09:01:27 <jakeyip> #startmeeting magnum 09:01:27 <opendevmeet> Meeting started Wed Jul 17 09:01:27 2024 UTC and is due to finish in 60 minutes. The chair is jakeyip. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:01:27 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:01:27 <opendevmeet> The meeting name has been set to 'magnum' 09:01:36 <jakeyip> #link https://etherpad.opendev.org/p/magnum-weekly-meeting 09:01:40 <jakeyip> Please put your topics into to Agenda 09:01:44 <jakeyip> #topic Roll Call 09:01:46 <jakeyip> o/ 09:01:48 <dalees> o/ 09:01:51 <travisholton> o/ 09:02:42 <jakeyip> dalees: travisholton: thanks for coming! 09:02:56 <jakeyip> #topic Review 09:03:26 <jakeyip> https://review.opendev.org/c/openstack/magnum/+/906086 - control plane resize 09:03:34 <jakeyip> I believe this is yours dalees :) 09:04:31 <dalees> Yeah, pleased to get some feedback from mnaser on it. Awaiting his reply as to exactly what he means but otherwise it should be in reasonable shape for review 09:05:40 <dalees> it works well with Heat (and does not permit cp resize), and helm driver (another patch to this is needed, once the above lands) 09:08:09 <jakeyip> yeah I saw his notes, possibly he supports size >7? 09:08:54 <jakeyip> I reviewed this previously, will take a look at the new PS 09:09:07 <dalees> it's probably more the function signature - a driver can override that function and return any values it likes. 09:09:35 <dalees> but yeah, i'm not sure. 09:10:12 <jakeyip> ok 09:11:52 <dalees> I'd appreciate the review to move it forward (without negating mnaser's feedback of course, until he has a chance to reply) 09:13:11 <jakeyip> will do 09:13:28 <jakeyip> anything else? 09:14:11 <dalees> that's all today 09:15:49 <jakeyip> ok my turn :P 09:16:20 <jakeyip> https://review.opendev.org/c/openstack/magnum/+/920725 - need help with CI for v1.28.9 09:16:26 <jakeyip> oh hmm the test is passing now, wth 09:17:56 <jakeyip> oh updated wrong thing 09:18:25 <jakeyip> anyway there are two CI things I'm stuck with 09:18:36 <jakeyip> one of them is getting heat driver passing for v1.28.9 09:18:46 <jakeyip> another one is getting capi working 09:18:55 <jakeyip> anyone able to help with CAPI ? 09:20:10 <dalees> was mnasiadka starting to work on CAPI at one point? it's quite a job to start up that CI with Zuul 09:20:26 <jakeyip> I was hoping to catch up with him but he's been busy 09:20:31 <mnasiadka> We have that task in our backlog, but don't count for it soon ;-) 09:20:39 <jakeyip> hi mnasiadka :P 09:20:42 <mnasiadka> hello 09:21:00 <jakeyip> I am concentrating with getting heat v1.28.9 working 09:21:08 <mnasiadka> I might have more time for Magnum in August 09:21:24 <jakeyip> it's a Caracal promise 09:21:42 <jakeyip> it's a D promise 09:22:19 <jakeyip> it 09:22:30 <jakeyip> it's working in our cloud but not in CI so a bit annoying 09:22:49 <jakeyip> and the motivation to get heat working isn't great :D 09:24:42 <jakeyip> just want to flag early to the team this might be missed, if I can't win CI 09:25:59 <dalees> I've mostly cleaned Heat off my hands, here. 09:26:17 <dalees> though the CI issue isn't really Heat specific, it's strange. 09:26:51 <jakeyip> yeah it looks like issues talking to the API from inside VMs 09:27:23 <jakeyip> we don't really test anything like create PVC in our test so that may never have worked 09:27:39 <jakeyip> similarly we may run into it testing CAPI in CI too 09:29:48 <jakeyip> let's go on 09:30:22 <jakeyip> mnasiadka: have you seen this bug? https://bugs.launchpad.net/magnum/+bug/2067345 09:30:46 <jakeyip> seems like related to the changes for SQLA 09:31:22 <mnasiadka> yeah, I thought it was too easy ;-) 09:31:36 <mnasiadka> I can have a look in August if this doesn't get fixed by somebody else 09:32:55 <opendevreview> Jake Yip proposed openstack/magnum master: DNM: Test CI for v1.28.9 https://review.opendev.org/c/openstack/magnum/+/920725 09:33:05 <jakeyip> mnasiadka: cool thanks 09:33:26 * dalees has to depart, apologies. 