16:30:11 <sdake> #startmeeting kolla
16:30:13 <openstack> Meeting started Wed Mar  9 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.
16:30:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:30:17 <openstack> The meeting name has been set to 'kolla'
16:30:19 <sdake> #topic rollcall
16:30:22 <akwasnie> hi
16:30:31 <sdake> \o/
16:30:32 <dave-mccowan> \o/
16:30:36 <nihilifer> hi
16:30:43 <pbourke> o/
16:30:46 <ajafo> o/
16:30:47 <vhosakot> o/
16:30:51 <britthouser> 0/
16:30:55 <Jeffrey4l> \o/
16:31:05 <openstack> Jeffrey4l: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
16:31:21 <sdake> Jeffrey4l I already did
16:31:28 <jpeeler> hi
16:31:57 <Jeffrey4l> ( i have a very bad network now. sorry)
16:32:13 <Jeffrey4l> saw that.
16:32:15 <SamYaple> o/
16:32:17 <sdake> #topic announcements
16:32:32 <inc0> o/ sorry I'm late
16:32:39 <sdake> 1. akwasnie was unanimously voted into the core review team
16:32:40 <elemoine_> (/
16:32:42 <sdake> yay :)
16:32:47 <elemoine_> o/
16:32:59 <nihilifer> congrats!
16:33:01 <elemoine_> congrats akwasnie
16:33:03 <inc0> congrats:)
16:33:08 <akwasnie> thank you all!
16:33:26 <sdake> 2. the backport upgrades plan was approved by a majority of the core reviewers
16:33:41 <sdake> 3. I added our plan and a warning not to use liberty to the documentation
16:34:05 <sdake> #link http://docs.openstack.org/developer/kolla/liberty-deployment-warning.html
16:34:26 <sdake> 4. akwasnie and inc0 and elemoine_ hae a talk accepted on diagnostics
16:34:34 <sdake> please attend if you can make it ;)
16:34:35 <coolsvap> hello
16:34:48 <sdake> sdake and inc0 got jammed into what may be an upgrade panel
16:35:07 <elemoine_> nice
16:35:08 <sdake> it is unclear what will be happening with this talk as there are 9 people reigstered for it
16:35:38 <inc0> and topic is....not what we intended to talk about
16:35:38 <sdake> ok thats about it from me
16:35:47 <sdake> anything from the community?
16:35:49 <sdake> yes topic is blue/green upgrades
16:36:07 <sdake> for those that don't know, that is having two datacenters and switching them via routing after an upgrade
16:36:17 <sdake> as coreos does with distro upgrades
16:36:26 <sdake> we are going to either get that changed or I wont be at that talk :)
16:36:47 <sdake> #topic Dependencies inside one host in kolla-mesos
16:36:48 <sdake> this rolled over from lsat week
16:36:57 <sdake> I'm not sure who added it to agenda, but whoever did, you ahve the floor
16:37:00 <sdake> was it you nihilifer ?
16:37:10 <nihilifer> no, it wasn't me
16:37:15 <nihilifer> kproskurin: ?
16:37:17 <sdake> SamYaple ?
16:37:53 <sdake> well I dont know what the purpose of this one is, so I'll move on if nobody steps up to lead the discussion :)
16:37:54 <rhallisey> hello
16:37:59 <rhallisey> sry I'm late
16:37:59 <sdake> hey rhallisey
16:38:10 <nihilifer> ok, i think we may go on to the other topic
16:38:18 <sdake> #link Kolla Austin Summit discussion [timebox 5 minutes]
16:38:21 <nihilifer> this is already implemented and has +2 ;)
16:38:22 <sdake> #undo
16:38:23 <openstack> Removing item from minutes: <ircmeeting.items.Link object at 0xaeb6910>
16:38:25 <sdake> #topic Kolla Austin Summit discussion [timebox 5 minutes]
16:38:39 <sdake> we have 24 proposed sessions
16:38:56 <sdake> if you add sessions, please add them at the end, not in the middle
16:38:57 <elemoine_> link?
