16:30:11 <sdake> #startmeeting kolla 16:30:12 <openstack> Meeting started Wed Feb 24 16:30:11 2016 UTC and is due to finish in 60 minutes. The chair is sdake. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:30:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:30:15 <openstack> The meeting name has been set to 'kolla' 16:30:16 <sdake> #topic rollcall 16:30:17 <SamYaple> o/ 16:30:18 <pbourke> o/ 16:30:20 <akwasnie> hi :) 16:30:22 <britthouser> 0/ 16:30:25 <sdake> yo folks 16:30:25 <pbourke> so close to being first 16:30:27 <pbourke> cheers SamYaple 16:30:31 <pbourke> :p 16:30:35 <inc0> o/ 16:30:42 <dave-mccowan> o/ hola 16:30:44 <SamYaple> so you know what they call almost winning pbourke ? 16:30:47 <SamYaple> losing 16:30:53 <pbourke> lol 16:30:54 <sdake> age before beauty pbourke ;) 16:31:32 <rhallisey> hello 16:32:00 <jpeeler> hi 16:32:07 <sdake> #topic announcements 16:32:34 <sdake> 1. meeting times have changed, even weeks are 16;30, odd weeks are 23:00 16:32:41 <carolbarret> #openstack-meeting 16:32:46 <carolbarret> #openstack-meetings 16:32:48 <sdake> ISO weeks - to find the week run date "+%V" 16:32:55 <akwasnie> 1630 timezone? 16:33:02 <akwasnie> ah, ok 16:33:12 <SamYaple> akwasnie: UTC 16:33:34 <sdake> carolbarret is there an issue with meeting conflct? 16:33:54 <sdake> the meetings are staying in the channel openstack-meeting-4 16:34:22 <sdake> the irc channel gate sorts out conflicts and -1's the patch if its wrong ;) 16:35:04 <sdake> 2. we need to get cracking on our austin summit sessions - ttx has requested the info in the next few weeks 16:35:06 <sdake> so the next two meetings, we need to sort that out 16:35:11 <vhosakot> sorry, I'm late 16:35:14 <vhosakot> o/ :) 16:35:24 <sdake> 3. mitaka releases between 2nd-4th of march, 1 week to go - and we got a ton of work to do 16:35:57 <sdake> 4. I am going to kick all medium/low blueprints to newton if they are not in the good progress stage, so please make it so in launchpad - if you dont' have permissions ping me and i'll add you to the drivers group to change your blueprint states 16:36:06 <SamYaple> sdake: you mean mitaka-3 right? 16:36:16 <sdake> SamYaple yes mitaka-3 16:36:25 <sdake> then we have 1 month for stablization via rcs 16:36:48 <sdake> the final date I believe is aprril 13th, but not certain - i'll double check and get backt o folks that want o know our actual release date 16:37:06 <sdake> any announcements from community folk? 16:38:08 <sdake> #topic Kolla Austin Summit topics [timebox 15 minutes] 16:38:50 <sdake> #link https://etherpad.openstack.org/p/kolla-newton-summit 16:43:40 <elemoine_> o/ 16:43:49 <xionglingfeng> hi everyone :-) 16:43:53 <elemoine_> sorry for being late, I thought there was no meeting 16:44:51 <dave-mccowan> we're collaborating in https://etherpad.openstack.org/p/kolla-newton-summit 16:45:29 <xionglingfeng> So this conference is only about the design of Newton cycle? 16:47:17 <vhosakot> dave-mccowan: add your name under "Contributors" at the top 16:48:31 <xionglingfeng> me? 16:49:03 <vhosakot> xionglingfeng: you too :) 16:49:30 <xionglingfeng> I am still thinking what shall I write on the pad LOL 16:49:52 <sdake> wheee :) 16:49:54 <rhallisey> xionglingfeng, anything you want to discuss 16:50:08 <sdake> ok lets keep going until the hour mark or until we run out of ideas ;) 16:50:24 <xionglingfeng> Im struggling about the cinder NFS thing :-( 16:51:07 <xionglingfeng> since nova_compute and nova_libvirt r split into two component, it is really difficult to make cinder NFS working 16:52:14 <xionglingfeng> any suggestions on this? 17:00:25 <sdake> ok folks 17:00:34 <sdake> we went past our timebox and our extended timebox 17:00:40 <sdake> lets finish this up next meeting 17:00:57 <sdake> #topic risk item status 17:01:36 <sdake> https://etherpad.openstack.org/p/kolla-mitaka-midcycle-roadmap 17:02:14 <sdake> elemoine_ akwasnie1 diagnostics status? 