshov
16:16:57 <evgeniyl`> It looks like he is not here.
16:17:36 <evgeniyl`> Ok, lets ask him directly about status of this bp.
16:17:48 <evgeniyl`> #link https://blueprints.launchpad.net/fuel/+spec/move-naily-to-astute-repo
16:18:00 <evgeniyl`> Vladimir Sharshov moved naily to astute!
16:18:01 <meow-nofer_> yay!
16:18:42 <vkozhukalov> evgeniyl`: great
16:19:10 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/provision-ironic
16:19:22 <vkozhukalov> about Ironic integration.
16:20:05 <vkozhukalov> Last week we agreed with other guys from Ironic about main point of ironic-python-architecture
16:20:20 <vkozhukalov> agordeev sent pull request
16:20:41 <evgeniyl`> vkozhukalov: I see 'how to' link in the bp and I can't open it.
16:20:44 <vkozhukalov> #link https://review.openstack.org/83087
16:22:41 <vkozhukalov> evgeniyl`: it is not actual any more, will remote it
16:23:43 <vkozhukalov> im working on reimplementing comprehensive partitioning stuff in terms of ironic-python-agent
16:24:10 <vkozhukalov> #link https://review.openstack.org/#/c/72950/7
16:24:27 <vkozhukalov> this pull request is about adding uuid into node model
16:24:48 <vkozhukalov> it has 2 pluses, so let's merge it
16:25:21 <evgeniyl`> vkozhukalov: ok, I'll do it.
16:25:31 <xarses> why do we keep updating current migrations?
16:26:16 <vkozhukalov> xarses: why not? i just didn't understand the point
16:26:16 <xarses> i thought the point of alembic was so that we could make incremental changes?
16:26:33 <xarses> well fist off, no one cut 4.1
16:26:43 <vkozhukalov> xarses: yes, you are right
16:26:46 <xarses> so thats bad
16:27:38 <vkozhukalov> evgeniyl`: do you have some comments about alembic and incremental updates?
16:28:14 <xarses> i think we need to talk about alembic strategy
16:28:16 <evgeniyl`> xarses, vkozhukalov we need to have single migraion file for single release.
16:28:39 <vkozhukalov> evgeniyl`: why?
16:28:57 <vkozhukalov> evgeniyl`: is it about kinda stability?
16:29:06 <evgeniyl`> xarses, vkozhukalov and we thought we don't have to make upgrades from 4.1 to 5.0 , but maybe we will have to do it.
16:29:27 <meow-nofer_> actually, I see no problems in keeping many files in repo and just merging them into one before release
16:29:29 <xarses> evgeniyl`: we at least need to cut 4.1 for release
16:29:37 <xarses> from current
16:29:47 <meow-nofer_> yeah, we need to make one for 4.1
16:29:53 <meow-nofer_> not current
16:29:55 <meow-nofer_> 4.0
16:30:10 <vkozhukalov> ok, can we discuss it in details later in mailing list?
16:30:15 <meow-nofer_> they are applied one by one from the first one, which is 4.0
16:30:45 <evgeniyl`> meow-nofer_: I see problems, 1. it will require a lot of time to run a lot of migration file, 2. it's messy and difficult to support
16:30:45 <vkozhukalov> meow-nofer_: could you please send a letter about this topic in fuel-dev mailing
16:30:49 <vkozhukalov> ?
16:30:56 <xarses> correct, and we released 4.1, so it should have been cut out of current. Anyway, we need to start mail list about this
16:31:10 <meow-nofer_> evgeniyl`, 1) no 2) no
16:31:20 <meow-nofer_> but okay
16:31:30 <meow-nofer_> I’ll send a letter
16:31:39 <vkozhukalov> meow-nofer_: great
16:31:44 <vkozhukalov> let's move on
16:32:02 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/fuel-containerization-of-services
16:32:29 <aglarendil> this the essential part of fuel-upgrade feature
16:32:51 <vkozhukalov> aglarendil: status?
