16:00:12 #startmeeting Fuel 16:00:13 Meeting started Thu Jun 26 16:00:12 2014 UTC and is due to finish in 60 minutes. The chair is vkozhukalov. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:16 The meeting name has been set to 'fuel' 16:00:19 Hi everyone 16:00:22 Hi 16:00:23 who is here? 16:00:28 o/ 16:00:29 Hi 16:00:31 hi fuelers! 16:00:43 Hello! 16:00:59 as usual agenda is here 16:01:05 #link https://etherpad.openstack.org/p/fuel-weekly-meeting-agenda 16:01:16 I'm here 16:01:17 #chair vkozhukalov 16:01:18 Current chairs: vkozhukalov 16:01:26 hi 16:01:36 #topic blueprints process proposal 16:02:07 #link http://lists.openstack.org/pipermail/openstack-dev/2014-June/038501.html 16:02:11 hi 16:02:26 hi 16:02:39 hi 16:02:42 dpyzhov has a series of suggestions about how to improve blueprints flow 16:02:50 dpyzhov: around? 16:03:07 hi 16:03:18 Yes, I send proposals, please review and respond 16:03:44 I need your opinion before updating our process 16:05:05 The main ideas are: do not track features as bugs and do not track future blueprints in current release 16:06:00 you mean even very minor features must be treated as blueprints, right? 16:06:09 dpyzhov: ok 16:06:09 Yep 16:06:16 but no spec 16:06:18 +1, the proposal needs a bit more clear-cut criteria to tell blueprints from bugs, and bluprints from work items, but I support the proposal itself 16:06:25 mattymo: exactly 16:06:30 +1 16:06:38 +1 16:06:40 so guys, please give your opinions in ML 16:06:46 moving 16:07:01 #topic bug squashing day results 16:07:31 angdraug: could you share status of bug squashing day, please? 16:08:16 dpyzhov: angdraug: would you like to share any stats about how squashing day was this week 16:08:51 summary is: we've improved our process from last week, but we didn't catch up with new bugs opened during the week 16:09:08 we are also inflating bug priorities a little 16:09:19 angdraug, we've got everyone on feature work and still no dedicated bug team 16:09:27 it was proposed last cycle, but not actually implemented 16:10:17 mattymo: we will have feature freeze in a week. I hope we will have more free hands for bugfixing after it 16:10:38 #link https://docs.google.com/a/mirantis.com/spreadsheets/d/10mUeRwOplnmoe_RFkrUSeVEw-__ZMU2nq23BOY-gzYs/edit#gid=1683970476 16:10:39 i believe it was much better than last week 16:10:39 maybe we can expose stats in ML 16:10:39 dpyzhov: will you share stats in ML? 16:10:54 stats are in the link ^ 16:11:37 angdraug: did adding 30K lines of code help or bugs per K? 16:11:50 marginally, we're still ways above 1.0 16:12:24 we should be more mindful of bug priorities, and we should look to make bug squashing day have more impact 16:12:28 moving on? 16:12:38 looks like vkozhukalov is offline 16:12:50 #topic 5.0.1 release schedule 16:13:01 I propose to set hard code freeze for 5.0.1 to 7/2 (day after next bug squashing day) 16:13:17 if all goes well, this will allow us to release 5.0.1 on 7/7 16:13:30 angdraug: actually, we depend on upgrade feature 16:13:35 #chair dpyzhov 16:13:43 #topic 5.0.1 release schedule 16:13:53 #topic 5.0.1 release schedule 16:14:02 dpyzhov: good point, what's the status of that? 16:14:36 we still don't have dedicated person to work on 5.0.1 upgrades 16:15:03 I suggest to move akasatkin1 back to upgrade team 16:15:21 in order to fix this asap 16:15:23 there should not be problems to backport upgrade system from 5.1 to 5.0.1 16:15:39 I'm agree 16:15:43 can be done by end of next week? 16:15:49 by 7/2? 16:15:58 I believe so 16:16:03 evgeniyl: ? 16:16:04 akasatkin1: can you confirm? 16:16:06 I think yes 16:16:07 no objections 16:16:25 ok, any other details? 16:16:38 angdraug: what about bugs status? 16:16:48 are we ok with 5.0.1 bugs? 16:17:13 we won't fix all of them, but the fixes we have so far are already enough to push out a release 16:17:27 nice 16:17:30 that's why I want to freeze after next bug squashing day, catch up even more 16:17:32 angdraug: great 16:17:51 moving? 16:17:57 there are several steps, 1. backport upgrade system, 2. fix UI to use latest version of openstack for new cluster (since we don't backport patching) 3. tesing 16:18:02 moving 16:18:20 #topic Features 16:18:32 #topic KVM+NSX 16:18:53 izinovik, please, share your progress 16:19:41 izinovik: around? 16:19:46 most part of BP is implemented, except Testing section 16:19:51 yes, I'm here 16:20:00 right now I do manual tests 16:20:16 BP is 16:20:21 #link https://blueprints.launchpad.net/fuel/+spec/neutron-nsx-plugin-integration 16:21:02 izinovik: thanx 16:21:05 moving 16:21:09 ok 16:21:20 #topic vCenter HA 16:21:34 #link https://blueprints.launchpad.net/fuel/+spec/vcenter-hv-full-scale-support 16:21:46 Implementation of nova-compute HA for vCenter will be started on the next week 16:22:31 ! 