16:01:33 <sdake> #startmeeting kolla 16:01:34 <openstack> Meeting started Wed Aug 19 16:01:33 2015 UTC and is due to finish in 60 minutes. The chair is sdake. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:36 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:38 <openstack> The meeting name has been set to 'kolla' 16:01:40 <sdake> #topic rollcall 16:01:41 <SamYaple> first 16:01:53 <bmace> hi 16:02:02 <sdake> o/ - kolla rockin! 16:02:20 <rhallisey> hey 16:02:22 <jmccarthy> o/ 16:02:23 <jpeeler> hey 16:02:37 <inc0> afternoon/evening guys 16:02:40 <sdake> waiting 2 more mins for folks 16:03:04 <pbourke> hi 16:03:37 <SamYaple> lets make all the important decisions now 16:03:43 <sdake> #topic Announcements 16:04:16 <sdake> 1. There is an operator midcycle this week - lots of good news on kolla front - will get to that later in agenda 16:05:00 <sdake> 2. I kicked all "Not started" blueprints to rc1 which has a deadline of September 25th (HARD DEADLINE) - more on this in the meeting 16:05:12 <sdake> anyone else have announcements? 16:05:19 <inc0> meeting time? 16:05:30 <sdake> inc0 yup we are in the meeting now 16:05:42 <inc0> I meant to remind that we're moving 30mins forward 16:05:52 <sdake> oh right 16:06:12 <inc0> and remove GMT-night one 16:06:26 <sdake> 3. i am posting an email aftr this meeting to propose we move all meetings to 16;30 including our 22:00 meeting 16:06:35 <sdake> would appreciate fast feedback on that 16:06:53 <sdake> on the mailing list 16:06:59 <sdake> #topic gate status 16:07:16 <sdake> jpeeler has done a fantastic job of integrating build.py with the gate 16:07:30 <sdake> i've done some work to start building all combos of templates + docker + ubuntu + source + binary 16:07:34 <sdake> this is 8 gates total 16:07:40 <sdake> we can't do ubuntu binary 16:07:44 <sdake> so it is actually 6 gates 16:07:57 <pbourke> good job guys 16:08:06 <rhallisey> nice job jpeeler 16:08:08 <sdake> we only have 2 gates operatoinal at this time 16:08:15 <pbourke> really looking forward to having gates that actually... gate :) 16:08:17 <sdake> but the rest are coming shortly 16:08:33 <sdake> on the gates that actually gate front, we need a --retry because the mirrors re flakey 16:08:46 <inc0> whats plan of CI-checking that services actually run and work well? 16:08:46 <sdake> once the gate works reliably we can possibly make the gates voting 16:09:06 <sdake> but delorean has no mirros so if delorean repo goes down, our voting gate would block 16:09:22 <jpeeler> inc0: that's after image building gets sorted 16:09:22 <sdake> so mirrors of delorean will be the trigger point for voting gating (pus retry) 16:09:44 <sdake> ya after image building is working reliably we can tackle functional ci 16:10:03 <inc0> can we make sources builds block at least? 16:10:11 <sdake> i'd like to start off simple with an AIO deploy on centos binary/source that nova boots and checks the nova instance enters active state 16:10:25 <sdake> inc0 we pull openvswitch from delorean repos 16:10:33 <sdake> the only one we can make voting is ubuntu-source 16:10:36 <sdake> but i'd be game for that 16:11:03 <inc0> well centos-binary is our main distro still, so that wouldnt block too much 16:11:17 <SamYaple> sdake: ubuntu-source isnt ready 16:11:26 <SamYaple> you mean centos-source? 