@bastelfreak thanks for quick reply! • I know that the module is old, and its usefulness becoming outdated due to network-scripts being optional in CentOS 8 and not at all in CentOS 9 (I use CentOS in my envs), however while I do have a custom class to switch to networkd (as we have discussed elsewhere) I need to continue to support network-scripts for the meantime • I prep and deploy to physically different environment, so using R10k directly is not feasible, though I can add the

puppet generate types

command in my deploy script • I do have the puppet/filemapper module installed, and the issue self-corrects after a few runs, wo it seems like a resource actualising/timing thing? • Thanks for adding info for what

generate types

does, but I'm still unclear on why it solves the problem and allows successful runs from the first time

@bastelfreak thanks for quick reply! • I know that the module is old, and its usefulness becoming outdated due to network-scripts being optional in CentOS 8 and not at all in CentOS 9 (I use CentOS in my envs), however while I do have a custom class to switch to networkd (as we have discussed elsewhere) I need to continue to support network-scripts for the meantime • I prep and deploy to physically different environment, so using R10k directly is not feasible, though I can add the

puppet generate types

command in my deploy script • I do have the puppet/filemapper module installed, and the issue self-corrects after a few runs, so it seems like a resource actualising/timing thing? I have confirmed this behaviour in different settings • Thanks for adding info for what

generate types

does, but I'm still unclear on why it solves the problem and allows successful runs from the first time

do you have different versions of the network or filemapper module in your environments maybe?

no, the dependencies are also deployed from dev to prod environments

the whole filemapper and boolean module is so ancient and useless nowadays, there's maybe a single person that knows how it works

ouch!

I don't recommend systemd-networkd without a reason 🙂

yeah good call. I guess unless I get other info I will just add

puppet generate types

to my deploy script and try to bring forward migration to networkd

yes, that should always be executed after a deployment

https://www.puppet.com/docs/puppet/6/environment_isolation.html also some background

sorry, are you advising that running

puppet generate types

on a deployment IS best practice regardless?

yes

ahh, thanks

that's also the default in puppet enterprise

can I get a review for https://github.com/voxpupuli/puppet-varnish/pull/15 plz. Ty

can I get a review for https://github.com/voxpupuli/puppet-varnish/pull/15 plz. Ty

https://github.com/voxpupuli/voxpupuli-acceptance anyone knows why it lists 2 packages, but those end up on the same page? is that a GH bug?

ok. So it's taken me a while, but I finally got around to unbreaking

ensure_packages

in stdlib. https://github.com/puppetlabs/puppetlabs-stdlib/pull/1300 🙂

ok. So it's taken me a while, but I finally got around to unbreaking

ensure_packages

in stdlib. https://github.com/puppetlabs/puppetlabs-stdlib/pull/1300 🙂

Are there any plans on making vox rubygems ruby 3.2 compatible? e.g. like https://github.com/voxpupuli/puppet-blacksmith/pull/107

Are there any plans on making vox rubygems ruby 3.2 compatible? e.g. like https://github.com/voxpupuli/puppet-blacksmith/pull/107

I guess for all the gems that use rubocop, this specific issue wouldn't exist?

there is indeed a cop to do that

I like the look of the new regex timeout stuff. Puppet could definitely make use of this. At the moment, it's very possible to accidentally denial of service your own puppetserver with a

Pattern

data type.

See https://tickets.puppetlabs.com/browse/PUP-11213

Wow. Ruby 3.0 dropped

URI.escape

The update to Puppet 8 is going to become interesting.

it certainly is

wait what?

you know the loud deprecation warnings that Ruby 2.7 shows for URI.escape? those are for Ruby 3

stares awkwardly in ruby 2.5...