[11:24] <SDes91> Is there a way to avoid a lock-out on first boot with cloud-init? I am trying to create a config for users for raspberry pi and since the boot messages go by really quickly I cannot see what went wrong and cannot login into the device
[11:26] <waldi> SDes91: if you are this deep, yank the sd card and read the log afterwards?
[11:58] <SDes91> waldi that works thanks.
[11:59] <SDes91> Is there a better way to actually force cloud-init to create an admin user for me. I keep getting an error  `useradd: group admin exists - if you want to add this user to that group, use -g.`
[12:18] <SDes91> answer to my query setting `no_user_group: true` will create an `admin` user
[13:07] <CorvetteZR1> hello.  how do i edit an existing cloud-init config?
[13:08] <CorvetteZR1> i set up a mount with noexec option.  i removed it manually from fstab and remount; but when i reboot the server, i see noexec in fstab again.  how can i modify it in cloudinit, so when it overwrites fstab, it no longer applies noexec; without having to reprovision the server?
[13:14] <meena> Does the bi-weekly meeting still happen? I'm not seeting any updates in about a year now: https://cloud-init.github.io/
[13:15] <meena> CorvetteZR1: hi. How you edit cloud-init config heavily depends on which cloud you're on. Most have web interfaces. and many don't allow editing once the machine is created.
[13:16] <CorvetteZR1> meena, ah, i see.  this is in azure.  i was wondering if it can be done in the vm os itself?
[13:17] <meena> Nope. cloud-init fetches it anew before applying it
[13:19] <CorvetteZR1> ok, gotcha
[13:20] <CorvetteZR1> so i guess next question is will azure let me modify it then?  i guess that's out of scope for this channel, but if anyone knows, that'd be very helpful :)
[13:23] <meena> CorvetteZR1: yeah, my experience is mostly with niche clouds
[13:24] <meena> OpenStack, Nebula, libvirt (lol), DigitalOcean and Hetzner
[13:24] <meena> Soon to change…
[13:29] <CorvetteZR1> meena, ok, thanks.  i just searched for it, and according to azure documentation you can't change it for single vm deployments; only scale sets which doesn't apply in my case.  all good though, thanks for your help
[13:37] <waldi> CorvetteZR1: you can disable the correct module in /etc/cloud/cloud.cfg
[13:38] <CorvetteZR1> waldi, good to know, thanks!
[13:39] <waldi> azure is special as it re-runs the mounts module on every boot and even have some kind of workaround for it
[13:43] <minimal> SDes19: don't forget if you're specifying a "users:" section in your user-data that you need to include "- default" also otherwise the default user won't be created
[13:43] <minimal> SDes19: BTW I use cloud-init for initial config of my Raspberry Pis :-)
[13:55]  * meena is back to contributing to cloud-init
[13:56] <meena> minimal: of course you do…
[13:56] <meena> minimal: which OS do you put on them?
[13:59] <SDes91> minimal I don't require a default user, I am relying on the `users` to be configured appropriately to be made accessible for the Pis
[14:00] <SDes91> i.e. `admin` user will be the one configured to access via SSH and other users are anyways locked through password so they can't access the Pis via SSH
[14:00] <minimal> meena: Alpine Linux
[14:01] <meena> that… sure is… minimal 
[14:02] <minimal> SDes91: ok, was just pointing out a common issue regarding the default user (which is defined in /etc/clodu/cloud.cfg)
[14:02] <minimal> meena: yupe, my Alpine VMs with cloud-init installed are around 200-300MiB in size :-)
[14:03] <SDes91> minimal do you happen to know if the Ubuntu Autoinstall configuration for ubuntu Server works for RPis? https://ubuntu.com/server/docs/install/autoinstall-reference
[14:05] <minimal> SDes91: no idea, never tried it, but I'm sure one of the Canonical guys can comment
[15:35] <falcojr> meena: meeting no longer happens. We didn't really see any engagement or value from it. If you found value in it, I think we'd be open to bringing it back
[15:44] <meena> falcojr: the main reason i couldn't participate, even when i was working on cloud-init (tho not paid) was the awkward time
[15:45] <meena> but i do think that just looking across issues and PRs and the (very quiet) mailing list should quickly bring someone up to date.
[15:46] <falcojr> meena: Yeah, PRs are probably the best way to get a sense of what has happened. I think the meeting content was more or less a summary of those
[15:53] <meena> anyway, in case y'all missed it: my project with the FreeBSD Foundation has been approved, and I'll be working over the next six or seven months to bring better *BSD support to cloud-init
[15:54] <meena> (ambitiously calling this: Making FreeBSD a Tier-1 Platform)
[15:58] <meena> the first step is to overhaul our networking organisation, as proposed here: https://github.com/canonical/cloud-init/blob/main/WIP-ONGOING-REFACTORIZATION.rst
[15:59] <blackboxsw> > "If you found value in it, I think we'd be open to bringing it back"        Yeah I'd +1 that if desirable since meena is going to be organizating some significant 'community' efforts that may need to involve a lot of communication.  The awkward time could be "resolved" my just posting updates/announcements to discourse and let comments/questions flow through that at folks' reasonable waking/working hours.
[16:00] <blackboxsw> *resolved by just posting to discouse for async communication/updates*
[16:01] <falcojr> meena: Glad to have you back on the project!
[16:04] <meena> i'd be very happy to finish what i started *checks notes* 4 years ago.
[19:16] <meena> I have a question of… Form…
[19:17] <meena> In the past, when the FreeBSD Foundation sponsored work on external projects, people added Sponsored by: The FreeBSD Foundation to their commit messages. See for example: https://github.com/llvm/llvm-project/commit/bdb4468d39496088fc05d8c5575647fac9c8062a
[19:17] <ubottu> Commit bdb4468 in llvm/llvm-project "[gdb-remote] Move broadcasting logic down to GDBRemoteClientBase"
[19:17] <meena> Is it okay for me to do that in my PRs?
[21:00] <falcojr> meena: yeah, that should be fine. Other organizations do similar things
[21:01] <falcojr> well, maybe not similar, but not uncommon to see downstream bug trackers or signed off statements
[22:01] <blackboxsw> cjp256: open question for you as to how Azure wants to handle potential absence of IMDS https://github.com/canonical/cloud-init/pull/1725#issuecomment-1244558085
[22:01] <ubottu> Pull 1725 in canonical/cloud-init "define new attribute for pre-22.3 pickles" [Merged]
[22:09] <cjp256> thanks blackboxsw! I will get on it first thing in the morning