16:11:32 <sdake> yup its a problem that centos-binary is our main distro 16:11:43 <inc0> SamYaple, centos-source still uses delorean openvswitch 16:11:44 <sdake> centos-source also pulls from delorean repos for openvswitch 16:12:07 <inc0> question tho - won't this work without delorean up? openvswitch is in other repos too 16:12:12 <sdake> my action is to ride the delorean people to get them to ssetup mirrors 16:12:17 <SamYaple> ubuntu-source isn't ready though 16:12:30 <sdake> yes we wont make the gates voting untit he code works samyaple ;) 16:12:42 <SamYaple> ok just making sure everyone knows 16:12:46 <SamYaple> ubuntu isnt ready like at all 16:12:57 <SamYaple> im going through it now 16:13:31 <sdake> one thing i need sooner rather then later for gating is for setup_gate.sh to setup an ubuntu environment gate 16:14:10 <sdake> once the gating patches are complete in project-config, we are removing f21-functional gate job (build.sh gating) 16:14:27 <sdake> we also need --retry on the gates if someone wants to tackle that 16:14:51 <sdake> #topic Operator Midcycle 16:15:11 <sdake> #link #link https://etherpad.openstack.org/p/PAO-ops-meetup 16:15:16 <inc0> anyone from kolla is there? 16:15:20 <sdake> #link #link https://etherpad.openstack.org/p/PAO-ops-meetup 16:15:24 <sdake> #link https://etherpad.openstack.org/p/PAO-ops-meetup 16:15:28 <sdake> #undo 16:15:28 <sdake> #undo 16:15:28 <openstack> Removing item from minutes: <ircmeeting.items.Link object at 0xa4c2690> 16:15:29 <openstack> Removing item from minutes: <ircmeeting.items.Link object at 0xa4c25d0> 16:15:30 <sdake> #undo 16:15:31 <openstack> Removing item from minutes: <ircmeeting.items.Link object at 0xa4c2390> 16:15:33 <sdake> #link https://etherpad.openstack.org/p/PAO-ops-meetup 16:15:51 <sdake> #link https://etherpad.openstack.org/p/PAO-ops-containers-for-deployment 16:16:11 <sdake> #link https://etherpad.openstack.org/p/PAO-ops-upgrades 16:16:23 <sdake> inc0 operators are keen to use Kolla 16:16:37 <sdake> I wasn't there but got a report back 16:16:47 <inc0> I know...I pitched kolla a lot;) 16:16:57 <sdake> there was some discussion on osad vs kolla 16:17:18 <sdake> apparently immutability is huge with the operators and this was a big point on the osad vs kolla discussion in the midcycle 16:17:33 <inc0> we had good discussion about kolla in Vancouver ops summit 16:17:35 <sdake> kolla was well represented, people seemed super interested in what we havet o offer 16:17:36 <inc0> upgrades session 16:18:16 <sdake> since it takes us about a week to implement a new container + ansible code for it assuming there is no keystone domain work erequired, we scale extremely quickly 16:18:26 <sdake> the links above (for your later viewiing pleasure) 16:18:28 <sdake> key requirements 16:18:39 <sdake> need upgrade from one version to the next 16:18:59 <sdake> operators want to be able to run rabbitmq/galera on their own vms 16:19:07 <sdake> in some cases rather then in containers 16:19:21 <sdake> there was alot of quesitons about whether docker was production ready 16:19:28 <inc0> it's not a problem with a bit of tinkering of inventory file 16:19:36 <sdake> my answer to that is "for kolla use cases it absolutely is" 16:19:46 <sdake> for more advanced use casewss it is not 16:20:05 <SamYaple> inc0: no that isnt inventory file tinkering, that is enable_rabbitmq: no 16:20:18 <sdake> yup 16:20:24 <SamYaple> but we _can_ is the point 16:20:28 <sdake> I think we can support that use case easily 16:20:36 <SamYaple> if thats a major concern i can write some docs around it 16:20:46 <SamYaple> I wrote the playboosk to support just that 16:20:49 <inc0> tools being used are tripleo and ansible 16:20:57 <sdake> there wasn't alot of discussion about haproxy+keepalived being abstracted out into bare metal like rabbitmq/galera 16:20:57 <inc0> and we'll have both 16:21:11 <sdake> docs will help samyaple 16:21:34 <sdake> any questions? 