16:22:35 don't forget, next week is feature freeze 16:22:41 AndreyDanin: ha is out of scope for 5.1? 16:23:15 We decided to keep nova-compute and nova-network on the controller nodes and move them under Pacemaker 16:23:17 it's a "should" priority, looks like it's not going to make it 16:23:49 AndreyDanin: if it's not merged by feature freeze, it's out 16:24:06 angdraug: exactly 16:24:41 Ok. We will rearange tasks to people today. 16:24:59 ok, thanx 16:25:02 moving on 16:25:17 #topic Cinder VMware VMDK driver 16:25:25 again AndreyDanin 16:25:36 I checked VMDK driver for fuel. I can create and destroy volumes via VMDK manually also attach and detach their to/from instances. Therefore vmdk driver works good. 16:25:48 And next week I will modify fuel to make cinder.conf and nova.conf with support vmdk driver on deploying 16:26:47 igajsin: great, looks like we have good chance to see it in 5.1 16:27:02 any other details here? 16:27:41 moving on? 16:27:48 1sec 16:28:51 igajsin: please take under consideration this https://review.openstack.org/101471 , upstream cinder module has a backend for vmdk 16:29:29 alex_didenko: good point 16:29:43 moving? 16:30:06 #topic vCenter+KVM 16:30:09 We need a lab to start developing of vcenter+kvm. Unfortunately, we will get the lab on next week. It means, we won't provide any code before FF. 16:30:50 I thought we had a vcenter env already? 16:31:10 why cant we use virtual KVM like in devops? 16:31:10 yes. we use it to test and develop other vcenter features 16:31:17 AndreyDanin: what about design? have you already designed feature? 16:31:36 #link https://blueprints.launchpad.net/fuel/+spec/vcenter-nsx-support 16:31:56 AndreyDanin: thanx 16:32:11 hrm, is topic wrong then, url is venter+nsx 16:32:16 topic is vcenter+kvm 16:32:23 but it is nsx 16:32:44 yes 16:32:45 * xarses was confused 16:32:50 me too 16:32:52 topic is vCenter+NSX 16:33:01 topic should be vcenter+nsx 16:33:07 ok 16:33:24 moving 16:33:36 #topic Access control master node 16:34:10 dpyzhov: do you have any info about this feature? 16:34:22 or maybe mattymo has 16:34:32 salmon isn't here 16:34:37 I have an update I sent today to Fuel Dev 16:35:04 The short summary is nailgun work is proceeding well, but there are bugs in the UI and with nailgun-agent (bootstrap and deployed nodes) for unauthenticated requests 16:35:14 #link https://blueprints.launchpad.net/fuel/+spec/access-control-master-node 16:35:42 the fuelclient and fuel-ostf portion of the work is behind schedule and I'm nervous it won't make feature freeze without any help from Fuel Python team 16:36:05 mattymo: what is your estimate about those issues? 16:36:14 keystone implementation is ready and can be merged today 16:36:22 mattymo: is it going to be landed in 5.1 16:36:43 nailgun we can fix in 1-2 days. If we get help from python team, we'll meet the feature freeze date for sure 16:37:09 otherwise it will slip feature freeze and we'll see if it can be still included in 5.1 16:37:28 dpyzhov: do we have a person who can help here? 16:37:38 mattymo: are guys from Poland working on nailgun side? 16:37:58 yes. Lukasz Oles and Kamil Sambor 16:38:23 so, they need some help with it? 16:38:29 in addition, Bogdan D, Vitaly K, and Aleksandra M did UI work 16:38:32 mattymo: Ok, you can contact with a_teem, will see if he can halp 16:38:47 dpyzhov: thanx 16:38:47 *help 16:38:53 great 16:39:08 #topic Secure Fuel Master services 16:39:18 I have an update on that too :) 16:39:23 #link https://blueprints.launchpad.net/fuel/+spec/secure-fuel-master-services 16:39:35 Code is all QA'd and ready to merge. https://review.openstack.org/#/q/topic:bp/secure-fuel-master-services,n,z 16:39:36 mattymo: please share 16:39:48 I just need a generous core reviewer to push them all 16:40:39 =) 16:40:46 mattymo: don't look at me, I'll only merge fuel-web commits that already have 2 +1's from fuel-web developers 16:41:09 angdraug, you can collaborate with dpyzhov 16:41:18 dpyzhov: any suggestions? 