| *** diablo_rojo is now known as Guest7 | 04:34 | |
| opendevreview | sean mooney proposed openstack/governance master: remove outdated refences to 20.04 and python 3.8 https://review.opendev.org/c/openstack/governance/+/895160 | 10:59 |
|---|---|---|
| *** Guest7 is now known as diablo_rojo | 11:57 | |
| *** tosky_ is now known as tosky | 14:06 | |
| *** diablo_rojo_phone is now known as Guest96 | 15:00 | |
| clarkb | fungi: frickler: sqlalchemy 1.something is the transition release for sqla and made the transition to sqla 2.0 much easier. It is unfortunate that we couldn't achieve similar with oslo.db as that basically negates any of the slqa migration path for openstack and forces a hard cutover | 15:14 |
| JayF | A project can be compatible with SQLA 2.0 and still work on the older oslo.db; Ironic does for instance. I don't think it's adding any more of a "flag day" than flipping the upper-constraints value would, other than it blocking any oslodb releases until we get there | 15:58 |
| frickler | clarkb: yes, but we should not release a oslo.db version for bobcat that doesn't work with most anything else in bobcat IMO | 15:59 |
| JayF | (and none of this is intended to excuse for putting that newer release up for bobcat) | 15:59 |
| clarkb | JayF: that implies the comment in the chagne isn't correct then | 15:59 |
| clarkb | as it says oslo.db old is only sqla<2 and oslo.db new is only sqla>2 | 15:59 |
| JayF | sqla<2 can be sqla1.4, which is the transition release of sqla which supports both ways AIUI | 16:00 |
| clarkb | JayF: yes. But if oslo.db has a cutover (which the comment says it does) then everyone must update to sqla at the same time (when oslo.db bumps) | 16:01 |
| JayF | Yes; but we all already had to move to sqla 2 at the same time anyway because they are mutually incompatible | 16:01 |
| JayF | (AIUI) | 16:01 |
| clarkb | JayF: not if you are compatible with 1.4 then it is up to the installer aiui | 16:01 |
| JayF | clarkb: I'm just operating under the 'openstack is coinstalled' assumption we generally assert with requirements; you can't run ironic with 2.0 coinstalled with opesntack projects that don't do 2.0 | 16:02 |
| JayF | I don't have full context on all of this, but did just want to make sure there was a specific data point of: we don't need to bump the oslo.db to the version that needs >sqla2 for projects to write sqla2-compatible code | 16:03 |
| JayF | and that AIUI because we assume coinstallability that we were mostly limited to <2 until all the projects were ready | 16:03 |
| clarkb | ya maybe a better way of looking at it is "everyone must be 1.4 compatible using the 2.0 flags" then $distro that is still on 1.4 can run 1.4 and $distro that is on 2.0 can run 2.0. And people doing container installs can pick and choose. | 16:04 |
| clarkb | and then in a year remove 1.x support entirely and switch everyone to 2.0 | 16:04 |
| clarkb | rather than 1.x -> 2.0 right this isntant | 16:05 |
| JayF | I see, you're saying as a project we should allow *operators* a release of slack that works with 1.4/2.0 too | 16:05 |
| JayF | which is reasonable and not something I had even thought about up to this point tbh | 16:05 |
| fungi | well, operators and distributors | 16:05 |
| JayF | downstreams generally is what I meant, yeah | 16:06 |
| fungi | linux distros have similar challenges where they may only want to ship one version of a library | 16:06 |
| clarkb | and maybe that side of the issue is relatively small to doing the updates whwether you do them in 1.4 mode or on 2 proper | 16:06 |
| fungi | so other things that are not openstack may also not be ready for sqla v2 | 16:06 |
| clarkb | *small compared to | 16:06 |
| JayF | fungi: I'm a gentoo contributor and everytime they package openstack-related something I get a grumble about our requirements, so I'm extremely aware :/ | 16:06 |
| fungi | you have a target painted on your back | 16:07 |
| fungi | a big red squarish one | 16:07 |
| JayF | Of course I do, I am in that community and I can speak Enterprise Linux(tm) | 16:07 |
| JayF | :D | 16:07 |
| clarkb | but to do what I describe you need sqla that works with both (that exists its the 1.4 transition series) and oslo.db that exists with both and I'm not sure this is actually a thing | 16:08 |
| clarkb | it might be that oslo.db supports the 2.0 things in 1.4 properly and it is a non issue. I don't know | 16:09 |
| JayF | I suspect that's the case and the commit was badly worded | 16:09 |
| JayF | and stephenfin is likely the person who'd know, and he's not in here | 16:09 |
| JayF | I /invited him a few minutes ago but he won't have context | 16:09 |
| JayF | Changing topics; for release notes this cycle we expect the 2023.1 and 2023.2 release notes to be together in the 2023.2 release notes page, yes? | 16:29 |
| JayF | That's how it built for me locally in Ironic when testing; but then in the gate https://review.opendev.org/c/openstack/ironic/+/895007 it seems to only have 2023.2 changes in the release notes | 16:29 |
| fungi | i thought it was the other way around, 2024.1 (slurp) release notes should include the 2023.2 (previous not-slurp) entries | 16:30 |
| fungi | or did we decide that was going to be too hard to differentiate and started just always including the previous cycle's notes regardless of slurp/not-slurp status? | 16:31 |
| JayF | Honestly I was pretty confident yesterday we were always including, but seeing the different behavior in the gate I can't find anything written about this in governance | 16:32 |
| JayF | which likely means I can't find it, not that it doesn't exist | 16:32 |
| fungi | but anyway, the idea was that people upgrading from 2023.1 to 2024.1 (slurp to slurp) shouldn't have to look at more that one release notes page | 16:32 |
| gmann | tc-members: retired winstacker project content removal are up for review and 2 tc-members can verify and merge them to proceed next - https://review.opendev.org/c/openstack/networking-hyperv/+/894413 https://review.opendev.org/c/openstack/os-win/+/894411 https://review.opendev.org/c/openstack/oswin-tempest-plugin/+/894414 https://review.opendev.org/c/openstack/compute-hyperv/+/894412 | 16:32 |
| JayF | oh yeah, this is the mid-slurp release, because it's Y->A->C | 16:32 |
| JayF | I was confused, I am no longer confused, thank you fungi | 16:32 |
| fungi | np | 16:33 |
| JayF | gmann: I'll add them to my list, thank you | 16:33 |
| gmann | JayF: thanks | 16:33 |
| JayF | probably just a few minutes, I wanna get the Ironic prelude stuff fixed | 16:33 |
| JayF | gmann: +2'd all of them :( Thank you for taking that on. If there's further legwork needed to help those retire, feel free to delegate some in this direction if you want | 16:47 |
| gmann | thanks, pretty much all is up but sure https://review.opendev.org/q/topic:retire-winstackers | 16:47 |
| gmann | after content removal it is infra cleanup and removal from requirements | 16:48 |
| -opendevstatus- NOTICE: The lists.airshipit.org and lists.katacontainers.io sites will be offline briefly for migration to a new server | 16:48 | |
| gmann | and deps is being done separately as they need at least one cycle deprecation phase | 16:48 |
| gmann | tc-members: also this one, anyone would like to hit +W as it has enough +2 https://review.opendev.org/c/openstack/project-team-guide/+/894807 | 17:13 |
| JayF | +2A | 17:22 |
| JayF | thank you for the ping | 17:22 |
| gmann | thanks again :) | 17:22 |
| opendevreview | Merged openstack/project-team-guide master: Add call for Extra AC in PTL guide https://review.opendev.org/c/openstack/project-team-guide/+/894807 | 17:32 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!