16:39:14 <sdake> 1 of them had no votes, 2 sets of 2 sessions were merged
16:39:32 <sdake> #link https://etherpad.openstack.org/p/kolla-newton-summit
16:39:41 <elemoine_> thanks
16:39:45 <inc0> let's vote for them...then sum up votes and calculate average, whoever gets 9.2 wins
16:39:47 <sdake> The most slots we can hvae are 16
16:39:56 <sdake> 9.2 mlilion!
16:40:30 <sdake> I asked for a minimum of 4 fishbowl and 10 design sessions
16:40:39 <sdake> and a full day contributor meetup on friday
16:41:04 <sdake> I asked for a maximum of 6 fishbowl and 17 dessign sessions with a half day contribtor meetup on friday
16:41:38 <sdake> we will priortize the talks going forward and I'll put the agenda together - anything that doens't make tue-thur will flow into friday
16:41:51 <sdake> we don't have 16 confirmed spots yet
16:41:58 <sdake> so we may get less then our minimum ask
16:42:01 <sdake> any questions?
16:42:24 <elemoine_> what are the goals of the design discusions?
16:42:35 <elemoine_> I mean
16:42:35 <sdake> designing
16:42:44 <elemoine_> is this just discussions?
16:42:48 <sdake> elemoine_ have youbeen to a summit before?
16:42:53 <elemoine_> no :)
16:42:58 <elemoine_> hence my question
16:43:01 <sdake> ok well its nothing to be worried about
16:43:09 <sdake> its design discussions about how the implementation should work
16:43:13 <pbourke> are we voting? what are [F] / [D]
16:43:13 <sdake> we take copious notes
16:43:17 <sdake> look at the midcycle notes
16:43:28 <rhallisey> pbourke, fishbowl & design
16:43:30 <sdake> pbourke those are vote counts F=fishbowl d=design
16:43:30 <inc0> pbourke, fishbowl and design
16:43:31 <akwasnie> pbourke: fishbowl, design
16:43:34 <ajafo> pbourke: on top is explanation [d][f]
16:43:38 <sdake> please don't vote twice ;)
16:43:39 <pbourke> thanks
16:43:51 <inc0> elemoine_, sitting around the table and figuring out how to fix stuff together:)
16:44:11 <elemoine_> ok
16:44:12 <sdake> stuff wlth low votes will flow over, i don't think we will miss any agenda topics unless there were zero votes
16:44:25 <elemoine_> I guess it can be related to future work we'd like to propose
16:44:47 <sdake> elemoine_ if you hae sessions you wnt o propose add them after 24
16:44:57 <sdake> if someone messes with the numbering my brain will explode :)
16:45:17 <sdake> #topic Mitaka rc1 blueprint review [timebox 15 minutes]
16:45:43 <elemoine_> diagnostics is what I'm interested in the most, and it's there already
16:45:47 <sdake> oh one more announcemnet
16:45:48 <sdake> we released mitaka-3 on the 5th
16:45:50 <sdake> grats everyone!
16:46:00 <sdake> nice work
16:46:07 <sdake> #link https://launchpad.net/kolla/+milestone/mitaka-rc1
16:46:17 <sdake> now its time to get to work on rc1
16:46:29 <sdake> I sent an email to the mailin list this mornin explaining how the rc schedules and priorties work
16:46:58 <sdake> lets go through the blueprints now
16:47:10 <sdake> since those need to be done before the 16th IIRC
16:47:16 <inc0> mariadb upgrade needs love
16:47:26 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/mariadb-lights-out owned by SamYaple
16:47:29 <inc0> who have spare love to give to it?
16:47:47 <sdake> inc0 lets go in order ok :)
16:47:51 <inc0> ok
16:47:54 <sdake> so this one isn't started
16:47:54 <inc0> sorry
16:48:02 <sdake> no reason to apologize
16:48:22 <sdake> SamYaple has 5 blueprints for rc1
16:48:36 <sdake> plus he is taking ressponsiblity for backporting liberty (thanks sam!)
16:48:41 <sdake> anyone want to take this oen off his hands?