16:21:39 <sdake> waiting for 30 seconds for irc lag :) 16:21:48 <harmw> hi guys! 16:21:53 <SamYaple> hey harmw 16:22:01 <inc0> now, that's a irc lag harmw :D 16:22:01 <SamYaple> we already discussed moving the time for next week 16:22:11 <SamYaple> so hopefully you wont get cut off anymore 16:22:18 <sdake> hey harmw 16:22:21 <harmw> please excuse me for beeing not as active past weeks 16:22:22 <sdake> harmw check the logs plz :) 16:22:26 <harmw> will do sdake 16:23:07 <sdake> #action all developres to review last two ops midcycle etherpads for ideas 16:23:15 <inc0> I'll talk with Will Auld (moderator of upgrades session) and ask if there is more to be done than things in etherpad 16:23:25 <SamYaple> namedropper 16:23:27 <SamYaple> boo 16:23:33 <sdake> #topic Liberty-3 planning 16:23:34 <SamYaple> :) 16:23:47 <inc0> we 16:24:00 <inc0> we're in middle of l3...is this a good time to plan stuff?;) 16:24:04 <sdake> #link https://launchpad.net/kolla/+milestone/liberty-3 16:24:12 <sdake> please everyone open that page up 16:24:20 <rhallisey> inc0, we're doing status 16:24:22 <sdake> i know clicking is hard ;0 16:24:33 <harmw> ok 16:25:10 <sdake> anything good progress, implemented, needs code review i'm going to skip 16:25:25 <sdake> we have a logging section in the agenda so skipping that one 16:25:49 <sdake> the goal of this session is to identify any bluepritns that wont make aug 31 16:25:56 <sdake> so they can be kicked to rc1 16:25:58 <SamYaple> ok sdake give me a second to update some blueprints 16:26:05 <SamYaple> go over other peoples first 16:26:09 <rhallisey> sdake, kick anything that isn't ceph for me 16:26:20 <sdake> rhallisey kicking 16:26:25 <sdake> 2 mins for people to update their status 16:26:31 <sdake> starting again at 9:29 16:26:36 <harmw> sad, thats easy for me 16:26:39 <harmw> :p 16:26:51 <inc0> any chance to bump ironic priority up? 16:27:21 <inc0> if we land this, we might actually impact tripleo heavily with a bit of luck 16:27:25 <jpeeler> inc0: i know you've pinged me a few times about ironic. i'm on it starting today 16:27:44 <jpeeler> already starting converting my old work to config-external 16:27:46 <inc0> jpeeler, anything we can help you with? 16:28:01 <inc0> I hate tripleo devtest, let's make it better 16:28:10 <SamYaple> sdake: was the core/contrib structure kciked? 16:28:18 <sdake> samyaple yes 16:28:25 <jpeeler> inc0: not yet, learning exercise for me for now since i've not done a new container in a while. but it's being worked on, so i'll let you know more later 16:28:27 <SamYaple> to Mitaka? 16:28:27 <sdake> we will get to rc1 next promise :) 16:28:32 <sdake> i think to rc1 16:28:39 <sdake> keep in mind we can make whatever changes we like 16:28:45 <inc0> appreciate that jpeeler 16:28:50 <sdake> lets focus on l3 for now plz :) 16:29:01 <pbourke> a quick question on the ansible swift bp 16:29:03 <sdake> rhallisey I am going to leave your bugs in liberty-3 ok? 16:29:17 <pbourke> how do people feel about us leaving it up to the op to provide the rings 16:29:18 <sdake> pbourke let me go through them in ordern ow 16:29:18 <rhallisey> sure 16:29:23 <pbourke> sdake: ah 16:29:31 <pbourke> sure 16:29:48 <sdake> ok 3 mins is up 16:29:57 <sdake> hopefully you have had a chance to udate all your blueprints 16:30:18 <sdake> if something is in good progress but it wont land until rc1 in compelted state lets keep it in l3 16:30:20 <sdake> and we can kick it at the end of the cycle 16:30:39 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/kolla-ansible-script 16:30:57 <sdake> diga is responsible for reporting status on this one 16:31:06 <sdake> since he isnt around i'll do that 16:31:18 <sdake> i'm not sure what else is needed here 16:31:22 <sdake> kolla-ansible deploy works ok 16:31:36 <SamYaple> sdake: it needs to be written in python 16:31:38 <sdake> I would appreciate moe discussion on the blueprint so we know hat people want out of it 16:31:42 <SamYaple> we talked about this in L2 16:31:47 <sdake> right 16:31:56 <sdake> can you add discussion to the blueprint on that point 16:32:00 <SamYaple> sure 16:32:03 <harmw> SamYaple: could you add the wy to it? 16:32:06 <sdake> what we wnt here i really undefined 16:32:24 <sdake> https://blueprints.launchpad.net/kolla/+spec/ansible-mongodb 16:32:29 <sdake> coolsvap|away is responsible for this one 16:32:37 <sdake> its a dependency of ceilometer 16:32:40 <sdake> it is not started 16:32:49 <sdake> he said he would knock it out this weekend along with ceilometer 16:32:51 <harmw> is the original even in a working state? 