16:33:22 <mattymo> vkozhukalov, still working on separation of services from core nailgun puppet module
16:33:41 <aglarendil> @mattymo is leading this feature and current status is that we have several services not contained and puppetized
16:33:54 <mattymo> but the main gist is we will have every daemonized service in its own docker container (except cobbler + apache which cannot be split)
16:34:55 <vkozhukalov> mattymo: aglarendil: ok, good progress
16:34:59 <mattymo> I have a handful of reviews to look at
16:35:14 <mattymo> https://review.openstack.org/#/c/82497/ https://review.openstack.org/#/c/82506/ https://review.openstack.org/#/c/82542/ https://review.openstack.org/#/c/83061/
16:36:05 <mattymo> I still need to work on separating rabbitmq, mcollective, and rsyslog
16:36:08 <aglarendil> but we are rather optimistic on this feature implementation in the upcoming month
16:36:23 <evgeniyl`> aglarendil: ++
16:36:48 <vkozhukalov> ok, what about
16:36:53 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/patch-openstack
16:37:23 <aglarendil> this feature is intersecting with fuel upgrade
16:37:43 <aglarendil> currently we have puppet provider implementation completed that does full installation and rollback
16:38:15 <aglarendil> thus, the only things we need to implement is repository  management and tying of all this stuff with nailgun engine
16:38:28 <evgeniyl`> akasatkin: is working on implenetation of this feature in nailgun
16:38:30 <aglarendil> thus allowing user to download the update and apply it by simple redeployment
16:38:46 <akasatkin> yes. https://blueprints.launchpad.net/fuel/+spec/openstack-patching-nailgun-part
16:39:21 <aglarendil> we are going to achieve this also by moving closer to upstream puppet manifests of stackforge/puppet-*
16:39:28 <vkozhukalov> aglarendil: akasatkin: afaiu, it is still on designing stage
16:39:45 <aglarendil> which part do you mean?
16:40:11 <vkozhukalov> puppet part
16:40:28 <vkozhukalov> or we have kinda POC?
16:41:12 <aglarendil> there is not much to do to upgrade from POC to product version
16:41:17 <akasatkin> vkozhukalov: I started with it yesterday (Nailgun part). AFAIC, I'll finish design tomorrow.
16:41:35 <vkozhukalov> akasatkin: awsome
16:41:38 <aglarendil> we have providers implemented, thus there is not much to implement. all the other stuff is on the puppet side
16:41:51 <aglarendil> *nailgun side
16:41:52 <aglarendil> sorry
16:42:10 <vkozhukalov> aglarendil: great
16:42:29 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/puppet-3-support
16:42:44 <vkozhukalov> it is already implemented
16:42:48 <vkozhukalov> and merged
16:43:06 <vkozhukalov> #link https://blueprints.launchpad.net/fuel/+spec/relocate-haproxy-to-its-own-network-namespace
16:43:26 <aglarendil> this blueprint is essential for HA architecture
16:43:50 <aglarendil> there was quite an interesting bug that after moving of virtual IP addresses from a controller to another one
16:44:11 <aglarendil> there were some connections retained, like VIP -> VIP
16:44:28 <aglarendil> this blueprint puts haproxy and VIPs into separate networking namespaces
16:44:43 <aglarendil> and utilizes proxy_arp in order to implement connectivity
16:45:02 <aglarendil> such architecture makes our HA much more robust and resilient
16:45:18 <vkozhukalov> afaiu, and it looks like moving haproxy to a separate namespace can help to solve this problem, right?
