| holmanb | @blackboxsw: I put up a PR to fix a couple of things that cause the ansible integration test to always fail on older systemd&&python versions. Additionally it tweaks the repo setup code in a way that should make it more robust and makes failing assertions more useful for debugging. https://github.com/canonical/cloud-init/pull/1691 | 00:46 |
|---|---|---|
| === ananke_ is now known as ananke | ||
| tilt | hi :) | 07:37 |
| tilt | if a vmware instance of ubuntu20 does not use the vmware datasource, what could be wrong / are there specific things to be done to make it use this datasource? | 07:38 |
| tilt | like, a special boot parameter oslt, can't find it in the docs | 07:38 |
| tilt | i think we got it, we did not follow the customization steps described in https://williamlam.com/2022/06/using-the-new-vsphere-guest-os-customization-with-cloud-init-in-vsphere-7-0-update-3.html | 07:53 |
| esposem | if I upgrade cloud-init to a new version and then reboot, is my cloud.cfg file updated too or left intact? I would say intact, as there could be custom settings in it | 13:15 |
| falcojr | I think that's one of those situations where with apt at least you'd get prompted for which version to use, but it would default to keeping the changed settings | 13:48 |
| falcojr | not sure about other distros | 13:48 |
| esposem | falcojr: thanks, I think for RHEL happens the same too | 14:42 |
| blackboxsw | falcojr: minor changelog omission on https://github.com/canonical/cloud-init/pull/1693 something else the new-upstream-snapshot(git-buildpackage based) needs to fix probably. | 17:39 |
| ubottu | Pull 1693 in canonical/cloud-init "Ubuntu/jammy" [Open] | 17:39 |
| blackboxsw | going through focal now | 17:39 |
| falcojr | blackboxsw: that cherry pick only ever existed in the hotfix branch, so it's not anything new-upstream-snapshot would ever see | 17:48 |
| blackboxsw | falcojr: ohh right, the whole reason we created the separate hotfix branches to avoid the pulling cherry picks into the base upstream/ubuntu/jammy branch great. I thought it implies another short coming in the new-upstream-snapshot changes | 17:50 |
| falcojr | based on what we said previously about not including the hotfix changelog info, I wouldn't think it would need to be included | 17:50 |
| blackboxsw | nevermind nothing to see here. | 17:50 |
| falcojr | but maybe it should? I guess we just need a consistent way to approach it | 17:50 |
| blackboxsw | falcojr: I'm on the fence there. the changeset you have currently, for SRU reviewers, ends up representing this next upload to focal as an aggregation of the previous hotfix22.2-0ubuntu1~20.04.2 being consumed within the 22.3 upstream version https://paste.ubuntu.com/p/Pp6MW3FBvq/ . | 17:54 |
| blackboxsw | so it makes the review of that upload a bit more time-intensive, but I think it's a reasonable representation of the changeset delivered. I guess the problem in SRU review tracking is that if we have a long-lived hotfix branch with multiple releases, eventually we'll lose visibility to those commits/patches and release history in debian/changelog once we new-upstream-snapshot for a full SRU again | 17:56 |
| blackboxsw | but it is strange to represent adding cherry-picks and removing them in the latest changelog when it didn't really happen in this release branch | 17:57 |
| blackboxsw | falcojr: I'd stay let's stick with what you have up for review now as policy. undocument the cherry-picks that were applied to hotfix branches as that content is included in the d/changelog as part of the new-upstream-snapshot anyway | 17:58 |
| blackboxsw | falcojr: actually I'll take the opposite approach I just suggested. I think it'll be easier to review and understand the history of the release debian package if we captured the previously released d/changelog entries and manually recorded the d/changelog sections that were released from any hotfix branches prior to our new-upstream-snapshot section and manually add the dropped patches comment. | 18:02 |
