20:00:25 <johnsom> #startmeeting Octavia
20:00:26 <openstack> Meeting started Wed Apr 26 20:00:25 2017 UTC and is due to finish in 60 minutes.  The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:28 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:31 <openstack> The meeting name has been set to 'octavia'
20:00:31 <johnsom> Hi folks
20:00:32 <xgerman> o/
20:00:33 <JudeC> o/
20:00:37 <nmagnezi> o/
20:00:59 <johnsom> #link https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda#Meeting_2017-04-26
20:01:27 <johnsom> Hopefully rm_work can join
20:01:52 <johnsom> #topic Announcements
20:02:15 <johnsom> The foundation lost our e-mail about the mascot....
20:02:41 <johnsom> They are concerned that two other groups now have trees
20:02:45 <johnsom> #link https://etherpad.openstack.org/p/octavia-mascot
20:03:10 <johnsom> So we need to talk about picking something else or pushing back to have them focus no the roots.
20:03:42 <johnsom> We can talk about that in a section below
20:04:15 <johnsom> There is a patch up to start switching to using systemd by default in devstack instead of screen.
20:04:37 <johnsom> You may see some changes to the gates/logs due to that.
20:05:00 <johnsom> #link https://review.openstack.org/#/c/460062/
20:05:41 <nmagnezi> what's the incentive?
20:05:58 <xgerman> shiny
20:05:59 <johnsom> There is also a discussion about log message IDs starting up again.  See Doug's e-mail on the dev list if you are interested in jumping into that.
20:07:03 <johnsom> nmagnezi From what I have seen there is a perception that it is more reliable, some folks like the systemd log system better (I'm not so sure myself), and they like being able to do group process restarts
20:07:09 <johnsom> +shiny
20:07:45 <johnsom> I just thought I should mention those items.
20:07:45 <nmagnezi> not sure how I like it for development env.. but..
20:07:50 <nmagnezi> i assume shiny as well
20:07:52 <johnsom> The summit is also just a few weeks out
20:08:24 <johnsom> Any other announcements?
20:08:33 <johnsom> The TC vote is done and posted as well.
20:08:39 <nmagnezi> I was just curious, didn't mean to interrupt. please go ahead :)
20:08:58 <johnsom> nmagnezi No worries
20:09:24 <johnsom> #topic Brief progress reports / bugs needing review
20:09:53 <johnsom> A lot has been going on over the last week.
20:10:43 <johnsom> I did take some time out to work on slide decks for the summit.  101 session is in progress and the project update for has been submitted to the foundation, so we should get a basic slide deck back from that today or tomorrow.
20:11:09 <johnsom> Any other updates from folks?
20:11:18 <johnsom> Patches that need attention?
20:11:28 <jniesz> will try to have patch set two for the L3 active/active spec up by end of week
20:11:32 <jniesz> based on your feedback
20:11:38 <johnsom> I finished testing out the proxy protocol patch (it still needs some work)
20:12:13 <johnsom> I will move on to sorting/paginiation again, then single-call-create, and finally back to api-ref
20:12:20 <johnsom> jniesz Sounds good
20:12:44 <johnsom> JudeC How is the OpenStack Client plugin coming?
20:13:08 <johnsom> I plan to try to get at least the listener bugs up today
20:13:19 <JudeC> I am getting through it as fast as I can, about half way done. I have someone else who is helping me on it too so things should speed up.
20:13:45 <johnsom> Cool!  Let us know when there are patches ready for testing/review.
20:13:56 <johnsom> Sounds like I need to get busy on those bugs too...  grin
20:13:59 <JudeC> johnsom that would be great getting the listener stuff would be very helpful
20:14:28 <johnsom> Ok
20:14:49 <johnsom> #topic Boston summit planning
20:15:01 <johnsom> Any thing we need to discuss on that this week?
20:15:43 <johnsom> rm_work xgerman ????  grin
20:15:58 <xgerman> I like to have the TLS bug fixed so we can demo that
20:16:01 <johnsom> Maybe that is a no...