16:49:04 <coolsvap> i can try but I am not completely sure about it
16:49:15 <SamYaple> ive got mariadb, even if its slow going
16:49:18 <SamYaple> itll be done
16:49:25 <dave-mccowan> i've got cycles (and WIP) for Fernet
16:49:27 <sdake> SamYaple sounds good
16:49:44 <SamYaple> fernet isnt going to happen
16:49:47 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/upgrade-rabbitmq
16:49:50 <SamYaple> oh i see dave-mccowan
16:49:58 <SamYaple> i dont think ferenet is going to happen dave-mccowan
16:50:07 <SamYaple> it requires stuff we just dont have yet, scripts playbooks
16:50:09 <SamYaple> its not just configs
16:50:14 <SamYaple> but i wont stop you for doing it
16:50:22 <SamYaple> i thin kthere are more important thigns to do though
16:50:24 <sdake> SamYaple dave already hs an implementation :)
16:50:47 <sdake> SamYaple would you mind giving that to dave then?
16:50:53 <dave-mccowan> everything but the rotation, which is the only hard part.
16:50:55 <SamYaple> id love to see it, and how he is rotating the tokens
16:50:59 <SamYaple> yes that was my point
16:51:11 <sdake> dave-mccowan can you take that one then in launchpad please
16:51:13 <SamYaple> im not sure we can squeeze that in
16:51:28 <sdake> well we wont merge it until its all complete in one big patch series, agreed?
16:51:32 <dave-mccowan> it won't be worthy to become default for sure
16:51:32 <SamYaple> external-deps are not likely to be eradicated this cycle
16:51:50 <sdake> should I bounce externl deps to n1?
16:52:04 <SamYaple> you can untarget it, im not sure when it would happen
16:52:18 <SamYaple> we need infra to do things so who knows
16:52:21 <sdake> i'll put it in n1 then and we can decide about newton priorities at summit
16:52:23 <SamYaple> they are working on them
16:52:31 <sdake> sounds good
16:52:53 <SamYaple> just back to fernet for a minute dave-mccowan
16:52:53 <sdake> inc0 did I hear you need hlep with rabbitmq?
16:53:03 <inc0> sdake, no, I'll finish it up
16:53:09 <SamYaple> the issue is we have no way to reinject new rotated tokens to running containers
16:53:10 <sdake> inc0 sounds good
16:53:18 <inc0> implementation is in review, only thing was backport voting
16:53:19 <SamYaple> and im not sure thats somethign we want to shove in, default or not
16:53:34 <inc0> which passed, so I'll just write this damned script in container:P
16:53:47 <sdake> inc0 :)
16:53:53 <SamYaple> inc0: feel free to change it for the better first thing newton
16:53:58 <sdake> inc0 thanks for that :)
16:54:02 <SamYaple> i agree its not amazing
16:54:15 <inc0> oh I will change it
16:54:17 <inc0> :P
16:54:25 <SamYaple> we need to change it in alot of places
16:54:31 <inc0> yeah
16:55:02 <inc0> I have other ideas as well, but that's for different discussio
16:55:03 <inc0> n
16:55:09 <SamYaple> sounds good
16:55:54 <sdake> pbourke around?
16:56:00 <pbourke> sdake: yup
16:56:06 <sdake> oracle linux gates on track?
16:56:22 <inc0> coolsvap, so, you're taking on mariadb? or anyone else wants to give it a shot?
16:56:40 <akwasnie> mariadb upgrade?
16:56:40 <sdake> inc0 SamYaple said he woudl do it although it would be slow going
16:56:50 <sdake> mariadb upgrade and mariadb lights out
16:56:54 <sdake> both need to be tackled
16:56:59 <sdake> we can probably split that up
16:57:07 <pbourke> sdake: they are. source are passing, binary deploys are failing for unknown reason, binary builds just need a few dependencies fixed
16:57:14 <inc0> akwasnie, yeah
16:57:15 <coolsvap> SamYaple: i can provide a helping hand
16:57:32 <akwasnie> I have to investigate predefined dashboard for kibana but can take also mariadb upgrade if you need any help
16:57:41 <inc0> SamYaple, coolsvap akwasnie yeah, maybe it would be good to spread the load
16:57:46 <SamYaple> mariadb upgrade and lights out are the same thing
16:57:47 <sdake> pbourke cool if you wnat me to help you debug the gate, point me at a gate log
16:57:48 <inc0> timing is short
16:57:49 <SamYaple> same mechanism needed
16:58:11 <inc0> SamYaple, maybe you could just describe mechanism and let coolsvap and akwasnie deal with it?