17:02:19 <sdake> I know heka is msotly merged 17:02:28 <akwasnie1> elasticsearch is merged 17:02:44 <elemoine_> what remains is integrating the two 17:02:45 <akwasnie1> we will start integration this week, i think 17:02:51 <akwasnie1> yup :) 17:02:51 <elemoine_> I created a wip CR for that 17:03:15 <elemoine_> I hope akwasnie1 and I can collaborate on it 17:03:24 <xionglingfeng> quick question: since we already have ELK and heka for logs, why we still keep rsyslog? 17:03:27 <akwasnie1> sure :) 17:03:34 <akwasnie1> rsyslog is removed 17:03:39 <xionglingfeng> cool 17:04:04 <sdake> we have EHK - special snowflakes ftw ;) 17:04:11 <xionglingfeng> LOL 17:04:21 <xionglingfeng> What is the time schedule for kolla release? 17:04:32 <sdake> Jeffrey4l reconfig 17:04:33 <xionglingfeng> I mean, what time is feature freeze, something like this? 17:04:42 <elemoine_> makes me wonder if we should rename the enable_elk var to something else 17:04:45 <sdake> feature freeze is theoretically march 2nd-4th 17:04:54 <sdake> i say theoretically because we are on releease-indepndent schedule 17:05:01 <elemoine_> I guess it could just be enable_elasticsearch in fact 17:05:01 <xionglingfeng> okok got u 17:05:04 <sdake> so we set our own schedules and release criteria 17:05:16 <xionglingfeng> ok. Is there any page for the schedule? 17:05:18 <sdake> however, i'd like to follow upstream release-managed as much as possible 17:05:25 <xionglingfeng> sure 17:05:34 <sdake> xionglingfeng google for openstack mitaka schedule 17:05:36 <elemoine_> sdake: makes sense to me 17:05:43 <sdake> Jeffrey4l ping re reconfig 17:05:57 <sdake> inc0 your up on infrastructure upgrades 17:06:11 <inc0> HAProxy is up for review 17:06:13 <xionglingfeng> I have OS time schedule. Because you mentioned kolla has own schedule so I just wondering if it is documented. 17:06:17 <inc0> rabbitmq is inprog 17:06:22 <inc0> mariadb is on SamYaple 17:06:39 <sdake> keepalived is also in the haproxy review 17:06:51 <sdake> ceph remains 17:06:53 <inc0> we need upgrades of logging stuff, but that's lower priority 17:06:53 <sdake> what else? 17:06:56 <xionglingfeng> what about n-1 support? 17:07:06 <SamYaple> xionglingfeng: that is unlikely to ever happen 17:07:15 <xionglingfeng> lol 17:07:16 <xionglingfeng> kk 17:07:24 <sdake> xionglingfeng add to the etherpad for austin summit discussion 17:07:26 <inc0> but elemoine akwasnie1 you might want to check on this one, hopefully is as simple as rebuild containers 17:07:39 <sdake> its not going to happen in mitaka - mitaka needs more capacity or less features 17:07:46 <elemoine_> inc0: I can have a log at upgrade of logging stuff, although I am not sure what that implies at this point 17:08:07 <sdake> sounds like diags is not really in risk any longer 17:08:09 <inc0> elemoine_, at very least implement naive redeploy with serial plz 17:08:15 <sdake> sounds like infrastructure upgrades is not at risk any longer 17:08:26 <inc0> ceph is still for the taking 17:08:29 <inc0> ceph upgrades 17:08:34 <SamYaple> nah ceph is fine 17:08:34 <sdake> migration path - pbourke ? 