20:16:05 <xgerman> also need to record some demos
20:16:12 <johnsom> Right, I should test that.
20:16:35 <xgerman> I think I promised some OpenStack Ansible stuff
20:16:39 <johnsom> Yeah, I plan to also record some demos for the 101 session.  Horizon for example
20:16:49 <xgerman> +1
20:17:35 <johnsom> #topic Mascot
20:17:44 <johnsom> #link https://etherpad.openstack.org/p/octavia-mascot
20:17:54 <johnsom> The is the brainstorming etherpad we had
20:18:20 <xgerman> pick 32
20:18:22 <johnsom> I think peacock was our second choice after a tree with roots
20:18:24 <xgerman> pick #2
20:18:33 <xgerman> peacock ;-)
20:18:40 <johnsom> Any comments or ideas?
20:18:49 <xgerman> I like peacock…
20:19:26 * nmagnezi looks at the etherpad again
20:19:26 <johnsom> It has the "fan out" concept...  I guess.
20:20:52 <nmagnezi> peacock is fine by me :)
20:21:03 <xgerman> yeah, I like to have stickers in Boston
20:21:07 <johnsom> Any other votes/comments?
20:21:24 <johnsom> Yeah, with the foundation dropping the ball here I'm not sure we will have anything for boston
20:22:18 <johnsom> Ok, with no other comments I will reply to Heidi with peacock
20:22:32 <xgerman> cool
20:22:46 * johnsom Thinks it had better not be lame
20:23:10 <johnsom> #topic Discuss asynchronous API and response handling
20:24:29 <johnsom> We have had this on the agenda for a week or two, I would like to make a call on this but I also want all of the cores the be here....
20:24:56 <johnsom> Hmm, darn.  Ok, let's bump this again.
20:25:08 <johnsom> #topic Open Discussion
20:25:18 <johnsom> Any other topics today?
20:25:47 <nmagnezi> nothing much on my end. just wanted to share that the triplo support for Octavia is very close to be finalized
20:25:55 <xgerman> Cool!
20:25:57 <johnsom> Oh cool.
20:26:14 <nmagnezi> I basically need to finish to fix ovs_cleanup script which destroys the hm port on reboot
20:26:15 <johnsom> Someone was just asking me about the different ways Octavia can be installed.
20:26:52 <johnsom> Send me a link when that is ready, I should probably put a section on our wiki page for links
20:26:54 <nmagnezi> yeah.. btw for Queens we plan to deploy octavia services via tripleO in containers.. but there's time for that
20:27:02 <nmagnezi> sure will
20:27:33 <xgerman> we already have that in OSA :-)
20:28:00 <nmagnezi> hah.. now you're telling me..
20:28:27 <jniesz> would it be possible to install Octavia components in containers, such as k8 without OpenStack?
20:28:45 <xgerman> no, we need nova, neutron, etc.
20:28:58 <nmagnezi> jniesz, i think xgerman is right..
20:29:11 <xgerman> though in theoru you can write drivers for something else
20:29:12 <jniesz> ok, I ask because we have 2 VMware hosts in each of our stores and there is k8 cluster on top
20:29:21 <jniesz> but not enough resources for openstack
20:29:25 <johnsom> Yeah, right now we are tied into keystone, nova, neutron, and barbican.  It could be done by writing drivers for an alternative, but it's not there today
20:29:47 <jniesz> ok, yea I was thinking there could be a driver implementation for other platforms
20:29:55 <jniesz> that is something we might be interested in helping out with as well
20:30:17 <johnsom> jniesz You can run the amphora without OpenStack access, it just needs the mgmt network
20:30:40 <johnsom> jniesz Is your vmware NSX enabled?