16:58:14 <pbourke> sdake: great thanks - nothing obvious as to why the binary deploy is failing to spin up an instance but I'm keeping an eye on it
16:58:17 <sdake> ya we have about 10 days
16:58:29 <SamYaple> inc0: i dont see how thats helpful since im working on it
16:58:41 <SamYaple> mariadb isnt sitting, it just needs lots o' testing
16:59:00 <sdake> at this point i'm ready to bounce the functional gating to newton
16:59:04 <inc0> ok, your call man, just holler if you need help
16:59:07 <SamYaple> will do
16:59:09 <sdake> people can work on it at their liesure
16:59:19 <sdake> i work on it occasionally
16:59:31 <sdake> one thing I'm working on is getting our security story in shape
16:59:42 <sdake> ok so onto the bug reports
16:59:58 <sdake> critical = feature that didn't land in mitaka-3
17:00:03 <sdake> work on those first if you hae spare cycles
17:00:17 <sdake> high = a big bad bug need fixing
17:00:23 <sdake> medium = annoying bug
17:00:32 <sdake> low = probably not worth bothering about this early in the rc cycle
17:00:47 <sdake> wishlist = docs or other features of low interest
17:00:47 <sdake> docs are of high interest
17:00:49 <sdake> just to clarify
17:01:02 <sdake> anything we can do to improve the docs improves our capacity
17:01:11 <sdake> because  we dont hae to spend hours on irc walking people through kolla deploys
17:01:23 <vhosakot> agreed
17:01:23 <sdake> so if you see a repeated quetison try to answer it in the quickstart
17:01:47 <sdake> we have 72 bugs
17:01:52 <sdake> and 9 blueprints
17:01:56 <sdake> lets tyr to knock out he lbueprints
17:02:00 <sdake> I hae alot of spare capacity
17:02:03 <sdake> someone put me to work ;)
17:02:24 <sdake> at this point i'm going to be working on making heat non-doa
17:02:28 <sdake> and then making magnum non-doa
17:03:00 <sdake> ok I guess we covered two agend topics in one topic :)
17:03:04 <sdake> so onto the next one
17:03:15 <sdake> #link Liberty backport planning
17:03:18 <sdake> #undo
17:03:20 <openstack> Removing item from minutes: <ircmeeting.items.Link object at 0xae36a50>
17:03:21 <sdake> #topic Liberty backport planning
17:03:25 * sdake hates computers
17:03:39 <sdake> I set a deadline of april 1st for our rc1 of liberty backports
17:03:47 <sdake> i'm not sure more then oen person can work on this at a time
17:04:01 <sdake> and that one person at this point is sam
17:04:17 <sdake> so anythign you can do to take stuff off sam's plate is a win for liberty backporting
17:04:28 <sdake> our liberty 1.1.0 release is april 15th
17:04:35 <sdake> these dates can slip
17:04:38 <sdake> nobody holds us accountable
17:04:47 <sdake> but we need to ship before summit!!
17:04:53 <sdake> oh also, we are publishing to pypi now
17:04:59 <sdake> thanks to a recent gate job change
17:05:19 <sdake> SamYaple do you need nay type of help for liberty backporting?
17:05:29 <sdake> nay/any
17:05:30 <SamYaple> yea there are patches in the queue
17:05:36 <sdake> you need them reviewed?
17:05:45 <SamYaple> reviews would be good
17:05:52 <SamYaple> they will not pass gate, but thts fine
17:05:52 <sdake> folks please review those patches
17:05:58 <sdake> ya, the deal wiht the gate
17:06:00 <SamYaple> we agreed on one patchstream with final patch passing gate
17:06:02 <sdake> we agreed early on
17:06:10 <sdake> the last patch in the patch will ben green
17:06:17 <sdake> and that is good enough
17:06:19 <sdake> patch stream
17:06:21 <sdake> ya what SamYaple said
17:07:00 <sdake> SamYaple if yu wnat ot give mariadb up it may make for a better opportunity to handle the backporting
17:07:25 <sdake> hard to do two major projects at one time i na constrained time period
17:08:02 <sdake> yay or nay?