17:08:38 <SamYaple> im working on it now 17:08:45 <SamYaple> its not doing storage infrace properly 17:08:49 <sdake> SamYaple can you assign the blueprint to yourself 17:08:56 <SamYaple> i did at one point... 17:09:02 <sdake> and set the state to started or good progress 17:09:05 <elemoine_> inc0: I can have a look, but I need to get more familiar with the exact needs 17:09:08 <sdake> i unset them at midcycle because we moved around who was working on what 17:09:21 <inc0> elemoine_ I'm here to help 17:09:34 <sdake> vm functional testing no longer at risk 17:09:35 <sdake> so that leaves reconfig 17:09:40 <elemoine_> inc0: ok, I'll ping you on IRC then 17:09:56 <sdake> #info reconfig is at risk and essential blueprint for mitaka 17:10:00 <elemoine_> inc0: any container I should look at as a model? 17:10:18 <inc0> elemoine_, keystone is the simpliest, nova is most complex 17:10:25 <elemoine_> is reconfig related to upgrade? 17:10:31 <inc0> not really 17:10:34 <sdake> elemoine_ sort of but not really 17:10:35 <elemoine_> or are the two independent? 17:10:44 <sdake> they are two seprate actions 17:10:47 <xionglingfeng> reconfig? 17:10:51 <elemoine_> ok, I'll figure it out 17:11:10 <xionglingfeng> do u mean that modify the configuration files then redeploy the containers? 17:11:11 <elemoine_> xionglingfeng: reconfig == restart container after config change 17:11:11 <SamYaple> hey i really dont like "reconfig" is there opposition to "reconfigure"? 17:11:14 <xionglingfeng> Or i misunderstood 17:11:17 <xionglingfeng> okok got u 17:11:17 <inc0> reconfig means I want to keep my containers as they are but change config files 17:11:28 <elemoine_> SamYaple: reconfiguration, sorry :) 17:11:37 <SamYaple> elemoine_: i meant for the name 17:11:43 <elemoine_> SamYaple: I agree with you 17:11:46 <sdake> reconfigure is a bit long but i dont care as long as it works ;) 17:11:54 <inc0> updtading_configuration_files_to_the_latest_version_based_on_what_is_currently_in_ansible_templates 17:11:57 <SamYaple> reload 17:12:04 <elemoine_> we don't have short names elsewhere 17:12:11 <SamYaple> i dont care, just the partital short name is annoying 17:12:16 <xionglingfeng> reconfig==write new version of configuration files and restart containers, isn't it? 17:12:21 <SamYaple> xionglingfeng: yes 17:12:28 <xionglingfeng> So the risk is? 17:12:46 <elemoine_> that it's not finished before mitaka I guess 17:12:46 <sdake> the risk is there is only one patch in the queue to do the job and it ahsn't been updatedd in 5 days 17:13:04 <sdake> and there needs to be a role for every service 17:13:16 <sdake> i'll speak to Jeffrey4l offline about it, he must be asleep ;) 17:13:27 <sdake> ok, any questions, or shall I move on? 17:13:28 <elemoine_> sounds worrying… 17:13:30 <SamYaple> sdake: its been updated... 