20:00:06 <johnsom> #startmeeting Octavia
20:00:06 <openstack> Meeting started Wed Jan 18 20:00:06 2017 UTC and is due to finish in 60 minutes.  The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:10 <openstack> The meeting name has been set to 'octavia'
20:00:23 <johnsom> Hi folks
20:00:26 <ankur-gupta-f2> howdy
20:00:43 <johnsom> Busy agenda today with announcements....
20:01:00 <johnsom> #topic Announcements
20:01:01 <abeda> hi
20:01:13 <johnsom> Ocata-3 is week of January 23rd
20:02:04 <johnsom> This is feature freeze week.  All non-bug fixes must be in by next week.  I will try to cut that earlier in the week, so if you have something that might be a day late, please work with me directly.
20:02:32 <johnsom> Just a reminder, we are working against the lbaas merge review priority list:
20:02:40 <johnsom> #link https://etherpad.openstack.org/p/octavia-ocata-merge-review-priority
20:03:03 <johnsom> I have created an etherpad for topics at the PTG
20:03:10 <johnsom> #link https://etherpad.openstack.org/p/octavia-ptg-pike
20:04:06 <johnsom> Also of note, it is PTL election time for Pike.  Nominations open tomorrow (depending on your timezone).
20:04:17 <johnsom> #link https://governance.openstack.org/election/
20:04:33 <abeda> #link https://etherpad.openstack.org/p/Active-_Active_Topology_commits
20:04:48 * blogan sneaks in late
20:04:54 <johnsom> Hi abeda.  We have that on the agenda a bit later.
20:05:02 <johnsom> #link https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda
20:05:16 <abeda> ok
20:05:44 <johnsom> FYI, I do intend to run again for the Octavia PTL, but encourage anyone interested to also run...
20:06:46 <johnsom> I have also completed the work to allow octavia to be translated.  This has been an agenda item for a while, but now all of the patches are in.
20:06:54 <johnsom> #link https://translate.openstack.org/project/view/octavia
20:07:26 <johnsom> Note, the translation team prioritizes horizon dashboard highest, so we may not get translations for some time.
20:08:04 <johnsom> However, everything is in place, so if you have folks interested in doing translations for octavia, they can via the i18n/translations teams/infrastructure.
20:08:31 <johnsom> The i18n PTL was very supportive, so that is cool.
20:09:00 <johnsom> Also an FYI, the TC is working on goals for Pike and Queens.
20:09:07 <johnsom> #link https://governance.openstack.org/tc/goals/pike/index.html
20:09:56 <johnsom> Right now it looks like python 3.5 support is approved for Pike.  We are pretty good on that, I think we just need a functional test gate and we are done thanks to the work diltram did.
20:10:17 <johnsom> Note, those py3x gates are failing as nova is not working under py35 at the moment....
20:10:34 <xgerman> o/ sorry for being late
20:10:38 <johnsom> There is also a proposed goal for tempest plugins:
20:10:45 <johnsom> #link https://review.openstack.org/#/c/369749/
20:10:59 <johnsom> and a proposed Queens goal for wsgi:
20:11:08 <johnsom> #link https://review.openstack.org/#/c/419706/
20:11:19 <johnsom> If you would like to comment on those.
20:11:44 <johnsom> These are things we are accountable for as an OpenStack project, so please comment if you have concerns.
20:12:05 <johnsom> Any questions about that long list of announcements?
20:12:15 <johnsom> (longest I think I have ever had...)
20:12:54 <johnsom> #topic Brief progress reports / bugs needing review
20:12:59 <johnsom> Ok, updates
20:13:24 <johnsom> I have been working to track down this subunit bug "length too long".
20:13:41 <johnsom> I am also working on the scenario/api tests for the v2 API.
20:14:07 <johnsom> abeda had mentioned the active/active review page has been updated:
20:14:16 <johnsom> #link https://etherpad.openstack.org/p/Active-_Active_Topology_commits
20:14:48 <johnsom> So, if you have reviewed everything on the merge list and have time these are ready for review.
20:15:15 <johnsom> Anything you wanted to add abeda?
20:15:35 <abeda> we like to have commits 1..6 from the etherpad
20:15:40 <abeda> in this release
20:15:57 <abeda> it doesn't include functionality
20:16:28 <abeda> it is up to Distributor Image creation
20:16:54 <johnsom> abeda I'm not sure that is going to happen.  We have been targeting that for Pike and we are in the last week before the freeze.  It might, but just an heads up that it might not make Ocata.
20:17:19 <abeda> perelman (she is in vacation) told me that this was agreed
20:18:31 <johnsom> Also, nmagnezi said he could not make the meeting today.  He is close on the amphora centos support.  That may be ready for review tomorrow.
20:18:50 <johnsom> #link https://review.openstack.org/331841
20:19:07 <abeda> it is passing and doesn't include any new functionality effect. infrastructure somehow
20:19:44 <abeda> i submitted the fulfillment of the reviewers notes
20:20:14 <abeda> from the latest review. it is important to us. please consider
20:20:16 <abeda> thanks
20:20:27 <johnsom> Yes.  It is still feature work though, so falls under the freeze.  We might get enough reviews on it to include.  Anything that changes code paths though, will be differed to Pike to reduce the risk of breaking Ocata in the last weeks.
20:21:21 <johnsom> Are there any priority patches for neutron-lbaas that we should be aware of?
20:22:01 <abeda> i dont know about any
20:22:23 <johnsom> Ok, any other updates from folks?