16:32:56 <sdake> harmw no idea 16:33:09 <harmw> (else he's in for more hard labor to get it even up and running) 16:33:35 <harmw> horizon feels like something that needs prio 16:34:18 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/ansible-horizon 16:34:28 <sdake> i just took this over from rhalllisey since he requested it be kicked 16:34:32 <sdake> there is a review up 16:34:44 <pbourke> is it still in wip? 16:34:48 <sdake> rhallisey mind giving an update on your WIP ansible code? 16:35:13 <rhallisey> I haven't looked at horizon since I thought you were going to take it 16:35:18 <sdake> pbourke yes I haven't touched it 16:35:27 <sdake> what state was it in then rhallisey ? 16:35:34 <SamYaple> hey guys im sorry I have to drop off here 16:35:34 <rhallisey> oh 16:35:46 <sdake> SamYaple understood 16:35:47 <rhallisey> basic infrastructure setup 16:35:48 <SamYaple> sdake was there anything else I was needed for? 16:35:49 <rhallisey> that's all 16:35:54 <SamYaple> ive got like 3 min 16:35:58 <sdake> the logging discussion and rc1 discussion :) 16:36:04 <SamYaple> :/ 16:36:06 <sdake> but we can tackle that next wek 16:36:13 <SamYaple> ok will ill read logs 16:36:16 <SamYaple> inc0 is leading that 16:36:18 <sdake> or the mailing lit 16:36:21 <sdake> ok 16:36:32 <inc0> we can discuss that tomorrow 16:36:36 <SamYaple> inc0 hit me up later if you haev questions 16:36:37 <harmw> regarding horizon, perhaps I can put in some effort there 16:36:39 <SamYaple> thanks guys im out 16:36:39 <sdake> so before weeks end i plan to have a horizon ansible code up and running 16:36:41 <sdake> or die trying 16:36:49 <rhallisey> SamYaple, see ya 16:36:56 <sdake> harmw if you want to take the wip blueprint that would rock - I am really overloaded 16:37:06 <sdake> but it absolutely needs to finish before the 31st 16:37:18 <sdake> all essential/high need to finish before the 31st plz :) 16:37:30 <harmw> it'd be realy sad if my contribution for L3 would be 0,0 16:37:37 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/replace-config-external 16:37:41 <sdake> harmw sometimes it happens 16:37:48 <harmw> lets not, ok :p 16:37:50 <sdake> harmw you are doing reviews, that is a contirbution 16:38:22 <harmw> sdake: I can't assign myself to it 16:38:43 <sdake> this blueprint is not started 16:38:52 <harmw> (horizon) 16:38:56 <sdake> the idea here is to use a json file to load into the container and remove config-eternal 16:39:02 <sdake> harmw i'll do after meeting ok? 16:39:05 <harmw> sure thing 16:39:18 <sdake> SamYaple if your still around any chance this woudl make liberty3? 16:39:43 <pbourke> does it need to? 16:39:45 <harmw> feels like a no, imho 16:39:56 <sdake> since this is in the not started state, I will speak to sam about it tomorrow when he returns 16:40:03 <harmw> it's a great feature, but I think there's higher prio work still to do 16:40:03 <sdake> and we may kick it to rc1 16:40:07 <harmw> ok 16:40:08 <pbourke> harmw: +1 16:40:09 <harmw> wfm 16:40:18 <sdake> i want the containers to be locked down after liberty 16:40:22 <sdake> as in the structure and config 16:40:28 <sdake> so we have an easy migration path 16:40:50 <sdake> so it is fairly high on my priority list 16:40:58 <sdake> but overall medium given our list of other work 16:41:04 <pbourke> fair point 16:41:07 <harmw> I'd say kickoff the last sprint with this one 16:41:11 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/run-data-container 