17:13:36 <SamYaple> its fairly active right now 17:13:39 <sdake> SamYaple cool - i'll check it out 17:13:39 <SamYaple> looks good 17:13:49 <rhallisey> well once there's a template the rest should fly up there 17:13:51 <sdake> SamYaple I haven't looked at the queue this morning yet 17:14:01 <rhallisey> we told him to make sure one merges first 17:14:02 <xionglingfeng> hav to leave meeting conflict 17:14:06 <xionglingfeng> see u guys :-) 17:14:09 <rhallisey> see ya 17:14:09 <sdake> xionglingfeng ttyl 17:14:12 <SamYaple> thanks xionglingfeng 17:14:12 <elemoine_> xionglingfeng: bye 17:14:34 <sdake> #topic essential blueprint status 17:15:08 <sdake> https://launchpad.net/kolla/+milestone/mitaka-3 17:15:12 <sdake> #link https://launchpad.net/kolla/+milestone/mitaka-3 17:15:14 <xionglingfeng> BTW, I think I have some spare time working on both core components and trove/murano in the following months. So if anything I can help, just let me know 17:15:16 <xionglingfeng> :-) 17:15:34 <sdake> essential is looking good 17:15:43 <sdake> we have many in the not started state however 17:15:46 <sdake> so lets focus on those first 17:15:55 <sdake> mariadb lights out recovery 17:16:00 <elemoine_> xionglingfeng: trove logs are not collected 17:16:06 <sdake> SamYaple ^ 17:16:16 <sdake> also playbook for neutron playbook for openvswitch for upgrades ^^ 17:16:21 <elemoine_> xionglingfeng: it would be good to add that, and that should be pretty straightforward 17:16:26 <sdake> are those not started state as well, or are they started? 17:17:20 <SamYaple> they are started 17:17:20 <sdake> remember, we have 1 week ;) 17:17:25 <SamYaple> i need to push them up for review 17:17:31 <SamYaple> im testing multinode neutron 17:17:34 <sdake> cool, mind i set them to started? 17:17:36 <SamYaple> it _seems_ to be working 17:17:40 <sdake> if they are ready to review, then theyare probably good dprogress 17:17:55 <sdake> or ready for review 17:18:10 <SamYaple> yea in neutrons case its basic service upgrade template 17:18:14 <sdake> SamYaple would you mind setting the state for those 3 to what you think is appropriate? 17:18:17 <SamYaple> mariadb is harder, still idempoteizing it 17:19:27 <sdake> nihilifer you ahve several upgrade blueprints in started state 17:19:32 <sdake> I see they have reviews up 17:19:38 <sdake> and your blocked on multinode testing 17:19:55 <sdake> folks - please don't block on mulitnode testing just test single node and we can sort out multinode qa in the rcs 17:20:11 <sdake> iwith qualifier (if you can't test multinode) 17:20:34 <sdake> the goal is to get things from not started->started->good progress->ready for review->implemented 17:20:42 <sdake> I leave it to you to judge what state your work is in 17:21:02 <sdake> but I can't read your minds, so please set the tracker appropriately so operators don't have a heart attack when they look at our tracker 17:21:19 <sdake> upgrade playbook for heat, half done 17:21:34 <sdake> rest are all good progress 17:21:45 <sdake> I'd define good progress as *it will be implemented by march 2nd-4th* 17:22:02 <inc0> how are we doing on Grand-Backport? 17:22:11 <sdake> inc0 lets get mitaka-3 out of the way first 17:22:23 <sdake> I had planned to discuss high prioirrty blueprints 17:22:27 <sdake> but there isn't enough time 17:22:28 <sdake> so :) 17:22:32 <sdake> #topic open discussion 17:22:35 <sdake> inc0 you have the floor 17:22:52 <elemoine_> inc0: grand backport == backporting thin containers ? 17:23:04 <sdake> elemoine_ no the first part of the proposal 17:23:12 <vhosakot> I have something to talk in open discussion.. should we add docker version and ansible version as well in requirements.txt ? 