| *** tobias-urdin7 is now known as tobias-urdin | 14:07 | |
| *** tobias-urdin0 is now known as tobias-urdin | 14:20 | |
| mnaser | Wow, this sucks very much for all users and feels very sketchy — https://www.stackhpc.com/magnum-clusterapi.html | 17:59 |
|---|---|---|
| mnaser | We’ve literally lead the whole damn effort on the cluster api only to find out some other implementation which relies on our code (that we changed to Apache 2 based on stackhpc request) gets to be “the default one” in magnum that “they led the effort on” | 18:00 |
| mnaser | including code that’s straight up copy pasted from our implementation into magnum | 18:00 |
| mnaser | And a significantly poorer implementation, working with another org and never pulling us into the convo? | 18:01 |
| mnaser | It feels like the magnum team is an absolute mess and they’re pushing for a worst user experience because of alterior reasons and then throwing jabs to do that too. It’s very frustrating. | 18:01 |
| mnaser | I’ve never got a single ping from them about the whole driver being merged as a main | 18:01 |
| fungi | mnaser: which commits was it? not entirely clear when skimming https://opendev.org/openstack/magnum/commits/branch/master | 18:35 |
| fungi | i do see some unreviewed proposed changes for magnum about adding a driver | 18:36 |
| fungi | though those are from mid-2022 and were last touched in october | 18:37 |
| * TheJulia reads and tries to digest | 18:44 | |
| mnaser | fungi, TheJulia: i've collected my thoughts here -- https://etherpad.opendev.org/p/magnum-state | 18:48 |
| TheJulia | I guess I have concerns with framing which has me wanting to ask questions, but I don't have backing context. I grok there is an aspect here of marketing and storytelling for StackHPC customers as well, but... I guess that makes the page feel more concerning overall to my own perception | 18:49 |
| JayF | My immediate thought is that this /really/ makes me wish we record PTG sessions. I understand why we don't, but much of the context around what happened last PTG is going to be in the form of eyewitness accounts that are months-old at this point. | 18:50 |
| fungi | got it, so the concern is that stackhpc says the magnum project is going to merge it next cycle, and is taking credit for that work | 18:51 |
| mnaser | _and_ magnum as a project gets a worst experience for upstream users | 18:52 |
| mnaser | which honestly is the thing that bums me out, because i hate that we wouldn't ship something that could be way better | 18:52 |
| fungi | seems like some clarification from magnum leadership is warranted, since it doesn't appear stackhpc can authoritatively speak on behalf of the magnum team anyway | 18:53 |
| mnaser | like users can get so much more if we used the driver we've been working hard on building, but instead .. this helm driver keeps getting pushed in | 18:54 |
| JayF | I'm going to note that clause 5 does require attribution for commits taken from an external source: https://review.opendev.org/static/cla.html -- at least in my non-lawyerly reading. | 18:54 |
| TheJulia | Yeah, the big thing which feels almost misleading is the link to the D cycle name page instead of any writeup on consensus or outcome | 18:54 |
| TheJulia | Maybe I just have the wrong expectation there | 18:55 |
| fungi | yeah, the opendev sysadmins were just talking this week about how we'd be interested in trying to use kubernetes again if any of the openstack public clouds who donate resources to us have a kubernetes-as-a-service offering that works well, but a few years back when we tried magnum was just not there yet (particularly when it came to being able to use modern linux distro versions and being | 18:55 |
| fungi | able to upgrade clusters) | 18:55 |
| fungi | would definitely love to see improvements in the user experience for that | 18:56 |
| fungi | for the record, a read-only link to the magnum ptg notes pad is https://etherpad.opendev.org/p/r.5b94a5a9dbf4e3bf8 | 18:57 |
| fungi | it does have a lengthy section on the clusterapi discussion | 18:58 |
| TheJulia | fungi: You do not have permission to access this pad | 18:58 |
| TheJulia | well, that is what etherpad tells me | 18:58 |
| fungi | oh, hah. lemme try that again | 18:58 |
| fungi | https://etherpad.opendev.org/p/r.5b94a5a9dbf4e3bf8202b5e57cb93770 | 18:58 |
| fungi | better? | 18:58 |