20:30:53 <jniesz> no, unfortuantly in stores everything is L2
20:30:59 <johnsom> Ok
20:31:01 <xgerman> I think vmware should not be that difficult to port to
20:31:24 <xgerman> but probably will tie up 4 people for 4-6 months
20:31:38 <johnsom> True, a driver for vmware would not be too hard.  It just depends on if the control plane services need to be in store or not
20:32:57 <johnsom> Hmm, yeah, I think it is possible, just would need some whiteboard time to chew through the details
20:33:25 <xgerman> yeah, pretty sure it is but I stand to my estimate for effort
20:33:36 <johnsom> Haha
20:33:37 <jniesz> ok thanks, agree we would have to map out the details and loe
20:34:17 <xgerman> once you rewrote the drivers I don;t think you will have much dependency on OpenStack
20:34:52 <johnsom> Yeah, just the keystone part for the API, but you could always go noauth on that
20:35:08 <xgerman> or write another driver ;-)
20:35:19 <nmagnezi> johnsom, speaking of noauth https://review.openstack.org/#/c/459654/
20:35:21 <johnsom> Yep
20:35:24 <nmagnezi> (needs work..)
20:35:25 <nmagnezi> :D
20:36:22 <johnsom> Yeah, I saw that.  Cool
20:36:59 <johnsom> Ok, anything else for today?
20:37:12 <nmagnezi> a question about nova lxd support
20:37:23 <nmagnezi> where does it stand nowadays?
20:37:27 <johnsom> Ah yes
20:37:35 <xgerman> frozen
20:37:42 <johnsom> Well, we don't have anyone working on it at the moment.
20:37:42 <xgerman> we lost that developer
20:38:06 <johnsom> Last status I knew, we were waiting for canonical to backport an lxd bug fix
20:38:13 <nmagnezi> makes that k8 even more important i guess..
20:38:26 <nmagnezi> k8 driver i mean
20:38:37 <johnsom> Power to you!
20:39:01 <nmagnezi> hah.. I'll get there.. i just really need to warp up with the tripleO support
20:39:02 <johnsom> Yeah, we still really need a working container system for the amphora.
20:39:40 <xgerman> nmagnezi why int this systemd thing
20:39:41 <xgerman> ?
20:39:59 <xgerman> nspawn and such?
20:40:04 <johnsom> Given all of the bugs we keep finding I really wonder how/why people are jumping into this stuff so fast.
20:40:11 <nmagnezi> xgerman, i don't follow :<
20:40:21 <xgerman> https://cloudnull.io/2017/04/rocket-past-the-white-whale-and-spawn-a-container/
20:41:07 <xgerman> another technology…
20:41:54 <nmagnezi> I'll read that blog after the meeting
20:42:09 <xgerman> K8 replicates some of our control plane functionality so I am not sure if that would be a perfect fit without some re-design on our part
20:44:12 <johnsom> Yeah, maybe.
20:44:29 <nmagnezi> xgerman, well last time we spoke about it I also said there's a lot research i'll need to do. I actually never actually learned k8 very deeply. So based on future findings I guess we will discuss our possibilities
20:45:11 <xgerman> yeah, true.
20:45:56 <xgerman> I think we might get back to sbalukoff’s original design to run multiple tenants in namespaces on a single host
20:46:17 <nmagnezi> anyhow lxd is not something I plan to look at.. since it is Ubuntu centric
20:46:26 <xgerman> k
20:46:35 <johnsom> Grin
20:47:14 <nmagnezi> johnsom, you already said distros can't agree on a lot of things :)
20:47:19 <johnsom> I have not allegiance to any of the container tech, just the one that works...  grin
20:47:38 <xgerman> moby!!!
20:47:48 <johnsom> nmagnezi yes, even simple things like file paths...  grin
20:48:02 <nmagnezi> so I've heard :)
20:48:11 <xgerman> we can always request a VM and load it up with containers
20:48:51 <johnsom> Yeah
20:49:21 <nmagnezi> i don't know what to think about that honestly..
20:49:21 <johnsom> Ok, if we don't have any more topics I will close out the meeting and we can get back to working on code!
20:49:31 <nmagnezi> johnsom, +1
20:49:34 <xgerman> agreed
20:49:50 <johnsom> Ok, thanks folks!  Chat with you next week.
20:49:54 <johnsom> #endmeeting