17:09:10 <sdake> well SamYaple must be afk or laggy
17:09:17 <sdake> #topic open discussion
17:09:26 <sdake> any topics folks want to discuss?
17:09:51 <vhosakot> are the droproot bugs high pri ?
17:09:57 <sdake> yes
17:10:11 <vhosakot> will pick one of them then
17:10:18 <sdake> upgrades, droproot, and reconfigure
17:10:21 <vhosakot> I'm done testing rabbit reconfig
17:10:26 <sdake> and of course diagnostics
17:10:27 <vhosakot> will send review today
17:10:57 <sdake> at thsi point diagnostics framework is done,but i'd like to see a dashboard json-ification of kibana
17:11:00 <sdake> so its actually usable ;)
17:11:36 <akwasnie> I will look at this sdake
17:11:37 <sdake> akwasnie elemoine_ i'd recommend working on that if you dont have other bugs to work on
17:11:52 <sdake> nice
17:11:56 <akwasnie> :)
17:12:22 <sdake> any other topics?
17:12:28 <elemoine_> I'll see what I can do
17:12:28 <sdake> before we close out then
17:12:32 <sdake> I'd just like to say
17:12:37 <elemoine_> but happy to help akwasnie
17:12:40 <sdake> the kolla engineering team absolutely rocks
17:12:41 <nihilifer> yes, regarding droproot stuff - i have haproxy on myself, but SamYaple shared a though with i agree with - so that dropping root when listering on 80 port is impossible
17:12:49 <sdake> we have done a kickass job in mitaka-3
17:12:54 <sdake> and mitaka is going to be fantastic
17:13:08 <sdake> nihilifer right it is
17:13:24 <sdake> unless you give it capabilties
17:13:28 <sdake> which we can do with docker 1.10
17:13:37 <sdake> you need NET_CAP_SYSADMIN
17:13:47 <sdake> or soething like that
17:13:57 <nihilifer> ok, that might be an option
17:14:05 <sdake> its the only way it can be done
17:14:14 <nihilifer> i thght about the other thing
17:14:43 <SamYaple> just fyi, haproxy is meant to run as root
17:14:44 <nihilifer> if horizon/apache will listen on some other port
17:14:51 <SamYaple> so is apache2
17:14:55 <SamYaple> it drops privs itself
17:14:59 <sdake> apache2 definately is
17:15:05 <sdake> i dont know anything about haproxy
17:15:07 <SamYaple> haproxy is as well
17:15:11 <sdake> ok
17:15:15 <SamYaple> it has mechanisms to jail itslelf
17:15:27 <sdake> jails in jails :)
17:15:40 <sdake> so what we need is for it to jail?
17:15:47 <sdake> and thats a config feature?
17:15:49 <SamYaple> just config options
17:15:59 <sdake> there ya go nihilifer
17:16:17 <sdake> although i'd try drop root with capabiltieis as well as a better option
17:16:31 <nihilifer> ok, will try then
17:16:41 <SamYaple> sdake: its a worse option
17:16:41 <sdake> whatever fits the deadline
17:16:45 <SamYaple> alot less secure
17:16:54 <sdake> ok well lets go for more security
17:16:54 <dave-mccowan> nihilifer check out chroot command at haproxy.org
17:17:19 <sdake> dave-mccowan can consult here, he has expertise in this area
17:17:41 <sdake> I hope people are getting pumped up for austin
17:17:47 <sdake> I can't wait to see the town
17:17:50 <sdake> i hear it rocks :)
17:17:55 <vhosakot> SamYaple: Jeffrey4l: I have a point to talk about rabbitmq reconfig (that does not have section-based *.ini config file that can be merged using Python's ConfigManager)
17:17:59 <sdake> or city rather
17:18:05 <vhosakot> sdake: sorry to interrupt :)
17:18:17 <sdake> vhosakot its all good
17:18:19 <vhosakot> ConfigParser*
17:18:51 <SamYaple> vhosakot: its not just rabbitmq, its all non-ini configs
17:18:58 <sdake> vhosakot I think SamYaple is busy - perhaps you can sync up late with jeffrey4l
17:19:04 <sdake> lets conclude the meeting
17:19:10 <SamYaple> it needs an action-plugin to be done correctly
17:19:10 <sdake> 12 minutes of your life back ;)
17:19:21 <sdake> thanks everyone for all the hard work up to this point ;)
17:19:28 <SamYaple> sdake im still talking here
17:19:32 <SamYaple> why are we ending early?
