Encourage users to classify using `simp::classes` rather than `classes`
This log records a significant past decision that hasn't been widely circulated.
Status | DECIDED |
|---|---|
Impact | LOW |
Driver | @Trevor Vaughan |
Approver | @Trevor Vaughan |
Contributors | @Trevor Vaughan @Chris Tessmer @Liz Nemsick |
Informed |
|
Due date | Jul 1, 2019 |
Outcome | We should encourage users to add their own classifications using the |
Background
Relevant data
Options considered
| Option 1: |
|---|---|
Description | We should encourage users to add their own classifications using the |
Pros and cons | The Reusing This avoids needing a lot of extra documentation to enable agentless deployments to either reuse the Combining users' and the |
Estimated cost | LOW |
Action items
Outcome
We should encourage users to add their own classifications using the simp::classes parameter instead of the (now-legacy) classes lookup key.
Comments
@Kendall Moore It's legacy as of today. Everything still works and will continue to work but, per the snippets above, we need to control the SIMP class load path for serverless/Bolt installations.
But yeah, you can still do whatever you want and we're not removing the support from site.pp.
This has a due date of a few weeks ago but I was just notified of it's existence and apparently tagged (I don't see where so the email is confusing). Is the assertion here from the description in Option 1 that it's considered legacy to do node classification via a classes array like in the example in these docs: https://puppet.com/docs/puppet/6.6/hiera_automatic.html? If so I'd like to understand why it's considered legacy as I don't see a deprecation notice anywhere about it (or anywhere else that I could find).