20:00:04 <johnsom> #startmeeting Octavia 20:00:05 <openstack> Meeting started Wed Oct 17 20:00:04 2018 UTC and is due to finish in 60 minutes. The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:08 <openstack> The meeting name has been set to 'octavia' 20:00:12 <xgerman_> o/ 20:00:13 <johnsom> Hi folks! 20:00:41 <johnsom> #topic Announcements 20:00:50 <johnsom> Just a few items today 20:01:06 <cgoncalves> hi 20:01:10 <johnsom> There is a discussion starting about setting up tempest plugins for the horizon dashboard plugins 20:01:16 <johnsom> #link http://lists.openstack.org/pipermail/openstack-dev/2018-October/135805.html 20:01:51 <johnsom> I commented in favor as I would rather have another repo than try to have tempest code for all of the dashboards in one place. 20:02:20 <johnsom> Also, the unified RBAC policy naming that I have been contributing to has merged: 20:02:21 <nmagnezi> o/ 20:02:25 <johnsom> #link https://docs.openstack.org/oslo.policy/latest/user/usage.html#naming-policies 20:03:16 <johnsom> We will probably want to switch over and deprecate our current names that don't match. 20:03:52 <johnsom> There is also a discussion of moving taskflow out of OpenStack. 20:03:55 <johnsom> #link http://lists.openstack.org/pipermail/openstack-dev/2018-October/135648.html 20:04:17 <johnsom> I'm not a huge fan of the idea, but it wouldn't be the end of the world either. 20:04:39 <johnsom> That is all I have for announcements today. Any others? 20:05:06 <nmagnezi> we will still be allowed to use it in OpenStack, right? 20:05:11 <nmagnezi> (I didn't read yet) 20:05:14 <johnsom> Yes 20:05:22 <johnsom> There are a number of services using it 20:05:47 <xgerman_> dims from the OpenStack TC got elected to the k8s steering committee 20:05:49 <johnsom> My concern is more about maintaining code quality and stability 20:05:54 <xgerman_> so Congrats! 20:06:18 <johnsom> Nice 20:06:37 <johnsom> #topic Brief progress reports / bugs needing review 20:07:23 <johnsom> So, over the last week I have been busy. 20:07:49 <xgerman_> respun #link https://review.openstack.org/#/c/585864/ and https://review.openstack.org/#/c/604479/ still has doc trouble 20:08:36 <johnsom> I have posted patches to enable API audit logging. This is a keystonemiddleware feature. Basically it gives a detailed audit log of who did what and when. There is bug in the keystonemiddleware that I have a patch up for that fixes the target service selection. 20:08:54 * cgoncalves is on a mtg call, ping if needed 20:08:59 <johnsom> I have also added it to our API noop tempest gate to have the code exercised. 20:09:22 <xgerman_> cgoncalves: I do sometimes three meetings at once — WIMP 20:09:31 <johnsom> I have also posted the octavia-lib initial patch and done a pypi release to setup the infra parts. 20:09:52 <johnsom> Reviews would be nice to keep this moving: 20:09:54 <johnsom> #link https://review.openstack.org/610133 20:09:56 <cgoncalves> johnsom, thanks for getting octavia-lib started! I had a quick look, LGTM 20:10:20 <johnsom> cgoncalves one +2+A away.... grin 20:10:30 <cgoncalves> done 20:10:38 <xgerman_> sweet 20:11:09 <johnsom> Then I also finished up the bits on the active/standby VRRP scenario test. I was missing an API test that is now there. 20:11:16 <nmagnezi> and I was starting to read it O_o 20:11:57 <johnsom> Finally, I have been working on adding IPv6 gates to make sure our IPv6 code is functioning. I have found a bug in Octavia that I'm working on now. 20:12:04 <johnsom> So this patch: 20:12:10 <johnsom> #link https://review.openstack.org/611405 20:12:28 <johnsom> Is a IPv4 VIP with a IPv4 member and a IPv6 member. 20:12:50 <johnsom> I plan to add an IPv6 VIP test as well. 20:13:02 <johnsom> These actually pass traffic, so complement the IPv6 API tests. 20:13:55 <johnsom> Looking forward it's flavors, maybe a UDP traffic test, and maybe a TLS with barbican test. 20:14:14 <johnsom> I still need to get back to the multi-node gate issues too, but needed a break 20:14:59 <johnsom> Oh, and I plan to start moving the provider driver code over to the octavia-lib.... 20:15:09 <johnsom> ha, so, not bored 20:15:20 <johnsom> Any other updates today? 20:16:11 <johnsom> #topic Open Discussion 20:16:22 <nmagnezi> I have a question 20:16:37 <nmagnezi> As a followup to a question a got from kuryer folks 20:16:51 <johnsom> I would ask that if you can, please spend some time on reviews. We are getting a bit of a backlog of un-reviewed patches across the repos. 