16:00:01 <dougwig> #startmeeting neutron lbaas
16:00:02 <openstack> Meeting started Tue Feb 10 16:00:01 2015 UTC and is due to finish in 60 minutes.  The chair is dougwig. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:05 <openstack> The meeting name has been set to 'neutron_lbaas'
16:00:10 <dougwig> #topic roll call
16:00:12 <dougwig> morning all
16:00:13 <johnsom> o/
16:00:13 <ajmiller> o/
16:00:16 <TrevorV> o./
16:00:19 <dougwig> agenda:
16:00:20 <dougwig> https://wiki.openstack.org/wiki/Network/LBaaS#Agenda
16:00:23 <TrevorV> (the dot is my bicep)
16:00:24 <blogan_> hi
16:00:26 <xgerman> o/
16:00:56 <dougwig> no wonder it's so tiny
16:01:17 <blogan_> zing
16:01:36 <blogan_> one day ill spend enough time to remove this scarlet underscore
16:01:42 <[1]evgenyf> Hi
16:01:43 <jorgem> o/
16:01:55 <dougwig> blogan_: run a supported OS.
16:02:10 <kobis> hi
16:02:12 <sballe__> o/
16:02:58 <blogan_> dougwig: bwahaha
16:03:11 <dougwig> #topic Announcements
16:03:21 <jorgem> blogan almost didn't make it in today
16:03:35 <jorgem> he got towed again
16:03:39 <kobis> towed again?
16:03:42 <kobis> nah..
16:03:42 <xgerman> ?
16:03:44 <johnsom> Hahaha, seriously?
16:03:47 <ajmiller> togan?
16:03:50 <ajmiller> towgan?
16:03:51 <johnsom> towgan
16:03:54 <jorgem> nah jk. I with though lol
16:03:56 <dougwig> just one today from me - for anyone not at the mid-cycle, lbaas v2, the non-agent driver, and the CLI merged.  there's a devstack review out, so deploying and testing v2 is simpler than ever.
16:03:58 <TrevorV> towgan towgan towgan
16:03:59 <jorgem> wish*
16:04:06 <jorgem> sorry had to :)
16:04:14 <jorgem> anyways sorry for the distraction
16:04:16 <dougwig> the interface for vendor drivers is likely pretty solid now, so i'd encourage driver submissions.
16:04:16 <jorgem> carry on
16:04:40 <sballe__> dougwig: +1
16:04:46 <dougwig> jorgem: i'm at home surrounded by screaming kids, so i didn't even notice.  :)
16:05:07 <dougwig> #link https://review.openstack.org/#/c/153079/
16:05:12 <dougwig> ^^ the devstack review
16:05:13 <sballe__> dougwig: What drivers do we think we have in the pipeline? a10. Who else?
16:05:14 <xgerman> dougwig - your's scream, too
16:05:16 <xgerman> ?
16:05:30 <dougwig> a10, brocade, vmware, radware, netscaler
16:05:35 <dougwig> and the agent driver
16:05:45 <sballe__> cool thx
16:05:57 <[1]evgenyf> radware driver is close to review submit
16:06:05 <kobis> vmware will be in within a couple of weeks
16:06:05 <sballe__> great
16:06:11 <kobis> with the ci
16:06:25 <dougwig> yes, version relevant ci is mandatory.
16:06:42 <dougwig> any other announcements today?
16:07:11 <ajmiller> There was a comment yesterday about the devstack review https://review.openstack.org/#/c/153079/ I would like to discuss when we have a chance.
16:07:22 <blogan_> i saw taht
16:07:23 <dougwig> ok, let's hit that in open discussion
16:07:28 <ajmiller> OK
16:07:31 <dougwig> or, let's do it now.
16:07:34 <dougwig> #topic devstack changes
16:07:40 <dougwig> ajmiller, you have the floor
16:07:42 <ajmiller> OK
16:08:08 <ajmiller> The comment is that we should split LBaaS out of the main devstack code and into a plugin.
16:08:25 <dougwig> i agree with him, but the current mechanism doesn't work to be fully external.  i'll respond on the review.
16:08:43 <ajmiller> I started to look into the mechanics of that yesterday, and yes, I'm not sure we can fully take advantage of that mechanism.
16:08:51 <blogan_> does that mean the lbaas code will not live in devstack?
16:08:55 <ajmiller> I to have some work in my repo to better modularize my changes.
16:09:15 <dougwig> (commenting done)
16:09:17 <ajmiller> They are working on a mechanism to deliver plugins as part of the product repo,
16:09:42 <ajmiller> Thanks Doug, you summed up the questions well.
16:09:45 <blogan_> so the lbaas devstack plugin woudl live in neutron-lbaas?
16:09:52 <dougwig> it means the devstack code lives in devstack/ in our repo, which i vastly prefer, but i think devstack will need an alias file of service tokens to external plugin locations, or something, because all the CI right now can only handle services, not plugins.