| fungi | seems my clipboard truncated the url the first time | 18:58 |
| TheJulia | much better, thanks | 19:02 |
| mnaser | fungi: we’re running our capi driver. It’s a world of improvement. I actually would recommend it to people now. | 19:06 |
| mnaser | I couldn’t with my good will recommend magnum for production before | 19:07 |
| fungi | mnaser: cool, sounds like it's worth revisiting in that case | 19:07 |
| fungi | thanks! | 19:07 |
| mnaser | we’ve got rolling restarts now, auto scaling, auto healing | 19:08 |
| mnaser | another tough point is deployment tools will now have to 'decide' on that should/can they support | 19:13 |
| mnaser | i think rather than this approach, the one i clearly pushed for was .. let people install whatever driver they want and leave it up to deployres/operators to choose what they want | 19:16 |
| mnaser | as opposed to force this through | 19:16 |
| fungi | maybe we'll see a heated race for magnum ptl this cycle. nominations are open for another 12 days 4 hours 16 minutes and 30 seconds! | 19:18 |
| fungi | er, that was supposed to be 26 minutes, but who's counting | 19:18 |
| mnaser | fungi: funnily enough our policy doesn't allow me to run :) | 19:22 |
| JayF | Howso? | 19:22 |
| fungi | lack of merged commits | 19:22 |
| mnaser | i work with my team to push things and i am mostly doing reviewing | 19:22 |
| fungi | presumably | 19:22 |
| JayF | Ah, if you had significant review contribution; you could've applied to be an extra-atc. | 19:22 |
| mnaser | so while i work with them to push things through and review them upstream (as core), i have no right to run | 19:22 |
| fungi | though you can still be an extra-atc for | 19:22 |
| fungi | yeah | 19:22 |
| JayF | I'll note such evaluations are done at the TC, not the project level. | 19:23 |
| JayF | (Although I'm fairly sure you're past the deadline for adding yourself at this point) | 19:23 |
| fungi | https://releases.openstack.org/caracal/schedule.html#c-extra-acs ends today | 19:25 |
| gmann | reading all these msg, I agree to have a leadership to think from user perspective and how better it be consumed by most of them than just a company or their marketing | 19:25 |
| gmann | I am also not sure how correct is to committing anything coming in D cycle. even thing were agreed in past those can be changed later or while doing implemention | 19:26 |
| gmann | I do not think/find where we define the extra-ac deadline | 19:27 |
| JayF | Yeah; there's clearly some missing pieces of context here, and we need to be careful about making conclusions until all that context is gathered. | 19:27 |
| gmann | fungi: not sure how it is coming to release schedule though | 19:27 |
| gmann | mnaser: even merge commit deadline is "2024-02-28T00:00" to be consider as APC and valid candidacy for PTL | 19:27 |
| fungi | gmann: in the past, it was tracked in the release schedule since the tc didn't publish separate schedules for such things | 19:28 |
| gmann | ohk | 19:28 |
| gmann | I think we should match those with email-deadline in election schedule | 19:28 |
| gmann | this one https://github.com/openstack/election/blob/master/configuration.yaml#L6-L10 | 19:28 |
| fungi | but that's something worth thinking about for the tc. i think the release team would be just as happy to stop tracking and mentioning election details | 19:28 |
| gmann | ++ | 19:29 |
| gmann | I will propose something so that it extra-ac and ac deadline matches. | 19:29 |
| fungi | came up as recently as last week's release meeting, that having election things in the release schedule feels very out of place | 19:29 |
| gmann | yeah, purpose was to market those election schedule via release schedule but not sure how much that is helping | 19:30 |
| opendevreview | Goutham Pacha Ravi proposed openstack/election master: Add Goutham Pacha Ravi candidacy for TC https://review.opendev.org/c/openstack/election/+/909239 | 19:40 |
| opendevreview | Goutham Pacha Ravi proposed openstack/election master: Add Goutham Pacha Ravi candidacy for TC https://review.opendev.org/c/openstack/election/+/909239 | 19:45 |
| gmann | tc-members: help me to merge the chef repo content removal changes, I have updated the acl to tc-members can review/merge them https://review.opendev.org/q/topic:%22retire-openstack-chef%22+status:open | 22:47 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!