17:19:32 <sdake> ok
17:19:33 <sdake> sorry
17:19:45 <inc0> more action plugins \o/
17:19:47 <sdake> it must be lag SamYaple
17:19:47 <vhosakot> SamYaple: right, so for such services, do we do a straight copy-paste and document that the user must update _entire_ file, not just parts like reonfig for *.ini services ?
17:19:55 <SamYaple> inc0: exactly
17:20:09 <SamYaple> vhosakot: it would be the entire file the way i have in mind
17:20:23 <SamYaple> so the user would need to copy out the emplate _then_ they could augment it
17:20:46 <SamYaple> the question is is this a required _required_ thing for mitaka?
17:20:49 <vhosakot> SamYaple: why even use action_plugin ? why not copy: module ?
17:21:14 <sdake> reconfig of eveything is realy realy nice ot have
17:21:17 <SamYaple> vhosakot: because we want copy_if_exists module and that can only be dont if its an action plugin
17:21:20 <sdake> required is a strong word
17:21:40 <vhosakot> SamYaple: so, I will submit rebbit reconfig work without action_plugin part (like Jeffrey4l did for horizon) and we cann  add non-ini action_plugin later
17:21:41 <SamYaple> well the action_plugin would be like a scaled down merge_configs plugin
17:21:41 <sdake> but if vhosakot has nothing else to work on might as well do that
17:22:14 <SamYaple> if we are ok with adding it, i would be ok because ive done alot to ensure merge_configs is 2.0 ready. it needs only 2 changes to work with 2.0
17:22:40 <vhosakot> cool, I will submit without action_plugin for now
17:22:50 <SamYaple> inc0: thoughts on above?
17:23:14 <inc0> SamYaple, you know where I stand, I think we should do action_plugins at once
17:23:23 <inc0> things are broken now
17:23:25 <inc0> without them
17:23:34 <inc0> but I'm ok with moving it for newton
17:23:35 <SamYaple> things are not broken without them
17:23:47 <SamYaple> things are harder without them
17:23:53 <inc0> ok, broken is wrong word, terribly ugly and hard to maintain
17:23:55 <SamYaple> and ansible doesnt do well with harder
17:23:57 <SamYaple> agreed
17:24:08 <SamYaple> (unless its a script in contianer, thats clean)
17:24:13 <vhosakot> yes, I see nothing broken without non-ini action_plugin.. just reconfig of non-ini
17:24:22 <SamYaple> ansible + multitask is bad news
17:24:54 <vhosakot> anyway, I will submit my rabbot reconfig work and get it merged before deadline... we cann add non-ini action_plugin later for all non-ini services
17:25:05 <SamYaple> sounds good vhosakot
17:25:10 <vhosakot> thanks!
17:25:29 <vhosakot> SamYaple: isn;t action_plugin wrapper about Python's native ConfigParser ?
17:25:44 <vhosakot> SamYaple: that copies files
17:27:39 <SamYaple> vhosakot: yea but thats not the hard part
17:28:00 <vhosakot> SamYaple: ah ok
17:29:06 <SamYaple> i think we are good here to wrap up
17:29:16 <sdake> and also out of time
17:29:19 <vhosakot> cool, bye everyone!
17:29:19 <sdake> thanks for coming folks :)
17:29:24 <inc0> thanks everyone
17:29:25 <akwasnie> thanks
17:29:34 <sdake> #endmeeting