16:10:34 <rm_work> ^^ barbican does this
16:10:35 <blogan_> ah okay, yeah id prefer it living in our repo for obvious reasons
16:10:36 <ajmiller> What I hope we can do is to get the V2 changes merged soon, then do the full plugin work in the (near) future
16:10:55 <blogan_> that makes sense
16:10:57 <dougwig> rm_work: barbican also has an extras.d hook to solve the services token issue.
16:11:05 <rm_work> yeah
16:11:37 <rm_work> I mean Barbican already does the whole "devstack plugin in-tree" thing
16:11:46 <rm_work> but yeah, also the extras.d hook
16:12:26 <rm_work> https://gist.github.com/rm-you/6feacb91182f5c011018#file-setup-sh-L11-L13
16:13:00 <dougwig> i expect ajmiller has enough to go on here, or is there more to discuss?
16:13:22 <dougwig> #topic Summit talks
16:13:23 <ajmiller> dougwig, this is enough to get me moving forward
16:13:27 <dougwig> we have two submitted:
16:13:32 <dougwig> https://www.openstack.org/summit/vancouver-2015/call-for-speakers/TalkDetails/2703
16:13:36 <dougwig> https://www.openstack.org/summit/vancouver-2015/call-for-speakers/TalkDetails/3093
16:13:51 <blogan_> can people see those?
16:13:58 <xgerman> I can't
16:14:08 <blogan_> i didnt think so
16:14:09 <dougwig> if you have items you'd like to present or demo during either, ping myself or brandon and we'll see about integrating.  i'd like to demo v2 with octavia 0.5 at vancouver.
16:14:10 <sballe__> I can't
16:14:13 <dougwig> oh, darn.
16:14:33 <dougwig> i'll get their contents into etherpads for folks to tweak on later today, then.
16:14:40 <sballe__> +1
16:14:48 <dougwig> here are the talk titles:
16:14:50 <sballe__> I like the demo idea.
16:14:57 <dougwig> Load Balancing as a Service, Kilo and Beyond
16:14:57 <dougwig> Neutron mitosis and the L4-7 services roadmaps
16:14:58 <xgerman> demos are great
16:15:11 <dougwig> i only want one of those talks to get selected, not both.  :)
16:15:28 <blogan_> what are you going to do if they both get selected?
16:15:46 <sballe__> We'll figured it out once they both are selected :-)
16:15:51 <sballe__> Just in time worruing
16:15:54 <sballe__> worrying
16:16:12 <dougwig> there is a selection process after voting, so i'll make my preference known.
16:16:12 <blogan_> i like it!
16:16:14 <xgerman> agreed
16:16:27 <dougwig> worst case, we have two talks.
16:16:35 <dougwig> this topic leads to....
16:16:36 <dougwig> #topic Design talks
16:16:54 <dougwig> be thinking about whether we have anything that needs design summit time.
16:17:16 <xgerman> we still didn't resolve the status me thinks
16:17:22 <xgerman> nor sharing
16:17:22 <dougwig> i know of one, which is how to resolve the inter-dependency mess between neutron and the services (and decomposed repos) and how to test them coherently and/or split out a real neutron library.
16:17:23 <blogan_> sharing entities, L7 maybe (if not finalized)
16:17:40 <blogan_> xgerman: pretty sure we did, status tree
16:18:00 <xgerman> ok
16:18:03 <sballe__> I would like for us to get moving on the flavor again
16:18:12 <dougwig> we might want a smaller session  just for brainstorming next features.
16:18:23 <sballe__> dougwig: I like that idea
16:18:30 <dougwig> sballe__: that's out for review: https://review.openstack.org/#/c/139758/
16:18:32 <dougwig> (flavors)
16:19:12 <dougwig> we'll bring this topic up again.  i just wanted to highlight this so that folks start thinking about anything that needs high-bandwidth face time.
16:19:16 <sballe__> dougwig: great! way ahead of me :-) I thought it had been put on the back burner. Will review
16:19:31 <dougwig> sballe__: it's approved for kilo, assuming I can get it through.
16:20:01 <blogan_> c0ffee_flavors.py (the last characters of the migration) makes me chuckle
16:20:03 <dougwig> ok, moving on.
16:20:26 <dougwig> that was a eugene easter egg.
16:20:28 <dougwig> :)
16:20:31 <dougwig> #topic v2 next steps
16:20:37 <kobis> :)
16:20:57 <dougwig> great progress at the mid-cycle.  next steps are the agent driver, tis, l7, and vendor drivers.  anything blocking anybody in any of those areas?
16:21:14 <dougwig> i think the owners are phil, adam/evgeny, evgeny, and misc.
16:21:24 <dougwig> respectively
16:21:35 <xgerman> I am also wondering how we can help the owners...
16:22:14 <dougwig> i'd suggest following up with them directly.  if you don't get a response, ping myself or brandon.
16:22:27 <xgerman> k
16:22:36 <dougwig> but please do.
16:22:54 <dougwig> anyone have anything else to discuss about v2 progress?
16:23:11 <ajmiller> I should have a WIP review for the provider extention to the create API soon.
