| === meena3 is now known as meena | ||
| blackboxsw | falcojr: holmanb LXD looks agreeable on implementation approach for dynamic instance-id changes which can trigger cloud-init re-run per boot for this cycle. https://github.com/lxc/lxd/issues/9814 so that should make sure we don't get blocked on anything LXDDatasource this cycle | 18:37 |
|---|---|---|
| ubottu | Issue 9814 in lxc/lxd "lxd API: request 1.0/meta-data.instance-id changes on any updated network, device or user config instance data change." [Open] | 18:37 |
| blackboxsw | Also for 22.04 Jammy: cloud-init-related LXD metadata and templates will likely be dropped per this bug as we expect Jammy to use the LXDDatasource by default instead of NoCloud. https://bugs.launchpad.net/cloud-images/+bug/1958460. With those two items solved. I think we are in good shape for driving NIC hot-plug based on LXD dev API 1.0/devices content. | 18:41 |
| ubottu | Launchpad bug 1958460 in cloud-images "LXD image metadata: No longer define nocloud-net templates in compressed LXD image metadata" [Undecided, Confirmed] | 18:41 |
| blackboxsw | holmanb: I made a couple of points on the ubuntu/bionic PR thanks for that. I know it's late and we can work it Monday. I'll have a concrete procedure we can talk over with falcojr now that I'm refreshing my context on this. | 21:08 |
| blackboxsw | that said we may have to rely on another new-upstream-snapshot despite your best efforts here because of the formatting drift introduced by running black on the code. | 21:12 |
| holmanb | blackboxsw: Thanks for the review, I can tackle that Monday. We can chalk this attempt up to a learning experience: "learn to love quilt, the bionic way!" ;) | 21:15 |
| falcojr | blackboxsw, I responded in the PR | 21:25 |
| blackboxsw | falcojr: good point and good image. and yes I do think `new-upstream-snapshot` will take us into the same minor diffs that Brett has already committed as well as retain the ability to build on upstream/ubuntu/bionic (without the need for an upstream/daily/bionic branch) | 21:39 |
| blackboxsw | As I'm walking through the quilt push/pop locally in ubuntu/bionic there is no way without running new-upstream-snapshot that our quilt patches in ubuntu/bionic will be able to ally locally in the branch and once main is merged because we didn't consume the formatting related changes locally in ubuntu/bionic yet. | 21:41 |
| blackboxsw | So, yes I think in this case we might need to run new-upstream-snapshot. And in the future, run new-upstream-snapshot anyway for any unreleased changes or commits which prevent daily builds from succeeding. Ultimately we are going to end up officially releasing all commits into stable Ubuntu releases anyway with or without minor debian/patches to retain original behavior for certain new features. | 21:44 |
| blackboxsw | Frequent new-upstream-snapshots in ubuntu/bionic doesn't preclude your idea of us releasing from hotfix branches if needed before we go through a full SRU process as our subsequent normal SRU will inevitably contain all upstream commits which includes the hotfix branch commit. | 21:45 |
| blackboxsw | ok, ok. so yes, let's fixup brett's branch by running new-upstream-snapshot which will allow us to build the unreleased content locally from upstream/ubuntu/bionic. And if we run into an unlikely feature request or hot-fix need, we can manage that with a hotfix branch as you mentioned. | 21:47 |
| falcojr | +1, I think that's a good plan | 21:49 |
| blackboxsw | ok falcojr holmanb ready for monday. build procedue on brett's branch looks good post a new-upstream-snapshot https://github.com/canonical/cloud-init/pull/1203#pullrequestreview-860094619 | 22:43 |
| ubottu | Pull 1203 in canonical/cloud-init "Fix patches" [Open] | 22:43 |
| holmanb | Sounds good to me, thanks for the help on this blackboxsw and falcojr | 22:52 |
| holmanb | falcojr: Thanks for the review on PR #1160. I responded to all of your comments. All except the async PUT question and the suggested tests should be a quick fixes IMO. blackboxsw: would you prefer to get in a review on it before I move forward with testing on AWS? | 22:56 |
| falcojr | holmanb: we might want to figure out the api token part first | 22:57 |
| falcojr | I'm not aware of how that works, so don't have any specific thoughts about it yet | 22:57 |
| holmanb | falcojr: sure that's fair | 22:58 |
| blackboxsw | holmanb: I'll get eyes on it by EOW, but won't have time for a full review. | 23:04 |
| holmanb | ok | 23:06 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!