| meena | I have a question about the init scripts: Is there every any reason to only enable one of the cloud-init daemons? | 10:10 |
|---|---|---|
| === esv_ is now known as esv | ||
| minimal | meena: you mean scripts rather than daemons? There's only 1 "daemon" with cloud-init, the hotplug script | 13:46 |
| blackboxsw | meena: not that I can think of. each init script corresponds to the 4 boot stages. Different parts of cloud-init discovery or configuration run during each of the boot stages. If you don't run init-local or init stages you risk not discovering certain datasources which are only detected in init-local or init stage. If you skip modules-config or modules-final you don't apply parts of the user-data configuration in the cc_* modules. | 14:13 |
| meena | blackboxsw: aye. | 14:36 |
| meena | meena: what's the hotplug script? and does that work on not-linux? (does it even work on Alpine?) | 14:36 |
| minimal | meena: it's for handling hotplugged network interfaces, i.e. on AWS | 14:38 |
| minimal | the shipped script does not work on (base) Alpine as it is written in Bash but I have packaged a derivative that runs on sh/ash | 14:38 |
| meena | how does a script run as daemon? | 14:45 |
| minimal | meena: in the sense that it runs and keeps running rather than runs for a few seconds and exits | 15:05 |
| minimal | it listens on a pipe (from memory) and processes data from it | 15:05 |
| meena | ah, | 15:05 |
| minimal | basically (again from memory) it listens for udev events and reacts to them | 15:06 |
| meena | minimal: this thing? https://github.com/canonical/cloud-init/blob/HEAD/tools/hook-hotplug | 15:09 |
| minimal | yeah | 15:10 |
| meena | how do people manage to write 20 lines of shell code, that's not sh compatible? | 15:13 |
| minimal | https://git.alpinelinux.org/aports/tree/community/cloud-init/03-hook-hotplug-sh.patch | 15:14 |
| meena | minimal: oh, right, no arrays, … brrr… | 15:18 |
| meena | but, still, easier to fix that than expected | 15:18 |
| meena | how come you didn't push it upstream? | 15:18 |
| minimal | I haven't tested in yet, hotplug is only support by Ec2, ConfigDrive2, and OpenStack datasources from memory | 15:19 |
| minimal | meena: in a related note I have a patch for doas support which I haven't upstreamed as doas is fragmented (Alpine is using opendoas) and it relies in functionality only currently present in opendoas | 15:23 |
| meena | meena: boooo | 15:25 |
| meena | i mean, that sounds great | 15:25 |
| minimal | basically doas only supports a single config file /etc/doas.conf whereas opendoas added also support for a directory where multiple files can be read from - so my doas support acts in a similar way to c-i sudo support of creating file in /etc/doas.d/ | 15:26 |
| meena | speaking of sudo / doas: somebody should really make the sudo thing idempotent | 15:27 |
| meena | https://gist.github.com/620feecb55db46cdeab5c12fa986ae61 | 15:28 |
| minimal | doesn't the contents of /etc/sudoers.d/90-cloud-init-users get replaced rather than appended to? | 15:33 |
| minimal | haven't checked the code yet... | 15:34 |
| meena | not on my machines | 15:34 |
| meena | every time i run cloud-init clean -sl i get a new entry | 15:34 |
| minimal | add, the code appends to file is it exists | 15:36 |
| minimal | function write_sudo_rules in distros/__init_.py or freebsd.py, netbsd.py | 15:37 |
| meena | minimal: indeed, i just saw that, too | 15:39 |
| meena | so it's a feature | 15:39 |
| meena | I think we should read the file, and if the contents is the same as what we have, skip writing it | 15:41 |
| minimal | I think just write the file regardless (unless the intention was for the image creator to prepopulate the file, in which case couldn't user-related entries in /etc/clould/cloud.cfg achieve the same?) | 15:42 |
| minimal | write as in overwrite | 15:42 |
| meena | yeah, i think overwrite would be less… surprising. | 15:46 |
| minimal | meena: BTW I noticed you'd raised a launchpad issue some time ago about freebsd support for cc_ca-certs | 15:47 |
| minimal | I'm doing some work with that code currently | 15:47 |
| meena | minimal: reminder to myself, to implement it | 15:47 |
| minimal | is there a clear document as to how freebsd handles certs? | 15:47 |
| minimal | I'm already trying to figure out if the existing RHEL support is fully functional or not | 15:48 |
| meena | minimal: usually, our man pages are pretty good ;) https://man.freebsd.org/certctl | 15:48 |
| meena | and if not, that's a bug. | 15:48 |
| minimal | ok, i'll have a look and try and add FreeBSD support while I'm at all | 15:49 |
| meena | it's on my radar, but my radar is currently beeping out of control | 15:50 |
| minimal | well I'm rejigging the code anyway so it's not much more effort to try and add it | 15:51 |
| minimal | plus wouldn't want my changes to make it any harder to add other os support in future | 15:51 |
| minimal | once I get a PR together then you can test FreeBSD support for me ;-) | 15:52 |
| blackboxsw | ok one more upstream release published to ubuntu Lunar: [ubuntu/lunar-proposed] cloud-init 22.4.1-0ubuntu1 (Accepted) thanks falcojr | 21:38 |
| blackboxsw | falcojr: plan is to now queue these uploads for ubuntu/kinetic|jammy|focal|bionic. so that when our 22.4 SRU clears the queue, we ping S-R-U release vanguards to assess and approve the 22.4.1 upload containing only the fix https://github.com/canonical/cloud-init/pull/1853 | 21:39 |
| -ubottu:#cloud-init- Pull 1853 in canonical/cloud-init "net: skip duplicate mac check for netvsc nic and its VF" [Merged] | 21:39 | |
| falcojr | blackboxsw: so we can just let them sit in approved queue until current SRU clears? | 21:40 |
| blackboxsw | @falcojr we can do either. let them sit or I can upload once we vet the ubuntu/* branches and we will only ping from S-R-U review once we've cleared solutions QA. Probably best if we ping/await solutions QA results on 22.4 before muddying the upload queue to unapproved (B, F, J, K) series | 21:42 |
| falcojr | I'll see if there's any updates from them, but tomorrow is technically the end of the waiting period | 21:43 |
| blackboxsw | yeah let's double check. I really don't want to push a release out if solutions QA doesn't have a +1 for us here | 21:43 |
| blackboxsw | especially given upcoming US-holiday | 21:44 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!