| *** brinzhang0 is now known as brinzhang | 01:36 | |
| *** brinzhang0 is now known as brinzhang | 02:51 | |
| opendevreview | Brin Zhang proposed openstack/os-service-types master: Update api_ref for Cyborg project https://review.opendev.org/c/openstack/os-service-types/+/856373 | 06:26 |
|---|---|---|
| frickler | gtema: seems you updated your zed candidacy instead of submitting one for antelope https://review.opendev.org/c/openstack/election/+/855313 , got me really confused because sdk is still listed as candidateless and I was sure I had seen that patch | 08:07 |
| frickler | also weird none of the officials noticed and it simply got merged | 08:07 |
| *** brinzhang_ is now known as brinzhang | 09:12 | |
| gtema | Yeah, i also noticed that and wrote today morning email to Amy | 10:24 |
| opendevreview | Stephen Finucane proposed openstack/openstacksdk master: compute: Fix '*volume_attachment' proxy methods https://review.opendev.org/c/openstack/openstacksdk/+/821733 | 11:02 |
| opendevreview | Stephen Finucane proposed openstack/openstacksdk master: compute: Add functional tests for volume attachments https://review.opendev.org/c/openstack/openstacksdk/+/822314 | 11:02 |
| opendevreview | Stephen Finucane proposed openstack/openstacksdk master: WIP: compute: Workaround to avoid breaking volume attachment APIs https://review.opendev.org/c/openstack/openstacksdk/+/856298 | 11:02 |
| stephenfin | frickler: wdyt of the approach in that WIP. Total hack, but it should get us where we want to go faster than near-duplicate methods? | 11:03 |
| frickler | stephenfin: oh, wow, that needed some heavy swapping to restore context. but I like that idea, yes | 11:19 |
| stephenfin | heh, yeah, I forgot I'd started on that and never finished it. Will finish that patch today | 11:27 |
| frickler | FYI OSC failure seems triggered by a Neutron patch, I pinged ppl over there https://zuul.opendev.org/t/openstack/build/96c981b70b5640c5ba2c21642fcf2b59 | 11:34 |
| frickler | cf. https://review.opendev.org/c/openstack/neutron/+/855850 | 11:34 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: WIP Use Python 3.10 on Ubuntu 22.04 LTS (Jammy Jellyfish) for Ansible's devel branch https://review.opendev.org/c/openstack/ansible-collections-openstack/+/852526 | 11:48 |
| opendevreview | Stephen Finucane proposed openstack/openstacksdk master: image: Add support for other metadef namespace operations https://review.opendev.org/c/openstack/openstacksdk/+/856271 | 11:51 |
| opendevreview | Stephen Finucane proposed openstack/openstacksdk master: tests: Improve functional testing for image methods https://review.opendev.org/c/openstack/openstacksdk/+/856272 | 11:51 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack stable/1.0.0: Release 1.9.1 version https://review.opendev.org/c/openstack/ansible-collections-openstack/+/856516 | 12:07 |
| stephenfin | gtema: when you've time, can you look at https://review.opendev.org/c/openstack/openstacksdk/+/856271/comments/eb0bc07e_8de9875b I expected there to be an easy way to do this but I haven't figured it out yet | 12:10 |
| opendevreview | Merged openstack/ansible-collections-openstack stable/1.0.0: Release 1.9.1 version https://review.opendev.org/c/openstack/ansible-collections-openstack/+/856516 | 12:56 |
| opendevreview | Lajos Katona proposed openstack/python-openstackclient master: Add address-scope to NDP proxy tests https://review.opendev.org/c/openstack/python-openstackclient/+/856521 | 13:15 |
| dtantsur | gtema: hey, re https://review.opendev.org/c/openstack/openstacksdk/+/849505: is your comment about maximum microversion blocking? in other words, could you check again? | 16:10 |
| gtema | the ball is rolling | 16:11 |
| opendevreview | Dmitry Tantsur proposed openstack/openstacksdk master: Migrate register_machine to use the proxy layer https://review.opendev.org/c/openstack/openstacksdk/+/849402 | 16:24 |
| opendevreview | Rajat Dhasmana proposed openstack/python-openstackclient master: compute: Add support for microversion 2.93 https://review.opendev.org/c/openstack/python-openstackclient/+/831014 | 16:25 |
| dtantsur | I hope I'm done with ^^ | 16:25 |
| opendevreview | Dmitry Tantsur proposed openstack/openstacksdk master: Update register_machine to use the Ironic format for ports https://review.opendev.org/c/openstack/openstacksdk/+/856549 | 16:37 |
| opendevreview | Merged openstack/openstacksdk master: baremetal: rework node creation to be closer to the backend https://review.opendev.org/c/openstack/openstacksdk/+/849505 | 17:31 |
| opendevreview | Rajat Dhasmana proposed openstack/python-openstackclient master: compute: Add support for microversion 2.93 https://review.opendev.org/c/openstack/python-openstackclient/+/831014 | 17:46 |