09:33:51 <jakeyip> I'll upgrade to Caracal for our cloud too and see if that turns up anything 09:33:54 <jakeyip> bye dalees 09:34:02 <jakeyip> I don't have anything else 09:34:30 <jakeyip> anyone has anythign else? 09:34:42 <travisholton> not me 09:34:44 <jakeyip> #topic Open Discussion 09:34:45 <mkjpryor> I only just managed to join due to another meeting, unfortunately 09:34:50 <mkjpryor> Did I miss anything? 09:35:43 <jakeyip> mkjpryor: I'm trying to persuade mnasiadka to do CAPI CI :P 09:36:12 <mkjpryor> CI for the CAPI Helm driver is on our list of things to get to 09:36:31 <jakeyip> mkjpryor: got a question - do you have patches to support AZ for servers and also maybe volumes 09:37:00 <jakeyip> I realised instances don't have AZ set and also storageclass have wrong AZ set 09:37:04 <mkjpryor> What exactly do you mean by volumes? Boot-from-volume? 09:37:30 <mkjpryor> I see what you mean I think 09:38:29 <mkjpryor> All those things are settable via the Helm charts. If you need them in the Magnum CAPI Helm driver, it should just be a case of a small patch that looks at the appropriate label or property on the Magnum cluster and updates the Helm values appropriately. 09:38:46 <jakeyip> cool 09:39:29 <jakeyip> TBH I haven't looked into this, it was just observed in our testing 09:39:43 <jakeyip> will get to it in the next couple of weeks, sounds like it's easy 09:39:46 <jakeyip> :) 09:39:50 <mkjpryor> There are also some settings in the Helm charts to prevent CAPO from setting AZs at all 09:40:01 <mkjpryor> Allowing placement to do its job 09:44:10 <jakeyip> hm, setting az is still valid for placement filter 09:44:20 <mkjpryor> Yes, but also not setting AZ is fine too 09:44:32 <mkjpryor> I'm just saying both options are available 09:44:34 <jakeyip> yeah, understand 09:45:00 <jakeyip> cool I'll take a look 09:46:08 <jakeyip> that's all from me 09:46:38 <jakeyip> anything else? 09:48:38 <mkjpryor> Just to say that we have started looking at filling out the paperwork to donate Azimuth and all its components to the CNCF sandbox 09:49:01 <mkjpryor> This will include the addon provider, the janitor and the CAPI Helm charts from the stackhpc namespace 09:49:11 <jakeyip> nice 09:49:17 <travisholton> cool 09:49:52 <mkjpryor> I know the CAPI Helm charts have already been forked into an opendev repo, but I was sort of hoping that this would give us a suitable vendor-agnostic home for the CAPI Helm charts? 09:50:06 <mkjpryor> One where we don't have to rewrite all the automation that we have 09:50:08 <mkjpryor> :D 09:50:42 <mkjpryor> How would people feel about Magnum consuming the CAPI Helm charts from a CNCF project as the default? It feels OK to me. 09:51:02 <mkjpryor> But I am biased 09:51:02 <mkjpryor> :D 09:51:53 <jakeyip> yeah this will prob help your case ;) 09:52:23 <mkjpryor> The repos will be being moved into a new GitHub org this week 09:52:27 <jakeyip> have you talked to the folks who maintain OCCM ? 09:52:28 <mkjpryor> azimuth-cloud 09:53:17 <mkjpryor> Not specifically about CNCF membership. Our membership would be slightly different as OCCM belongs to the Kubernetes project whereas we want a new project. 09:53:50 <jakeyip> I see 09:55:29 <travisholton> I think it sounds good. 09:55:53 <mkjpryor> Just a heads up anyway 09:56:07 <mkjpryor> We are hoping to do the paperwork over the next few weeks 09:56:17 <mkjpryor> Then we'll see whether they accept us 09:57:00 <travisholton> so after next week we should clone/fork from the new repo 09:58:24 <jakeyip> do keep us updated :) 09:58:36 <jakeyip> I'll end the meeting now, since we are almost at time 09:58:38 <jakeyip> #endmeeting