20:22:50 <johnsom> #topic Open Discussion
20:23:23 <ankur-gupta-f2> discussion over whether to do full alignment for API v2 or take fastest route to bringing up v2 API
20:23:24 <johnsom> abeda Just a note, when sending e-mail to the openstack-dev list, please include the [octavia] tag in the subject line so we see it.
20:23:56 <johnsom> ankur-gupta-f2 Yes, let's talk about that.
20:23:57 <abeda> sure
20:24:27 <johnsom> So for the neutron-lbaas v2 API work in the Octavia API(v2)...
20:24:43 <ankur-gupta-f2> I for the position that when we bring up the new controllers. We migrate the DB and make all the changes needed to be aligned
20:24:47 <johnsom> There are differences between neutron-lbaas and octavia, including in the data models.
20:24:58 <xgerman> +1
20:25:27 <johnsom> My assumption had been that we were mapping the API to the existing octavia data model and not updating the data model code.
20:25:58 <johnsom> ankur-gupta-f2 is proposing making the data model changes now to reflect the neutron-lbaas way.
20:26:18 <johnsom> An example is neutron-lbaas has "admin_state_up" where octavia has "enabled".
20:26:31 <johnsom> ankur-gupta-f2 is that a fair description?
20:26:40 <ankur-gupta-f2> correct
20:26:54 <blogan> i'd rather have enabled
20:27:30 <xgerman> wouldn’t that hose our backward compatibility?
20:27:32 <blogan> but i can' thave it
20:27:35 <blogan> i know
20:27:38 <johnsom> My preference is to do the mapping in the API and leave the existing  octavia model as-is until some point in the future that we feel the need to change it.  This seems like a lower risk/faster solution
20:27:40 <m-greene> johnsom: can we add [octavia] to the openstack mailers list to allow selection in/out of the mailer?  it’s not listed here http://lists.openstack.org/cgi-bin/mailman/options/openstack-dev
20:28:08 <ankur-gupta-f2> blogan: different discussion. this is whether when we bring up the controllers to do the db migration as well.
20:28:27 <ankur-gupta-f2> because there are other things that need to be renamed
20:28:43 <xgerman> mmh, so if I run Octavia straight but also have neutron-lbaas with the Octavia plugin…
20:28:55 <xgerman> would that still work after the db update?
20:29:13 <blogan> if you have the pass-through
20:29:24 <ankur-gupta-f2> yes full compatibility with v1, the new v2 which will be aligned to lbaas-v2
20:29:43 <xgerman> even if we change the DB?
20:29:49 <johnsom> m-greene In theory yes.  I'm not sure how to add octavia to that selector list.  I will look into that.
20:29:59 <blogan> the current octavia driver only works with v1
20:30:11 <blogan> v1 octavia
20:30:21 <xgerman> yep, and v1 will continue to work after the DB changes?
20:30:30 <johnsom> Right you would have to change the v1 API to do the mapping to the *new* data model.
20:30:36 <xgerman> or do we then spend the work we save on v2 on making v1 work?
20:30:45 <blogan> yeah so thats 2 things that would need to be changed
20:30:45 <johnsom> So, either way you are mapping it somewhere.
20:30:55 <blogan> whereas chaning v2 to work with v1 data model is jsut one thing
20:31:37 <xgerman> yeah, my expectation is that v1 continues to work (at least for the next three cycles)
20:31:49 <blogan> then i definitely agree that teh data model shouldnt' be changed
20:31:57 <johnsom> I just think changing the DB and data model introduces a lot of risk as that means going deep into the code to make the changes.
20:32:12 <johnsom> xgerman +1 v1 has to keep working
20:32:32 <blogan> there's a risk in the duct tape that is required to get the v1 model to work with teh v2 api
20:32:53 <ankur-gupta-f2> okay. I think its decided then. not touching the db
20:32:54 <xgerman> yeah, we need to proper deprecate v1
20:32:54 <blogan> but it is less risk
20:33:17 * blogan properly deprecates xgerman
20:33:29 <johnsom> blogan Agreed, but I would rather have duct tape on the V2 api then break both v2 and v1
20:34:22 <johnsom> Ok.  One other topic that was raised, https://bugs.launchpad.net/devstack/+bug/1655656
20:34:22 <openstack> Launchpad bug 1655656 in octavia "test_calculate_delta fails with message "TypeError: unhashable type: 'list'"" [Undecided,Incomplete]
20:34:40 <blogan> speaking of duct tape :)
20:34:49 <blogan> i'm sure that originates with me
20:35:11 <johnsom> I have been running --until-failure tests on master py35 all morning.  I'm 99.5% sure this issue is caused by a change in that patchset chain.
20:36:07 <johnsom> I will continue to run the test on master, but after the day if it doesn't show up again I am going to conclude that it is a problem in that patch chain.
20:36:38 <blogan> interesting
20:36:48 <johnsom> Ok, any other topics today?
20:38:13 <johnsom> abeda you just missed the discussion of the bug.  Please check the meeting log...
20:38:31 <abeda> sorry for the disconnect
20:38:43 <abeda> regards the bug
20:38:43 <johnsom> Ok, well if that is all we can end a bit early.
20:39:12 <johnsom> I just said that I have been running the test on master all morning and have not seen the failure.
20:39:14 <abeda> yes. i followed your messages
20:39:19 <johnsom> I think it is a bug in the patch chain
20:39:37 <johnsom> But will keep it running the rest of the day just to make sure.
20:40:18 <abeda> thanks
20:40:36 <johnsom> Ok, thanks folks.  Chat with you all next week or in the channel.
20:40:45 <johnsom> #endmeeting