"Vendor dependency"

documented vendor dependency

in this case you should have a link or a local copy

Question about the Puppet AI agent... if I have 2 seperate primaries, do they each need their own Azure OpenAI service or can they both connect to the same one?

ugh... on my workstation (macOS 26.0.1) i removed the older homebrew

pdk

and

puppet-agent

packages, downloaded PDK from the secret perforce server, and used

gem install eyaml

to install eyaml and its dependencies ... now whenever i try to use it, i get this:

$ echo testing | eyaml encrypt --stdin -o block -l xyz
/Users/jms1/.gems/gems/psych-5.2.6/lib/psych.rb:716:in 'File#initialize': Is a directory @ io_fillbuf - fd:7 xyz (Errno::EISDIR)
	from /Users/jms1/.gems/gems/psych-5.2.6/lib/psych.rb:716:in 'IO.open'
	from /Users/jms1/.gems/gems/psych-5.2.6/lib/psych.rb:716:in 'Psych.load_file'
	from /Users/jms1/.gems/gems/eyaml-0.4.4/lib/eyaml.rb:37:in 'EYAML.encrypt_file_in_place'
	from /Users/jms1/.gems/gems/eyaml-0.4.4/lib/eyaml/cli.rb:11:in 'block in EYAML::CLI#encrypt'
	from /Users/jms1/.gems/gems/eyaml-0.4.4/lib/eyaml/cli.rb:7:in 'Array#each'
	from /Users/jms1/.gems/gems/eyaml-0.4.4/lib/eyaml/cli.rb:7:in 'EYAML::CLI#encrypt'
	from /Users/jms1/.gems/gems/thor-1.4.0/lib/thor/command.rb:28:in 'Thor::Command#run'
	from /Users/jms1/.gems/gems/thor-1.4.0/lib/thor/invocation.rb:127:in 'Thor::Invocation#invoke_command'
	from /Users/jms1/.gems/gems/thor-1.4.0/lib/thor.rb:538:in 'Thor.dispatch'
	from /Users/jms1/.gems/gems/thor-1.4.0/lib/thor/base.rb:584:in 'Thor::Base::ClassMethods#start'
	from /Users/jms1/.gems/gems/eyaml-0.4.4/bin/eyaml:7:in '<top (required)>'
	from /usr/local/Cellar/ruby/3.4.7/lib/ruby/3.4.0/rubygems.rb:319:in 'Kernel#load'
	from /usr/local/Cellar/ruby/3.4.7/lib/ruby/3.4.0/rubygems.rb:319:in 'Gem.activate_and_load_bin_path'
	from /Users/jms1/.gems/bin/eyaml:25:in '<main>'

@csharpsteen do you know why postgresql on a primary is listening on *, when I've no replica, no standalone puppetdb. no compilers? That seems unnecessary / like a security issue to me. Tested on 2025.6.0 EL9

question: is there any way to "inspect" the contents of a

suite-license.lic

file, without installing it on a PE server and checking the web interface? (specifically, to see the start/end dates) ... it looks like an SSL certificate, i'm hoping there's a command like

openssl x509

that can read it (asking here first before i start poking at the file with a sharp stick)

new question: is there an easy way to tell, from a PE server's command line, whether code manager is enabled on the machine or not? i.e. is there a specific

puppet config print

setting that changes, or something added or changed in

/etc/puppetlabs/puppet/puppet.conf

, or maybe a file or directory which is created when code manager is enabled and deleted when it's disabled?

Does anyone know if it was intentional that almost all of the tasks shipped under enterprise_tasks in PE are now marked private since 2025.6? This means they don't show up in the console anymore. There's some common convenience tasks in there such as modifying puppet agent config settings (

enterprise_tasks::set_puppet_config_setting

) I can't see anything about it in the release notes / known issues

I was told in here before that the enterprise_tasks were never intended to be used by end-users from the console. For agent config, there is https://forge.puppet.com/modules/puppetlabs/puppet_conf/

Thanks @kenyon ... A mention in the release notes might be a good heads up

there's a lot that might be a good idea that seems to get overlooked

Yes, that change to

enterprise_tasks

was intentional. That module was always intended to be internal-use only and should have been named something like

pe_internal_tasks

. I think a sub-set of "safe" tasks has been left publicly exposed. There are some items in there that are outright dangerous if not run exactly as expected by

puppet-enterprise-installer

or one of the

puppet infra

sub-commands. For example, if you go by title alone,

enterprise_tasks::pg_basebackup

sounds like it would back up the Postgres database in some way. It does not. It nukes the database data directory and then runs the

pg_basebackup

CLI to pull a copy from some other Postgres instance. Does exactly what

puppet infra provision replica

needs it to do, and disastrous if naively run against the Primary.

that definitely falls into the "yikes" category. 😬

Anyone setup Infra Assist yet?

saw some demos, but I'm waiting for the new license to test it

@csharpsteen while I agree on some tasks/plans being internal, maybe

enterprise_tasks::pg_basebackup

isn't the best name? 😄

maybe

enterprise_tasks::pg_basebackup

isn't the best name

Well, yes and no. On the one hand, definitely not a good name from a "people make assumptions based on name" standpoint. On the other hand, the main thing that task does is run a Postgres CLI tool that is literally named

pg_basebackup

. So, the task name 100% makes sense from that standpoint.

I'm struggling to get Infra Assist to work... but don't know if I've done something dumb in Azure or if PE is being dumb

@RyChannel: What seems to be happening with the Infra Assist setup? Is it not connecting to your LLM provider?

Correct, I just keep getting 404s back

But not knowing how I'm suppose to setup Azure OpenAI is problematic too :D

@RyChannel: I'm assuming you followed these steps here: Enabling the Infra Assistant | Puppet Enterprise (PE) | 2025.6

So the Infra Assist is available and enabled, but when you try to ask queries, now you get a 404s back? (in the chat).

No, the 404 comes during the setup when clicking the button to Update API Settings

We just confirmed that our Azure OpenAI resource is reachable from On-Prem

I’d first tail the console-services and infra assistant logs when trying to update the settings through the console.

That should yield some more info

What is the URL you are entering? If you can share a screenshot or the URL. I ran into some issue yesterday but finally got it working with OpenAI (Not Azure one).

Hi folks I have a strange behavior impacting some of my Linux Ubuntu VMs, Any hints:

Could not retrieve catalog from remote server: Error 500 on SERVER: Server Error: Evaluation Error: Error while evaluating a Resource Statement, Evaluation Error: Operator '[]' is not applicable to an Undef Value. (file: /opt/puppetlabs/server/data/puppetserver/filesync/client/versioned-dirs/puppet-code/env_xxx/modules/apt/manifests/setting.pp, line: 60, column: 12) (file: /etc/puppetlabs/puppetserver/code/environments/env_xxx/modules/apt/manifests/pin.pp, line: 130) on node xxx