[09:34] smoser: blackboxsw hey guys, was busy last week but was able to update the PR#721, can take a quick look at it if it's missing anything else? Thanks a lot :) === cpaelzer__ is now known as cpaelzer [15:01] Hi, NetworkData is always performed before UserData ? in other words, if UserData did the last line of it, i can be sure that NetworkData was already finished ? Thanks [15:02] (without looking on boot-finished) [15:06] oshoval: User data is applied to the system by various config modules (https://cloudinit.readthedocs.io/en/latest/topics/modules.html is the full list). These run at different points in boot for a variety of reasons; one important one you've identified is that some require networking whereas others don't. The modules that run in the `init` stage will run before networking is up. The modules which run [15:06] in the init stage are configured in /etc/cloud/cloud.cfg; you can see the template used to generate that file upstream here (with the appropriate lines highlighted): https://github.com/canonical/cloud-init/blob/master/config/cloud.cfg.tmpl#L55-L82 [15:06] oshoval: So: no, network configuration is not always performed before user-data is processed. :) [15:09] Odd_Bloke: Thanks :) [15:18] :) [16:49] focal daily ppa isn't happy: Reading package lists... [16:49] W: GPG error: http://ppa.launchpad.net/cloud-init-dev/daily/ubuntu focal InRelease: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY 083D06FBE4D304DF [16:49] E: The repository 'http://ppa.launchpad.net/cloud-init-dev/daily/ubuntu focal InRelease' is not signed. [17:03] rharper: A rebuild earlier fixed that, I think. [17:04] Ah, ok; thanks [17:12] hi folks. is the release going ahead today? [17:23] minimal: Yep, we're lining up a couple of last things to perhaps land currently. [17:27] Odd_Bloke: any chance you could look at #817? Its minor so no fuss if it doesn't make this releases - equally as its minor it would be easy to slip into the release :-) [17:27] ah, it's time for my birthday release, and i didn't even do anything for it. [17:30] minimal: I think it's going to miss, I'm afraid; I recall conversations around modifying that output before and IIRC there were concerns about changing that logger name. We probably _can_ do it, but I think we'll regret slipping it in before getting to the bottom of that. [17:33] Odd_Bloke: no problem. Just noticed your comment on #811. Don't actually think they're compilimentary - there's no need to add a new config to disable the output completely when it can be achieved using existing config if #817 is applied. [17:35] perhaps add a new value for ssh_fp_console_blacklist and ssh_key_console_blacklist of "all" then? [17:57] minimal: I'm not super-keen on having a magic string (albeit a fairly straighforward one); YAML has a boolean type, which more closely matches what we want here. We'd also need to consider special handling so that people passing `ssh_fp_console_blacklist: all` don't end up disabling "a", "l" and "l" (and therefore, of course, disabling nothing). [17:59] Oh, let me reply on the PR. [17:59] Odd_Bloke: Thanks. Your call, just though that "all" was a logical case of a "which keys do you want to blacklist" setting [18:03] minimal: Your input is much appreciated. :) You can reserve the right to say I told you so. ;)# [18:04] Odd_Bloke: you get paid the big bucks to make the important decisions ;-) [19:03] falcojr: I've pushed integration tests to https://github.com/canonical/cloud-init/pull/811; could you take another look? [19:14] Odd_Bloke: approved with nit [19:22] Thanks! [20:26] falcojr: I'm seeing one of those integration tests on #800 fail for me locally, I'm investigating. [20:31] Odd_Bloke: I had to build a deb and set it as CLOUD_INI_SOURCE first [20:31] *CLOUD_INIT_SOURCE [20:31] Yep, have done that. [20:32] (_Both_ fail if you don't. :p) [20:32] heh, good point [20:33] I'm also not seeing /tmp/cloud_init_logs populated for these tests. [20:35] hmmm, yeah, that's part of the client fixture...we should probably move that somewhere else [20:38] Aha, right, that'd explain it. [20:38] We probably need to spend a bit of time working out what the interface for non-client-fixture tests look like: we have a good selection to consider now. [20:39] (`setup_image` being the other rough edge.) [20:39] yep...we kind of YOLOed it as we went [20:40] falcojr: Remember how there's a pycloudlib change required for these tests. *facepalm* [20:41] hah, was gonna mention that at first but didn't think one would pass without it [20:41] Odd_Bloke: wait before running a new test [20:41] about to push a commit that adds looking for devices/mounts [20:42] falcojr: Will do; yeah, the hardcoded instance type works for one of them. [20:43] Odd_Bloke: commit pushed, please review too :) [20:48] falcojr: Reviewed. :) [20:49] * falcojr puts on regex hat [20:49] falcojr: `readlink` may be your friend for this? [20:50] Or maybe `realpath` is better, according to readlink's manpage. [20:50] ah yeah, that will be. Thanks! === bswinnerton8 is now known as bswinnerton [21:20] folks, CX has finally approved posting the cloud-init logs on bug 1913354 [21:20] Error: Launchpad bug 1913354 could not be found [21:37] Odd_Bloke: updated [22:14] falcojr: Your review of https://github.com/canonical/cloud-init/pull/820/ for would be great; I think we probably also want blackboxsw to take a look too, given that this is my first time driving the process. [22:15] "for would" forsooth [22:16] falcojr: https://github.com/canonical/uss-tableflip/blob/master/doc/upstream_release_process.md is the process I've followed, to review I'd expect you to do the same and confirm that the output looks as I've pushed (modulo the described manual modifications). [22:17] will do