16:41:18 <sdake> this is unowned 16:41:46 <sdake> harmw wfm since need sam's expertise lets make sure hes on board with that 16:41:55 <harmw> ofcourse :) 16:41:56 <sdake> the idea of this blueprint is to use a data container for /run 16:42:05 <sdake> for things like neutron and haproxy 16:42:10 <sdake> i dont think it can be done for nova 16:42:18 <sdake> this woudl be a breaking change to our containers 16:42:26 <sdake> so again another thing that needs to be locked down before t he release 16:42:35 <sdake> which is september 25th (rc1) deadline 16:42:50 <sdake> anyone interested in giving this a spin? 16:43:26 <sdake> ok moving to rc1 16:43:32 <harmw> sounds like it needs some more investigation, so perhaps settle in having someone do that and do the works in the next sprint 16:44:23 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/neutron-third-party-plugins 16:44:29 <sdake> note rc1 is not really a sprint 16:44:39 <sdake> its supposed to be bug fixes only :) 16:44:41 <harmw> true true 16:44:48 <sdake> and mandatory features we missed in l3 16:44:54 <sdake> pbourke mind giving an update on this one 16:45:07 <sdake> in other words, will it make aug 31st 16:45:22 <harmw> pbourke: you've got requests for specific plugins here (from ppl)? 16:45:22 <pbourke> i haven't looked at it for a little while, but afaik it just needs a restructure, should not be a big deal 16:45:27 <pbourke> harmw: yes 16:45:30 <harmw> cool 16:45:37 <harmw> $work, perhaps? 16:45:37 <pbourke> sdake: so yes should not be a problem for l3 16:45:49 <sdake> pbourke if you could get a reworked patch in the queue for debate that woudl help 16:45:58 <sdake> the last patch hasn't been touched since the beginning of l3 :) 16:46:07 <sdake> #link https://blueprints.launchpad.net/kolla/+spec/ironic-container 16:46:09 <pbourke> well, I said a while :p 16:46:29 <sdake> jpeeler mind giving us an update on ironic in container and ansibilizing the work 16:46:32 <sdake> it says started 16:46:54 <jpeeler> well it was started from june and then i had to put it down 16:47:12 <jpeeler> so as i was saying earlier, i have started porting my work over to current practices 16:47:42 <jpeeler> and i'll be focused on it now that the gate is running the new build job 16:47:51 <sdake> cool 16:47:59 <sdake> well when it gets to good progress pleae update the tracker 16:48:06 <jpeeler> will do 16:48:18 <sdake> i'm going to start pushing starteds next week 16:48:25 <inc0> jpeeler, if you need any help with ansible, ping me 16:48:27 <sdake> so we can focus on the blueprints we can finish 16:48:43 <jpeeler> inc0: alrighty! 16:48:48 <sdake> #topic liberty rc1 planning 16:48:50 <sdake> we are running short on time 16:48:59 <sdake> so I just want to point out a link 16:49:18 <sdake> #link #link https://launchpad.net/kolla/+milestone/liberty-rc1 16:49:18 <sdake> #undo 16:49:19 <openstack> Removing item from minutes: <ircmeeting.items.Link object at 0xa5f8510> 16:49:20 <sdake> #link https://launchpad.net/kolla/+milestone/liberty-rc1 16:49:46 <sdake> you will notice all blueprints here are in not started state 16:49:51 <sdake> they were originally in liberty3 16:50:08 <sdake> if you think something should be pushed to mitaka, be prepared next meeting to make the case :) 16:50:15 <harmw> I'm seeing two already started 16:50:18 <sdake> #topic templating 16:50:30 <sdake> #link #link https://etherpad.openstack.org/p/kolla-dockerfile-template 16:50:35 <sdake> harmw those i just pushed this meeting 16:50:42 <harmw> k 16:50:42 <sdake> ok so good news everyone :) 16:51:03 <sdake> the templating is complete - everything needed is in the review queue 16:51:18 <sdake> please prioritize these reviews today 16:51:26 <harmw> cool 16:51:32 <sdake> i think just nova is left for an ack 16:51:34 <sdake> but not certain 16:51:45 <inc0> how about I make logging stuff in template-only containers? 