16:23:21 <TrevorV> Oh, uh, we're still in progress with the tempest tests
16:23:22 <[1]evgenyf> l7 is ready for review, please review
16:23:31 <[1]evgenyf> https://review.openstack.org/#/c/148232/5
16:23:46 <TrevorV> The real troublesome parts have been actually running the tests.  Devstack is a frustrating beast ha ha
16:24:04 <[1]evgenyf> I will also push TLS with some comments handled today or tomorow
16:24:05 <[1]evgenyf> https://review.openstack.org/#/c/145085/17
16:24:15 <dougwig> [1]evgenyf: great.
16:24:26 <jorgem> Still working on the agent manager code #WIP
16:24:32 <dougwig> are those going to need to change with the agent driver, or is that part of the code shared?  (blogan_ )
16:24:37 <[1]evgenyf> radware driver commit tomorrow
16:24:47 <rm_work> [1]evgenyf: so are you on those two commits 100% now? should I leave you to it?
16:24:49 <dougwig> by those, i mean the tis and l7 reviews
16:25:26 <rm_work> [1]evgenyf: I had been working out bugs in your TLS stuff during the midcycle while you were on L7, but seems that you've been committing work on TLS again
16:25:45 <blogan_> im going to try and get the code shared, but the tls and l7 implementatiosn will need to update the sync driver if they get in before the agent does
16:25:46 <[1]evgenyf> rm_work: it's almous done. L7 is waiting for reviews, TLS will be tomorrow. Will see how it's going with comments
16:26:07 <sballe__> rm_work: can you point me to the review for TLS?
16:26:31 <sballe__> oh sorry just saw it
16:26:40 <[1]evgenyf> sballe__: https://review.openstack.org/#/c/145085/
16:26:56 <dougwig> sballe__: this is a handy link:
16:26:57 <dougwig> https://review.openstack.org/#/q/(openstack/neutron-lbaas)+status:open,n,z
16:27:22 <xgerman> I put it on top of the review etherpad
16:27:39 <xgerman> (the old link failed)
16:27:59 <dougwig> #topic Open Discussion
16:28:17 <dougwig> anything anyone wants to discuss?
16:28:53 <ajmiller> Thanks to the Rackspace people for the hospitality last week!
16:29:04 <kobis> +1!
16:29:05 <rm_work> Castellan appears to be possibly moving along again, I hope we can get that wrapped up by the end of NEXT week, and then hopefully in the following weeks before the Kilo-3 freeze I can get it swapped in for the CertManager stuff
16:29:39 <rm_work> though my time on Neutron-lbaas and Octavia has been significantly reduced due to internal scheduling
16:29:41 <dougwig> ajmiller: +1, though -0.01 for no CEC.  :)
16:29:47 <dougwig> they spoiled us last time.
16:29:52 <rm_work> I will still have time to do reviews, but that may be it for a while
16:29:53 <blogan_> you still got all the benefits of the CEC for 3 days!
16:30:07 <blogan_> and you had the added entertainment value of me getting towed
16:30:15 <dougwig> the biggest benefit was seeing you towed.
16:30:48 <johnsom> Actually it was watching you drive by a parking spot and then get towed
16:30:55 <rm_work> ^^ this
16:30:59 <rm_work> I would have felt bad otherwise
16:31:35 <dougwig> ok, any real topics, or are we done today?
16:31:37 <pc_m> I added coverage test enabling for LBaaS repo: https://review.openstack.org/#/c/154208/. Question: Do you want tests included in coverage, by default?
16:31:38 <blogan_> :(
16:31:57 <dougwig> pc_m: what's the difference?  i note that neutron excludes them.
16:32:30 <pc_m> It just gives tells you if all the unit tests are actually being invoked.
16:32:36 <pc_m> Not a big difference.
16:32:44 <dougwig> i'm good either way.
16:32:50 <pc_m> I have it set to include UTs for VPN repo.
16:33:08 <pc_m> Easy to comment out or uncomment so I'm fine either way.
16:33:22 <pc_m> If there's a preference, I can do it as desired.
16:33:39 <blogan_> does that get run as part of tox?
16:33:46 <dougwig> i don't have a preference, i was just curious as to the difference.
16:33:49 <dougwig> tox -e coverage
16:33:53 <pc_m> blogan_: Yeah, you do "tox -e cover"
16:34:17 <blogan_> okay i dont have a preference either
16:34:20 <pc_m> Does coverage testing of production code (and, if not omitted, test code)
16:35:19 <blogan_> looks like its your call pc_m
16:35:30 <pc_m> I commented the omission out for VPN, that way it show all coverage. I can ignore the test files when looking at report, or uncomment the omit line to exclude the test area.
16:35:54 <pc_m> blogan_: OK. I'll push out the updated patch and do the same as VPN repo.
16:36:30 <dougwig> pc_m: thank you!
16:36:31 <blogan_> pc_m: thanks!
16:36:37 <dougwig> any other topics today?
16:36:46 <pc_m> np. Pushing in a few mins.
16:37:00 <dougwig> ok, thanks folks!
16:37:08 <rm_work> o/
16:37:18 <xgerman> o/
16:37:23 <dougwig> #endmeeting