[01:34] <minimal> Lvl4Sword: you mean the user-data?
[01:47] <Lvl4Sword> minimal, ?. I mean without a machine. Such as a website like regex101 or something to verify.
[02:35] <minimal> Lvl4Sword: to verify what exactly? the user-data YAML you are providing?
[02:42] <holmanb> LVL4Sword: It's not clear what you are talking about: user configurations, code, or documentation. We use static analyzers (black, pylint, flake, mypy) which would catch code typos in Python. We have some spell checkers for the documentation. We have some tooling that validates user user-data cloud-config. We have some code tests that catch a variety of things. To which of these do you refer?
[02:45] <minimal> holmanb: I assume he's referring to user-data he had hasn't been clear. For user-data beyond using "cloud-init schema" there's no real solutions, right?
[02:45] <minimal> s/had /just /
[02:51] <holmanb> Yeah I think you're right.
[02:54] <holmanb> I mean a script could be statically validated by external checkers (shellcheck, for example), but since scripts are freeform that would have to be validated with something external. I guess we could maybe add something for Jinja, to see if the template is valid (just a basic load with a try/except would be all it would take).
[02:57] <minimal> I meant more than the likes of yamllint will indicate if a user-data YAML file is valid YAML, but doesn't check that it is valid user-data
[02:57] <minimal> s/than/that/
[14:26] <Naaaa> I understand how cloud-init works on initial boot of a system, but is cloud-init supposed to be applied every time a system reboots?
=== dbungert is now known as Guest7975
[14:52] <meena> Naaaa: there are modules that run on every boot
[14:53] <meena> and depending on the cloud provider, this could be divergent parts: Azure for instance wants the network configured every boot, while most other clouds do that once per instance
[14:54] <meena> every module's documentation tells you when it runs
[15:40] <anatoly79> hello
[15:40] <anatoly79> I'd like to test my user-data file by running cloud-init in a temp direcory or docker and apply only "apt" section
[15:41] <anatoly79> how can  do it? what is a command line?
[15:42] <anatoly79> the distro is ubuntu 22.04
[15:42] <minimal> test the syntax of your user-data or test performing actions/commands based on the user-data?
[15:43] <anatoly79> I want to see mofification to the root fs
[15:43] <anatoly79> I know I can do it with VM, but using docker (or tmp dir0 is faster
[15:44] <minimal> you'd have to run cloud-init to see the results
[15:45] <minimal> where is your user-data coming from? a cidata ISO?
[15:45] <anatoly79> a local file
[15:46] <minimal> don't think you can test that then - where cloud-init gets the user-data from depends on the particular DataSource used
[15:46] <anatoly79> I tried this: 1) run ubuntu 22.04 in a docker 2) install cloud-init 3) run cloud-init --file user-data.yml init
[15:47] <anatoly79> but this command tries to modify disk (run udevadm)
[15:47] <anatoly79> I'd like to run only "apt" section
[15:47] <minimal> does your user-data have only a apt section?
[15:48] <anatoly79> yes, I have a version with only "apt" section. cloud-init still tries to do the other things
[15:48] <minimal> well cloud-init doesn't just handle user-data
[15:49] <minimal> it sets up network etc
[15:49] <minimal> so it is not clear what exactly you are trying to achieve
[15:49] <anatoly79> can I tell it to do only "apt" stuff?
[15:50] <minimal> well if it doesn't setup network then how can it "do" only apt stuff (which likely involves a network connection)?
[15:51] <minimal> cloud-init is not really designed for runnning in a docker "application" container, though it can run in a LXC "OS" container
[15:52] <anatoly79> it can just assume that current configuration works, and start from "apt", I guess
[15:52] <minimal> you said you want to do this as it is quicker than using a VM, but you could have used a VM by now and finished your test...
[15:53] <anatoly79> :-)
[15:54] <minimal> basically I think you're trying to use cloud-init in a way it is not designed to be used
[15:57] <minimal> also what are you doing in a "apt:" section that is so complicated that you need to test it?
[15:59] <anatoly79> I'd like to create an ISO with build-in pool of debs for offline (no internet) install. For that I'd like to apply "apt" section  in some isolated env, and download debs
[15:59] <anatoly79> I can do it with my particulat configuration, but I want to do it generic, so applying "apt" section whatever it might be sounded as a good idea
[16:02] <anatoly79> livefs-editor project just uses main mirror
[16:02] <anatoly79> I thought by applying custom apt I can get more options for download