16:45:26 <angdraug> yes
16:45:35 <aglarendil> yep, that what I am talking  about
16:46:05 <vkozhukalov> angdraug: aglarendil: ok
16:46:09 <xarses> yes it should help alot
16:46:22 <aglarendil> currently this blueprint is in "Beta Available" stage and is being tested
16:46:22 <vkozhukalov> #topic Bugs
16:46:56 <vkozhukalov> #link https://bugs.launchpad.net/fuel/+bug/1295131
16:47:20 <vkozhukalov> right now we have hard coded kernel parameters which are in pmanager.py
16:47:48 <vkozhukalov> we have been asked to remove some parameters
16:48:20 <vkozhukalov> so it seems to be suitable to bring them out of pmanager.py into settings.yaml
16:48:32 <vkozhukalov> i'll do it
16:48:39 <xarses> yes that would be great!
16:48:53 <vkozhukalov> #link https://bugs.launchpad.net/fuel/+bug/1294222
16:49:26 <vkozhukalov> this bug is connected with the fact that our discovery agent filters out all removable drives
16:49:50 <vkozhukalov> but some drives which are not removable are marked as removable
16:50:34 <vkozhukalov> right now we have a list of vendors
16:50:45 <vkozhukalov> which we don't have to filter out
16:51:15 <vkozhukalov> but it seems to be much more suitable to stop filtering out removable devices at all
16:51:21 <xarses> i think that we should move business logic like this into nailgun so that it's easier to update
16:51:36 <evgeniyl`> xarses: ++
16:51:43 <vkozhukalov> and allow user to decide whether he wants to use them or not
16:52:09 <xarses> if we remove the filter, we need to make deselecting disks a priorty
16:52:39 <xarses> becase often it hides random usb devices, or the 16gb compact flash card that some systems have
16:52:54 <xarses> which most people dont want to use, especally in a boot raid
16:53:03 <vkozhukalov> yes, i agree, this piece is also on me
16:53:26 <vkozhukalov> #link https://bugs.launchpad.net/fuel/+bug/1297792
16:54:04 <vkozhukalov> it turned out that in some cases creating partitions on mounted drives is not viable
16:54:31 <vkozhukalov> the idea of umounting come from one of our users
16:54:49 <vkozhukalov> it is implemented
16:55:05 <vkozhukalov> #link https://bugs.launchpad.net/fuel/+bug/1291692
16:55:25 <vkozhukalov> in 4.1 we had cciss regression
16:55:54 <aglarendil> the one related to old HP RAID controllers
16:56:10 <vkozhukalov> we tried to look for drive links by their cciss!bla-bla names
16:56:16 <vkozhukalov> aglarendil: exactly
16:56:28 <xarses> apparently, its used on may blade chassis, so not that old
16:56:36 <vkozhukalov> it has been fixed and merged
16:56:38 <xarses> s/may/many
16:56:57 <xarses> can we create some data tests to stop regression?
16:57:12 <xarses> this is the third time it's regressed by release
16:57:19 <vkozhukalov> xarses: so it is not only for old devices
16:57:26 <vkozhukalov> xarses: right?
16:57:49 <xarses> correct, dhblaz's servers are about 2 years old
16:58:01 <xarses> he is first indicator
16:58:03 <angdraug> vkozhukalov: most people with HP hardware come across this problem
16:58:24 <vkozhukalov> xarses: I am thinking of that, but it looks more suitable to implement this test inside ironic-python-agent
16:59:05 <vkozhukalov> ok, guys
16:59:12 <vkozhukalov> we have a couple minutes
16:59:26 <vkozhukalov> Im closing meeting
16:59:38 <vkozhukalov> thank you all for participating
16:59:42 <xarses> thanks
16:59:49 <meow-nofer_> thx
17:00:01 <vkozhukalov> if you have to discuss something let's move to #fuel #fuel-dev
17:00:08 <aglarendil> ерч
17:00:11 <aglarendil> thx
17:00:14 <evgeniyl`> thanks
17:00:17 <vkozhukalov> #stopmeeting Fuel
17:00:33 <vkozhukalov> #closemeeting Fuel
17:01:38 <vkozhukalov> #endmeeting