16:51:46 <sdake> so we have complete parity with docker and docker_templates 16:52:02 <sdake> inc0 logging is next up on agenda 16:52:05 <sdake> we will hae 5 minutes 16:52:07 <sdake> so for templating 16:52:12 <sdake> we need to nuke the docker directory 16:52:21 <inc0> I mean, do we accept new stuff in templating from start? 16:52:27 <sdake> inc0 yes 16:52:28 <inc0> without "normal" Dockerfiles? 16:52:30 <sdake> docker is deprecated 16:52:34 <inc0> ok 16:52:39 <sdake> the docker directory is deprecated now 16:52:42 <sdake> as of today 16:52:55 <inc0> that's what I wanted to hear (read) 16:53:00 <pbourke> +1 16:53:04 <harmw> sdake: is that in the docs already? 16:53:05 <sdake> i am going to make a blueprint for removal of the docker directory 16:53:18 <sdake> harmw no - needs docs changes 16:53:20 <sdake> harmw it is in some of the docs actually 16:53:22 <sdake> the README.md 16:53:26 <rhallisey> uh hold on.. 16:53:27 <sdake> iirc 16:53:34 <harmw> I mentioned doing that in some of the changes made in docs recently 16:53:42 <harmw> ok sdake 16:53:44 <inc0> what's up Ryan? 16:54:11 <harmw> building up suspense.. 16:54:27 <jpeeler> ha 16:54:31 <harmw> *drumroll* 16:54:46 <rhallisey> I just want to make sure we can consume these 16:54:47 <sdake> rhallisey speak or i'll move on :) 16:54:59 <sdake> rhallisey i will put a blueprint up in the discussion phase 16:55:00 <sdake> for rc1 16:55:03 <inc0> rhallisey, you mean tripleo? this shouldn't affect tripleo 16:55:11 <rhallisey> ya I haven't looked that them at all 16:55:14 <inc0> since its about building dockers 16:55:19 <inc0> not deployment 16:55:22 <sdake> rhallisey we hae cmplete parity with docker so tripleo is unaffecte 16:55:23 <jpeeler> right 16:55:23 <sdake> d 16:55:26 <rhallisey> I just want to be sure since we're using docker dir atm 16:55:56 <inc0> rhallisey, do you build images docker images in tripleo? 16:56:24 <rhallisey> we aren't, but we might 16:56:32 <sdake> #action sdake to make blueprint to remove docker directory 16:56:39 <inc0> I don't think we should tho 16:56:40 <sdake> if you aren't then you can use the docker templates directory and build.py 16:56:45 <rhallisey> kk 16:57:02 <sdake> i dont want to maintain two docker build systems or sets of files 16:57:08 <rhallisey> oh ya agreed 16:57:11 <inc0> let's just teach tripleo to consume registry images, and let build.py do its templating magic 16:57:25 <sdake> ok logging - unfortunately we are out of time 16:57:26 <rhallisey> inc0, I think that's likely the case 16:57:34 <sdake> #topic open discussion 16:57:37 <rhallisey> just was bringing it up 16:57:40 <rhallisey> that's all 16:57:40 <sdake> 3 minutes - sorry for short amount of time 16:57:50 <inc0> sdake, I'll just prepare prototype of loggin and we'll discuss it next week 16:58:19 <inc0> rsyslog container is inprog 16:58:22 <harmw> regarding logging, I know putting logs in a datacontainer makes us able to cater to everyones needs 16:58:23 <sdake> inc0 sounds good 16:58:38 <sdake> yes i'd like one data container for logs if possible per node 16:58:39 <harmw> but docker can log using fluentd, which talks logstash native 16:58:48 <harmw> perhaps nice to look into, some day 16:59:11 <sdake> and to put the log files in /var/log 16:59:14 <inc0> let's discuss it next week, I have few ideas I want to try out 16:59:18 <sdake> cool 16:59:20 <harmw> sounds like a plan 16:59:25 <sdake> have discussion on the logging blueprint pls 16:59:33 <sdake> that way there is a referencable recor 16:59:39 <sdake> if you want, create an etherpad for discussion 16:59:42 <sdake> and link it to the blueprint 16:59:48 <sdake> or link blueprint to it i mean 16:59:55 <sdake> ok we are outof time, we can overflow in #kolla 16:59:58 <sdake> #endmeeting