08:00:36 <dalees> #startmeeting magnum
08:00:36 <opendevmeet> Meeting started Tue Jul 22 08:00:36 2025 UTC and is due to finish in 60 minutes.  The chair is dalees. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:00:36 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:00:36 <opendevmeet> The meeting name has been set to 'magnum'
08:00:46 <dalees> #topic Roll Call
08:00:50 <dalees> o/
08:00:56 <mtnorthcott> o/
08:01:21 <sd109> o/
08:01:25 <dalees> Meeting agenda is at https://etherpad.opendev.org/p/magnum-weekly-meeting - please add to it if you have things we can discuss
08:01:28 <dalees> jakeyip: ping
08:01:56 <dalees> mnasiadka is on vacation, as I understand.
08:02:36 <mnasiadka> Yup, until 5th Aug
08:03:48 <dalees> ok, enjoy!
08:04:00 <dalees> we'll wait a few minutes for jakeyip, then get started.
08:04:28 <dalees> hi Jasleen_kaur
08:04:49 <Jasleen_kaur> Hi Dalees
08:06:01 <dalees> #topic Proposal: Adding endpoint to renew/recreate Application Credentials
08:06:05 <dalees> #link https://review.opendev.org/c/openstack/magnum-specs/+/955448
08:06:25 <dalees> so mtnorthcott proposed this yesterday, thanks for the review already sd109
08:07:26 <mtnorthcott> Yep, much appreciated sd109. I've made some amendments to the proposal that touch on the workload restarts
08:07:45 <dalees> mostly just wanted to socialise these 2 new specs in this meeting so folk could discuss or review at a later time.
08:08:59 <sd109> Makes sense, I haven't had a chance to look at your changes yet this morning but idea app cred rotation idea looks good in principle
08:10:21 <mtnorthcott> No worries, I'll keep an eye out for any further comments as they come through. All are welcome
08:12:04 <dalees> ok, let's move on and continue discussion in Gerrit.
08:12:51 <dalees> #topic API validation of input using Driver
08:12:58 <dalees> #link https://review.opendev.org/c/openstack/magnum-specs/+/955549
08:13:18 <dalees> again, just posting the link really. have a read and please give feedback.
08:13:52 <dalees> it should allow more specific validation as the driver knows what it supports or doesn't.
08:14:21 <sd109> Sounds like a nice addition, I'll try to have a first pass at reviewing that one later today
08:15:39 <dalees> #topic CAPO errors (e.g. hitting quota limits) visible to users in health_status_reason
08:16:06 <dalees> I remember this being discussed in an issue or on Slack recently.
08:16:22 <dalees> Was this sd109 or jakeyip's topic?
08:16:56 <sd109> I added this one, not so much a proposal but more a request for ideas. It's a frequent complaint we here from Magnum users that clusters get stuck creating or deleting for various reasons and they don't know why without access to the management cluster
08:17:39 <dalees> yeah, we've seen this a bit too. Easy to find for operators of the management cluster, but not Magnum users.
08:18:16 <dalees> the problem is that CAPO treats this as a temporary failure and doesn't set any state in the CAPO objects.
08:18:38 <sd109> Yeah, the kinds of simple errors I have in mind are things like quota errors which the user could probably resolve themselves by deleting other resources
08:18:46 <dalees> so it only logs the errors, and doesn't set the `Failure` fields.
08:19:22 <sd109> Yeah, I found this issue in CAPO yesterday which might help if/when it gets implemented: https://github.com/kubernetes-sigs/cluster-api-provider-openstack/issues/2379
08:19:37 <dalees> Yep - our users can also request automated quota changes, which they might know know they need to do without the feedback from Magnum.
08:20:57 <sd109> So right now the only place that those errors are surfaced is in the CAPO logs as far as I can tell, and I'm assuming there's not much appetite for trying to implement some fragile log message parsing thing to extract those errors...
08:21:39 <sd109> So maybe the answer is we just have to wait for (or help) CAPO to implement the changes so that non-fatal errors get recorded in the status fields of the CAPO resources?
08:21:57 <dalees> yeah, really not keen on the log parsing ;)
08:22:40 <dalees> that issue is very relevant though, I hadn't seen it would raise up transient errors better. That's promising and the better way forward for CAPO.
08:23:51 <dalees> yeah, CAPO might need help to implement with some folk moving on recently.
08:25:14 <sd109> Okay thanks, glad we agreed that log parsing is a bad idea :D happy to move on
08:26:01 <dalees> ok
08:26:22 <dalees> #topic Review: Token-based kubeconfig auth
08:26:24 <dalees> #link https://review.opendev.org/c/openstack/magnum-capi-helm/+/953956
08:27:09 <dalees> hm, that one is nearly complete. Just needs another set of eyes (who aren't here)
08:28:10 <dalees> this patchset did make me check our cert auth - it is indeed 1 year. /me makes note.
08:28:53 <sd109> Yeah that bit us on a couple of deployments which is why we proposed the patch...
08:29:07 <sd109> I added these next two reviews to the adenda just to try to push them over the line but you've already reviewed and other cores aren't available so probably not much to say on them here I think
08:29:48 <dalees> yeah, I was just checking those.
08:30:09 <dalees> I will leave a review on the last one, soon. https://review.opendev.org/c/openstack/magnum-capi-helm/+/951966
08:30:23 <dalees> but generally it's looking good now.
08:30:36 <sd109> Great, thank you!
08:30:45 <dalees> #topic Open Discussion
08:30:54 <dalees> any other topics to discuss?
08:31:22 <sd109> Nothing from me
08:32:44 <mtnorthcott> Nothing here either
08:34:41 <dalees> btw sd109 i uploaded several small patchsets for both magnum and magnum-capi-helm this week, if you had time for some more reviews ;)
08:34:48 <jakeyip> hi all, sorry I'm late
08:35:58 <dalees> otherwise, i'm working on a patchset for mutable flavor_id - allowing users to update a nodegroup with a new flavor (and CAPO will replace all nodes). The validation spec is part of this.
08:36:34 <dalees> hi jakeyip, i'll hang out a bit before closing meeting, but happy to revisit any topics above if you have more to add to them
08:38:00 <sd109> All sounds good, thanks dalees - I'll try to pick up some more reviews later this week
08:38:08 <jakeyip> thanks, nothing from me. I do have lots of reviews to do :P will get to it
08:39:09 <dalees> many thanks to both of you
08:39:36 <dalees> I'll close the meeting then. Thanks for joining all, and thanks for the recently contributions Jasleen_kaur
08:39:42 <dalees> #endmeeting