16:41:33 ok, guys, moving on 16:41:38 angdraug mattymo: I’m looking no your requests 16:41:43 move on 16:41:50 #topic Fuel Master firewall 16:42:10 all iptables rules have been refactored and is nearly ready for review, but I am still testing it. 16:42:25 #link https://blueprints.launchpad.net/fuel/+spec/master-node-iptables-ruleset 16:42:38 just a quick update. a single patch covers it all. I expect to be ready Monday for QA and merge 16:42:55 mattymo: great 16:43:00 moving on 16:43:19 #topic Nailgun ML2 16:43:31 Neutron ML2 - Making good progress, nearly done re-patching upstream puppet-neutron into fuel. Should have reviews up for testing shortly(maybe tonight / tomorrow). Spec is still not approved. No QA is identified as lead for testing yet. Progress can be tracked at 16:43:31 #link https://github.com/xarses/fuel-library/compare/bp/ml2-neutron?expand=1 16:43:41 corosync HA composition appears to work well without cs_shadow (tested sandboxed since all of fuel isn't plugged in yet). We should consider using it to replace other corosync usage so we don't have to mingle the code into the upstream manifests. 16:43:41 #link https://github.com/xarses/fuel-library/blob/bp/ml2-neutron/deployment/puppet/neutron/manifests/fuel_extras/corosync.pp 16:45:07 any questions here? 16:45:28 xarses: thanx for status 16:45:30 nurla: we need a qa here 16:45:48 #topic Nailgun plugins 16:45:49 move on 16:45:56 dpyzhov: as i know - noqa 16:46:03 meow-nofer__: your turn 16:46:13 so, on plugins there is a good progress 16:46:17 #link https://blueprints.launchpad.net/fuel/+spec/nailgun-plugins 16:46:32 but PR is kinda hard to review, so I arranged a meeting on tomorrow 16:46:52 I’ll go through code and say why it’s how it’s done 16:47:10 meow-nofer__: good idea to have meeting to explain all that stuff to reviewers 16:47:24 I don't see a meeting invite for that 16:47:26 it is a huge patch 16:47:40 so, after that I’ll fix some comments and write a spec for 5.1 16:47:47 not « 16:47:59 not»support plugins», but experimental stuff 16:48:03 #action meow-nofer__ sends invite to xarses 16:48:28 meow-nofer__: thanx 16:48:45 #topic Image based provisioning 16:48:55 good progress here 16:48:55 xarses: sent 16:49:18 moving on? 16:49:24 unfortunately agordeev is sick 16:49:47 so we have some delay in cloud-init stuff 16:50:09 but i believe we'll see this feature in 5.1 16:50:12 moving 16:50:25 #topic Mellanox 16:50:38 #link https://blueprints.launchpad.net/fuel/+spec/mellanox-features-support 16:50:43 Hi, Mellanox status is as follow: 16:50:52 Implementation of Mellanox features over Centos are mostly done and working locally. Finalizing now the coding over Ubuntu. Local manual tests are in progress. 16:51:08 However, I have several questions and I hope this is the correct place: 1. We are depending on the Ml2 package and the Qemu 2.0 upgrade. In our local code we are working with both, but since feature freeze is at 3/7, I wonder if there is a due date for uploading them upstream 2. We have sent a Mellanox Centos required packages list to the dev-list. What is the process we should follow? Should we open a task / bug some 16:51:49 Mellanox drivers are enable in the webui). 3. aviramb had committed our first change for the bootstrap (added Mellanox support for recognizing our HW). What should we do in order to push it? 16:52:13 Our team is new with the processes here so I hope the questions are appropriate. 16:52:21 nuritv: so far, for packages we have been creating a bug and it gets assigned to fuel-osci 16:52:53 Xarses: how can i follow? 16:53:28 nuritv: lets go over the package needs after the meeting in #fuel 16:53:42 Xarses: ok. thanks 16:53:50 rvyalov: around? 16:54:15 nuritv: create a bug in fuel project and add fuel-osci as assignee 16:54:16 nuritv: please paste reviews open so we can start chasing them 16:54:17 can you give comments about ml2 and qemu 2.0 packaging status? 16:54:40 Xarses: here? 16:54:46 yes 16:54:58 #link https://review.openstack.org/#/c/101126/ 16:55:39 vkozhukalov1: neutron ml2 should be in fuel packages already. nuritv is this an issue with the code bits not being present, or the manifests dont configure ml2 plugin yet? 16:56:00 nuritv: it is better to discuss all those questions in #fuel-dev and ML 16:56:14 yes, packages for qemu 2.0 building and they testing 16:56:17 Xarses: manifests 16:56:28 we are working locally with the packages 16:56:55 vkozhukalov1: ok 16:57:06 nuritv: ok 16:57:12 for other packages please create bug on launchpad 16:57:22 nuritv: That will be on me for ml2-neutron, we will need to discuss what entry points you need so that we can ensure it works together 16:57:25 looks like we have couple minutes for other questions 16:58:10 xarses: rvyalov: please stay in contact with nuritv 16:58:26 #topic Open discussion 16:58:31 1 minute 16:59:27 ok guys, looks like no one has any questions here, if has, let's move in #fuel-dev 16:59:33 ending 16:59:43 #endmeeting Fuel 17:00:13 nuritv: we building neutron-ml2 packages, please create bug for add their for ISO 17:00:19 #endmeeting