| falcojr | blackboxsw: so I think either all or nothing | 18:03 |
| blackboxsw | reason being SRU reviewer is going to see a debdiff of current upload vs previous published upload to pkg/ubuntu/focal-proposed in git-ubuntu | 18:03 |
| falcojr | if we're not going to record the interim releases, we shouldn't include changes between the interim release and now | 18:03 |
| blackboxsw | and they'll see the dropped cpick file in that diff and no documentation in d/changelog of actually dropping it | 18:03 |
| falcojr | if we are going to include the interim release, then we should also document what changed in the changelog | 18:03 |
| blackboxsw | falcojr: right so I flipped 180. from my prior stance. I think record them all . | 18:03 |
| blackboxsw | add the d/changelog entries verbatim from upstream/ubuntu/focal-22.2-hotfix prior to our top-most new-upstream-snapshot section | 18:04 |
| falcojr | blackboxsw: sounds good | 18:05 |
| blackboxsw | and then manually add the drop cherry-picks line to our new-upstream-snapshot section | 18:05 |
| blackboxsw | reworking my diff suggestion to see if that makes sense | 18:05 |
| blackboxsw | and then policy-wise anytime we have a hotfix release, we also need to cherry-pick the release changelog sections into upstream/ubuntu/<release> base branch in preparation for whenever the next SRU is | 18:06 |
| falcojr | so much time spent on a freakin changelog. It'd be faster if we manually curated it at this point | 18:08 |
| falcojr | (not a rant at you, just a rant :P ) | 18:08 |
| blackboxsw | diff updated https://github.com/canonical/cloud-init/pull/1693#pullrequestreview-1087280009 and yes I did that manually | 18:09 |
| ubottu | Pull 1693 in canonical/cloud-init "Ubuntu/jammy" [Open] | 18:09 |
| blackboxsw | no cherry-pick please. don't waste your time finding the commitish | 18:09 |
| blackboxsw | I'm with you on spending too much time on accounting basically | 18:09 |
| falcojr | blackboxsw: thanks, you're quicker on that than me | 18:09 |
| blackboxsw | falcojr: my patch was for focal, but I put in on the jammy branch :/ | 18:10 |
| blackboxsw | will add that diff to focal.... updating jammy now | 18:10 |
| falcojr | heh, all good. I'll get them worked out | 18:10 |
| blackboxsw | +1 | 18:10 |
| blackboxsw | faster and "wronger" | 18:11 |
| blackboxsw | patches corrected for focal and jammy | 18:18 |
| blackboxsw | rebuilding now to confirm no lints | 18:18 |
| blackboxsw | jammy w/ patch no debian/* diffs unaccounted for in d/changelog https://paste.ubuntu.com/p/sRgddwSsTP/ | 18:19 |
| blackboxsw | and makes our topmost d/changelog section "clean" when reviewing the supplemental diff | 18:19 |
| falcojr | blackboxsw: actually now I have duplicated changelog entries | 18:23 |
| falcojr | I should remove the postinst redact from 22.3 since we have it in the hotfix, right? | 18:24 |
| blackboxsw | falcojr: good pt. yes remove the duplcated lines that show up again in the top-most section. (because they were already officially released via hotfix publish)( | 18:24 |
| falcojr | blackboxsw: I think all 3 are up to date now | 18:28 |
| blackboxsw | ok today I'm documenting the hotfix process for uss-tableflip:doc/ubuntu_release_process.md too | 18:30 |
| blackboxsw | we had a breadcrumb in there, but not enough | 18:30 |
| blackboxsw | we had a breadcrumb in there, but not enough ... and nothing in cloud-init RTD content... probably want that too | 18:31 |
| blackboxsw | at least from a tag perspective | 18:31 |
| blackboxsw | so others can track whether upstream releases have followup hotfixes | 18:31 |
| falcojr | blackboxsw: part of it is that we're doing some new things so making it up as we go | 18:32 |
| blackboxsw | +1 | 18:32 |
| blackboxsw | bionic/focal/jammy uploads good | 19:23 |
| blackboxsw | PRs & tags pushed to upstream | 19:24 |
| holmanb | . | 19:34 |
| falcojr | blackboxsw or holmanb: if either of you are still around we have a request in #ubuntu-release (pinged here because we're all here) | 23:03 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!