| jm1 | gtema: is it safe to release the zed version of python-openstackclient with openstacksdk 0.61.0? | 18:14 |
| opendevreview | Emma Foley proposed openstack/openstacksdk stable/yoga: Fix incremental backups handling in project cleanup https://review.opendev.org/c/openstack/openstacksdk/+/856509 | 18:15 |
| gtema | Yes | 18:15 |
| jm1 | gtema: do we have other openstack code except python-openstackclient and ansible openstack collection which uses openstacksdk? | 18:15 |
| gtema | Zuul (nodepool), but it is now pinned, nova (inside) | 18:16 |
| gtema | And maybe some other projects | 18:16 |
| jm1 | gtema: wait, nova is using openstacksdk? will we break nova when releasing zed version of nova with openstacksdk 0.61.0? | 18:39 |
| gtema | Do you want to make proper release of 61 into Zed? I thought you ask about some particular usecase | 18:40 |
| jm1 | gtema: we pinned openstacksdk to 0.61.0 in rdo c9 master (openstack distribution for centos 9) and aoc to 1.8.0 because we have not finished porting aoc to sdk 0.99.0. we are wondering whether this would break any openstack component | 18:43 |
| jm1 | stephenfin: ^ fyi | 18:45 |
| gtema | Generally I thing it should not break, since we always said we are backwards compatible. We even want to get rid of release branches as such. But whether such step may influence nova I can't say. In 61 there were no critical changes, so chances are low. | 18:45 |
| gtema | From code it is hard to say confidently | 18:46 |
| jm1 | gtema: sdk 0.99.0 broke backward compatibility, so when openstack components begin to depent on changes from sdk 0.99.0+, then they could break when used with sdk<0.99.0 (e.g. 0.61.0). | 18:47 |
| jm1 | gtema: until know our rdo/tripleo ci jobs seem to work fine with sdk 0.61.0, aoc 1.8.0 and latest zed code for all openstack components. but this might be simply because we do not have 100% code coverage? | 18:49 |
| gtema | I heavily doubt anybody else was ever relying on cloud layer, since it doesn't support micro versions properly | 18:49 |
| gtema | Well, now it does | 18:49 |
| gtema | But because of that I doubt nova ever wanted to use cloud layer | 18:50 |
| gtema | Hard to say without seeing real usage | 18:50 |
| jm1 | gtema: hmm.. searching on opendev reveals several uses of openstacksdk, not only in nova, so in theory we would have to check that code as well https://codesearch.opendev.org/?q=import%20openstack | 18:56 |
| jm1 | gtema: unpinning openstacksdk and aoc would definitely cause some breakage, e.g. we will have to port tripleo-ansible etc. to new aoc and sdk | 18:57 |
| gtema | Uuuu, so much usage. That's cool | 18:58 |
| jm1 | gtema: yes and no... | 18:58 |
| jm1 | gtema: sdk 0.99.0 in zed is really causing us a headache.. | 18:59 |
| gtema | Tja, I guess you can really try it out in rdo | 19:00 |
| jm1 | gtema: we have quite a bit of ci coverage for tripleo deployments and since they are passing i guess pinning is not a huge issue for now | 19:03 |
| gtema | Yupp | 19:03 |
| jm1 | gtema: anyway, thanks for your take on this. good to see that openstacksdk is actually used more than we thought :) | 19:04 |
| jm1 | gtema: with that search i found another interesting thing: https://opendev.org/openstack/metalsmith/src/commit/cc8fe24d40e64e874bc3d4308cd76e50379729d3/metalsmith_ansible/ansible_plugins/modules/metalsmith_instances.py#L22 | 19:04 |
| gtema | Yeah, right. it's a positive surprise for me as well | 19:04 |
| gtema | 😌 | 19:05 |
| jm1 | gtema: so from skimming through the results it looks like openstacksdk is used in ironic (and related repos), tripleo-ansible, aoc, bifrost, tripleo-heat-templates, zuul nodepool, zuul-registry, networking-baremetal, neutron, openstack-virtual-baremetal | 19:12 |
| * jm1 hit enter while editing 🤦 | 19:13 | |
| jm1 | .., various tripleo related repos and other probably less rdo relevant repos. | 19:13 |
| jm1 | any volunteers to check whether a pinning of openstacksdk affects the zed releases of those repos? | 19:14 |
| jm1 | ;) | 19:14 |
| jm1 | gtema: for nova i cannot see how they use openstacksdk | 19:24 |
| jm1 | gtema: oh.. now i see | 19:25 |
| gtema | Jm1, me neither. It seems to be deep as adapter | 19:26 |
| jm1 | gtema: oh with "from openstack import" i found a couple of more usages: heat, metalsmith, nova, octavia-dashboard, ... | 19:29 |
| jm1 | gtema: lets hope they do not use cloud layer.. | 19:30 |
| gtema | ;-) | 19:31 |
| jm1 | gtema: not to be solved today. /me out for today, have a nice evening! | 19:31 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!