20:17:07 <johnsom> nmagnezi The floor is yours 20:17:20 <nmagnezi> What's the best practice in fetching our API minor version 20:17:27 <nmagnezi> We do, do that our tests 20:17:32 <nmagnezi> With johnsom's work 20:17:43 <nmagnezi> But they have the need to do that in their code as wekk 20:17:45 <nmagnezi> well 20:18:17 <nmagnezi> As for reviews. You're very much right about this 20:18:37 <nmagnezi> we have a lot of internal fires to put out lately 20:18:39 <johnsom> Well, we publish the versions at the root of the endpoint: 20:18:48 <nmagnezi> I hope things will improve soon 20:19:08 <johnsom> https://www.irccloud.com/pastebin/N1YrsV5z/ 20:19:13 <cgoncalves> johnsom, I know. my excuse is: tripleo 20:19:28 <nmagnezi> :D 20:19:30 <johnsom> So you get that version document back that lists the available version(s) 20:19:49 <johnsom> This is what we query in our tempest tests to figure out what API version the test is running against. 20:20:21 <nmagnezi> Sorry if that a silly question, but does it make sense to have it also as an API call? 20:20:23 <johnsom> (Note, 2.3 is the version bump for one of my posted patches.... 20:20:48 <johnsom> Well, technically that is an API call 20:21:01 <nmagnezi> I did not see a precedent for this in other projects 20:21:09 <johnsom> #link https://developer.openstack.org/api-ref/load-balancer/#api-discovery 20:21:29 <nmagnezi> johnsom, yeah I mean actually adding it to our python client 20:21:39 <johnsom> It is exactly the same as neutron uses 20:22:00 <openstackgerrit> Merged openstack/octavia-lib master: Initial cookie-cutter commit for octavia-lib https://review.openstack.org/610133 20:22:31 <johnsom> nmagnezi Ok, so like a CLI command to query it? 20:23:15 <cgoncalves> IIRC nova cli has that 20:23:20 <nmagnezi> johnsom, that too, yes. I was thinking about this to allow them to import the client and basically use it 20:23:30 <nmagnezi> cgoncalves, IIRC it is in nova-manage, no? 20:23:35 <nmagnezi> Is or was 20:23:35 <cgoncalves> https://docs.openstack.org/python-novaclient/latest/cli/nova.html#nova-version-list 20:23:51 <xgerman_> well, we are open source - so just submit a patch 20:24:24 <johnsom> Yeah, but does the modern openstack client have that anywhere? 20:25:19 <cgoncalves> johnsom, again IIRC it does. I actually used openstack client 20:25:26 <johnsom> I think to some degree openstacksdk is supposed to "abstract" that for you. Though I don't know if that is implemented for Octavia yet or not. 20:26:02 <johnsom> Huh, what do you know, "openstack versions show" 20:26:12 <johnsom> https://www.irccloud.com/pastebin/Hwg97syC/ 20:26:18 <johnsom> So, boom, done! 20:26:22 <johnsom> lol 20:26:34 <nmagnezi> wow 20:26:35 <johnsom> I guess we do support it 20:26:45 <nmagnezi> That's neat. 20:26:53 <cgoncalves> https://docs.openstack.org/python-openstackclient/latest/cli/command-objects/versions.html#versions-show 20:27:03 <cgoncalves> oh well you were faster :) 20:27:05 <nmagnezi> yboaron_, ^^ 20:27:24 <nmagnezi> yboaron_, when you guys need to fetch API versions 20:27:40 <johnsom> nmagnezi Does that provide what you were looking for? 20:28:00 <nmagnezi> johnsom, I think so, I will get their feedback but it does look like it 20:28:38 * nmagnezi thanks johnsom 20:29:46 <johnsom> Ok, cool. We followed the non-microversion documents for how to do version discovery. So it should be standard with other non-microversion services 20:30:47 <johnsom> Other topics for today? 20:31:02 <johnsom> Any light at the end of that internal work tunnel? 20:31:36 <nmagnezi> We want to say yes.. 20:31:38 <nmagnezi> :) 20:31:47 <cgoncalves> no 20:32:09 <johnsom> Maybe you should *be* saying no... To the other things.... grin 20:32:10 <nmagnezi> Right.. but we want.. :D 20:33:07 <johnsom> Just FYI, my review wishlist would be: 20:33:09 <johnsom> #link https://review.openstack.org/#/c/589292/ 20:33:26 <johnsom> the IPv6 fix, simply because it touches a lot of code 20:33:58 <cgoncalves> the things we've been doing will hopefully reflect back upstream in terms of octavia adoption, which is good for the project 20:34:36 <johnsom> Hmmm, intriguing. Sounds like eye candy 20:34:41 <johnsom> grin 20:34:51 <johnsom> probably not, probably boring tripleo stuffs 20:35:23 <johnsom> Ok, well if there are not any more topics I will close out the meeting this week. 20:36:03 <johnsom> Alright, thanks folks! 20:36:05 <johnsom> #endmeeting