17:23:12 <elemoine_> ah ok 17:23:13 <inc0> elemoine_ thin containers and named volumes 17:23:13 <sdake> the containers backport hasn't hit its voting deadline yet 17:23:25 <elemoine_> got it 17:23:27 <sdake> vhosakot one at a time :) 17:23:32 <inc0> sdake, but I think we have majority already 17:23:36 <sdake> whoever has the floor drives discussion in open discussion 17:23:48 <vhosakot> sdake: got it! 17:23:50 <sdake> inc0 cool wedidn't yesterday i'll check ml and make it so 17:24:24 <sdake> so inc0 anything to add re grand backport ? :) 17:24:27 <inc0> ok, I'd love to have 1.1.0 asap tbh...we need to really test this one 17:24:31 <rhallisey> ya I think we hit the majority 17:24:38 <inc0> its more important (way more important) than Mitaka-3 imho 17:24:45 <inc0> I'm done 17:24:54 <sdake> #action sdake to check for majority on thin container backport and post results on mailing list 17:24:56 <SamYaple> i agree inc0 17:25:19 <sdake> inc0 we need to finish the features for mitaka by eadline atleast essential/high 17:25:34 <sdake> I don't want features in the rcs if we can help it 17:25:46 <sdake> we have 80 bugs in the tracker, which is about 40 too many to handle in the rcs 17:26:02 <inc0> ok, but I want to say that while noone will deploy mitaka for next few months, 1.1.0 will get deployed as first kolla on prod 17:26:04 <sdake> vhosakot you have the floor 17:26:05 <sdake> 3 minutes ;-) 17:26:11 <vhosakot> sdake: thanks :) 17:26:13 <vhosakot> I have something to talk in open discussion.. should we add docker version and ansible version as well in requirements.txt ? 17:26:19 * SamYaple begs for reviews https://review.openstack.org/#/c/283887 17:26:21 <sdake> inc0 ack that, i think we need to do that during rcs 17:26:23 <SamYaple> that was holding up ceph upgrades ^ 17:26:35 <inc0> vhosakot, we are separating kolla-ansible from kolla 17:26:37 <sdake> SamYaple will review after meeting 17:26:54 <inc0> and we can't add docker version because it's not pip installable 17:27:13 <vhosakot> inc0: ah, ok.. agreed.. cool 17:27:15 <sdake> ansible can't be added to requiremetns.txt because of openstack policies 17:27:19 <sdake> there is a bug in the tracker on this piont 17:27:22 <inc0> that too 17:27:34 <sdake> i'll find it for you later if you want to read up on it 17:27:38 <inc0> ansible has wrong licence 17:27:39 <vhosakot> sdake: thats it.. I will ping in #kolla channel about a feature I need to start or help with 17:27:53 <SamYaple> ok im out guys 17:27:55 <SamYaple> lunch 17:28:01 <elemoine_> https://bugs.launchpad.net/kolla/+bug/1547533 17:28:01 <openstack> Launchpad bug 1547533 in kolla "Tech-Debt: Namespace the Docker volumes" [Medium,Confirmed] 17:28:04 <sdake> any other open discussion? 17:28:12 <elemoine_> would love a decision on this 17:28:19 <sdake> decision on what 17:28:24 <elemoine_> https://bugs.launchpad.net/kolla/+bug/1547533 17:28:27 <sdake> oh you rbug 17:28:35 <elemoine_> volume naming 17:28:39 <sdake> i'll tak a look elemoine_ 17:28:45 <sdake> I think we dont need namespacing on further thinking 17:28:55 <elemoine_> thanks, I've added a few ideas in the bug 17:29:07 <SamYaple> elemoine_: i know what youre asking, we can implement that in newton 17:29:13 <elemoine_> sdake: which means renaming kolla_logs to something else I guess 17:29:14 <inc0> if docker doesn't support volume rename 17:29:20 <SamYaple> mesos has this thing 17:29:25 <sdake> thanks for coming folks 17:29:26 <SamYaple> inc0: it doesnt need to 17:29:34 <sdake> our meeting time is up :) 17:29:41 <elemoine_> bye 17:29:45 <sdake> lets overflow into #kolla as needed 17:29:46 <vhosakot> bye :) 17:29:48 <sdake> #endmeeting