From aebe80e73f6ef198aaac73e6969a6dd39d02f066 Mon Sep 17 00:00:00 2001 From: Ben Vincent Date: Sat, 25 Jul 2026 00:04:11 +1000 Subject: [PATCH] Add profiles::puppet::migrate: repoint VM agents onto puppet-on-k8s (#493) ## Why Migrate the VM puppet agents off the legacy VM puppetmasters onto the new puppet-on-kubernetes servers, one wave at a time and reversible without re-enrolment. Changing `server`/`ca_server` alone is insufficient: each agent's ssldir holds a client certificate signed by the **old** CA, which the new k8s CA neither trusts nor recognises. Migrated nodes switch to a **fresh ssldir**, so the agent generates a new key + CSR on its next run (autosigned immediately by the k8s CA, which runs `autosign = true`), pulls its catalog from the new compilers, and leaves the old CA-signed creds intact on disk for rollback. The reconfiguration catalog is compiled by the **old** server (the node still points there when it applies the change). `profiles::puppet::client` rewrites `puppet.conf` with the new endpoints + ssldir and notifies `Service['puppet']`, so the very next agent run lands on k8s. **Single-boolean design (reworked):** cutover is now driven by one hiera key. `profiles::puppet::migrate` ships wired into `profiles::base` with `enabled => false`, so every node evaluates it but nothing changes until the boolean is flipped. `profiles::puppet::client` includes migrate and folds the migrate endpoints into the effective values it renders. Flip `profiles::puppet::migrate::enabled: true` at any hiera layer (node/role/common) to migrate a target -- no other keys required. ## Changes - `profiles::base` (`site/profiles/manifests/base.pp`): `include profiles::puppet::migrate` alongside the other puppet profiles (ships disabled). - `profiles::puppet::migrate` (`site/profiles/manifests/puppet/migrate.pp`): unchanged params (`enabled=false`, `new_server`, `new_ca_server`, `new_ssldir`); when enabled and the node is not a puppetmaster it owns the fresh ssldir directory. Class-header runbook rewritten for the single-boolean flow. No longer opt-in via `hiera_include`. - `profiles::puppet::client` (`site/profiles/manifests/puppet/client.pp`): `include profiles::puppet::migrate` and compute `$effective_server` / `$effective_ca_server` / `$effective_report_server` / `$effective_ssldir`, feeding those to the template. Keeps the existing optional `$ssldir` / `$report_server` params as manual overrides. - `puppet.conf.erb`: render from the `@effective_*` vars. ## Precedence - **ssldir / report_server** (`Optional`, default `undef`): explicit `profiles::puppet::client::ssldir` / `::report_server` in hiera wins; else the migrate-derived value when `enabled`; else the legacy default (ssldir omitted -> puppet default; report_server -> effective server). - **server / ca_server** (`String`, no undef sentinel): the migrate endpoint wins while `enabled => true`; otherwise the client param (legacy default unless set explicitly). To pin a bespoke server, leave `enabled => false` for that target and set `profiles::puppet::client::server` directly. - No include cycle: migrate.pp never references client.pp; its guard uses `$facts['enc_role']`. ## Runbook - **Canary (one node)** -> `hieradata/nodes/.yaml`: `profiles::puppet::migrate::enabled: true` - **Wave (a role)** -> `hieradata/roles//[/].yaml`: `profiles::puppet::migrate::enabled: true` - **Estate flip** -> `hieradata/common.yaml`: `profiles::puppet::migrate::enabled: true` - **Rollback** -> remove the key (or set `false`) at that layer; next run rewrites puppet.conf back to the legacy server + legacy ssldir (original CA-signed cert intact). - The k8s endpoints (`new_server` / `new_ca_server` / `new_ssldir`) are `profiles::puppet::migrate` param defaults; override them in hiera if the k8s names change. ## Verification - Render comparison: unmigrated node renders a **byte-identical** `puppet.conf` to `develop` (diff empty). Migrated render (only `enabled: true`) emits `server`/`ca_server`/`report_server` = k8s endpoints and `ssldir = /etc/puppetlabs/puppet/ssl-k8s`. - `puppet-lint` (repo args `--no-80chars-check --no-documentation-check --no-puppet_url_without_modules-check --fail-on-warnings`): clean on all 3 manifests. - `puppet parser validate`: OK. ERB syntax: OK. Repo pre-commit hooks (erb-validate, puppet-lint, puppet-validate): all Passed. Reviewed-on: https://git.unkin.net/unkin/puppet-prod/pulls/493 Co-authored-by: Ben Vincent Co-committed-by: Ben Vincent --- site/profiles/manifests/base.pp | 3 + site/profiles/manifests/puppet/client.pp | 67 +++++++++++--- site/profiles/manifests/puppet/migrate.pp | 87 +++++++++++++++++++ .../templates/puppet/client/puppet.conf.erb | 9 +- 4 files changed, 152 insertions(+), 14 deletions(-) create mode 100644 site/profiles/manifests/puppet/migrate.pp diff --git a/site/profiles/manifests/base.pp b/site/profiles/manifests/base.pp index f9d2f80..cc1cab5 100644 --- a/site/profiles/manifests/base.pp +++ b/site/profiles/manifests/base.pp @@ -12,6 +12,9 @@ class profiles::base () { # manage the puppet agent include profiles::puppet::agent include profiles::puppet::client + # k8s migration switch; ships disabled, flip its ::enabled boolean in + # hiera (node/role/common) to repoint a target onto puppet-on-k8s. + include profiles::puppet::migrate # include the base profiles include profiles::base::repos diff --git a/site/profiles/manifests/puppet/client.pp b/site/profiles/manifests/puppet/client.pp index b58c33a..9e1a6a7 100644 --- a/site/profiles/manifests/puppet/client.pp +++ b/site/profiles/manifests/puppet/client.pp @@ -4,19 +4,64 @@ # # site/profile/manifests/puppet/client.pp class profiles::puppet::client ( - Array $dns_alt_names = [$trusted['certname']], - String $server = 'puppetmaster', - String $ca_server = 'puppetca', - String $environment = 'develop', - Integer $runinterval = 1800, - Integer $runtimeout = 3600, - Boolean $show_diff = true, - Boolean $usecacheonfailure = false, - Integer $facts_soft_limit = 4096, - Boolean $splay = true, - Integer $splaylimit = 600, + Array $dns_alt_names = [$trusted['certname']], + String $server = 'puppetmaster', + String $ca_server = 'puppetca', + Optional[String] $report_server = undef, + String $environment = 'develop', + Integer $runinterval = 1800, + Integer $runtimeout = 3600, + Boolean $show_diff = true, + Boolean $usecacheonfailure = false, + Integer $facts_soft_limit = 4096, + Boolean $splay = true, + Integer $splaylimit = 600, + Optional[Stdlib::Absolutepath] $ssldir = undef, ) { + # Pull in the k8s-migration switch. When + # profiles::puppet::migrate::enabled is true at ANY hiera layer + # (node/role/common) the effective endpoints below flip to the k8s + # servers and a fresh ssldir -- no other hiera keys required. This class + # never includes client.pp, so there is no include cycle. + include profiles::puppet::migrate + + # Effective values fed to the template. Precedence differs by param type: + # + # ssldir / report_server (Optional, default undef): an explicit client + # param wins; else the migrate-derived value when enabled; else the + # legacy default (undef -> template omits ssldir; report_server -> server). + # + # server / ca_server (String, no undef sentinel): the migrate value wins + # when migrate::enabled, otherwise the client param (which is the legacy + # default unless an operator set it explicitly in hiera). To pin a + # bespoke server while enabled, leave enabled => false for that target + # and set profiles::puppet::client::server directly. + $migrate_on = $profiles::puppet::migrate::enabled + + $effective_server = $migrate_on ? { + true => $profiles::puppet::migrate::new_server, + default => $server, + } + $effective_ca_server = $migrate_on ? { + true => $profiles::puppet::migrate::new_ca_server, + default => $ca_server, + } + # report_server: explicit param wins; else follow the effective server. + $effective_report_server = $report_server ? { + undef => $effective_server, + default => $report_server, + } + # ssldir: explicit param wins; else the fresh k8s ssldir when migrating; + # else undef (template omits the key, i.e. puppet's built-in default). + $effective_ssldir = $ssldir ? { + undef => $migrate_on ? { + true => $profiles::puppet::migrate::new_ssldir, + default => undef, + }, + default => $ssldir, + } + # dont manage puppet.conf if this is a puppetmaster if $facts['enc_role'] != 'roles::infra::puppet::master' { diff --git a/site/profiles/manifests/puppet/migrate.pp b/site/profiles/manifests/puppet/migrate.pp new file mode 100644 index 0000000..5bf287e --- /dev/null +++ b/site/profiles/manifests/puppet/migrate.pp @@ -0,0 +1,87 @@ +# Class: profiles::puppet::migrate +# +# Repoints a VM puppet agent from the legacy VM puppetmasters onto the new +# puppet-on-kubernetes servers. +# +# The agent's existing ssldir holds a client certificate signed by the OLD CA; +# the new k8s CA will neither trust that cert nor recognise the agent. Rather +# than revoke/clean the old creds (which would burn the rollback path), this +# class switches the agent to a FRESH ssldir. On the next run the agent finds +# an empty ssldir, generates a new key + CSR, submits it to the new CA +# (autosign = true on the k8s side signs it immediately), and then pulls its +# catalog from the new compilers. The OLD ssldir is left untouched on disk, so +# reverting the hiera below restores the node to the legacy masters with no +# re-enrolment. +# +# The catalog that performs the switch is compiled by the OLD server (the node +# still points at it when it applies this change). profiles::puppet::client +# rewrites puppet.conf with the new server / ca_server / report_server / ssldir +# and notifies Service['puppet'], so the very next agent run lands on k8s. +# +# NOTE: this is deliberately hiera-driven per node/role so waves can be rolled +# out and rolled back one target at a time. The class only manages the fresh +# ssldir; the actual puppet.conf changes ride on profiles::puppet::client's +# existing template. client.pp `include`s this class and, when enabled, folds +# new_server / new_ca_server / new_ssldir into the effective server / +# ca_server / report_server / ssldir it renders. Do NOT add a second File +# resource for puppet.conf here -- client.pp already owns it. +# +# == Wiring (single boolean) +# +# This class ships wired into profiles::base with enabled => false, so every +# node evaluates it but nothing changes until the boolean is flipped. Because +# client.pp includes it and reads its params, setting the ONE key below at any +# hiera layer repoints a target -- no client.pp param overrides needed. +# +# The endpoints (new_server / new_ca_server / new_ssldir) are param defaults +# on this class; override them in hiera too if the k8s names ever change. +# +# --- Canary: one node -> hieradata/nodes/.yaml ------------------- +# profiles::puppet::migrate::enabled: true +# --------------------------------------------------------------------------- +# +# --- Wave: a whole role -> hieradata/roles//[/].yaml ----------- +# profiles::puppet::migrate::enabled: true +# --------------------------------------------------------------------------- +# +# --- Estate flip -> hieradata/common.yaml (retires the legacy masters) ----- +# profiles::puppet::migrate::enabled: true +# --------------------------------------------------------------------------- +# +# Manual override: profiles::puppet::client::ssldir and ::report_server are +# Optional and, when set explicitly in hiera, win over the migrate-derived +# value. server / ca_server have no undef sentinel, so while enabled => true +# the migrate endpoints win; to point a target at a bespoke server leave +# enabled => false for it and set profiles::puppet::client::server directly. +# +# == Rollback (per target) +# +# Remove the profiles::puppet::migrate::enabled key (or set it back to false) +# at the node/role/common layer. The next run rewrites puppet.conf back to the +# legacy server + legacy ssldir, which still contains the original CA-signed +# cert, and the node is back on the VM masters. +# +# site/profiles/manifests/puppet/migrate.pp +class profiles::puppet::migrate ( + Boolean $enabled = false, + String $new_server = 'puppet.k8s.syd1.au.unkin.net', + String $new_ca_server = 'puppetca.k8s.syd1.au.unkin.net', + Stdlib::Absolutepath $new_ssldir = '/etc/puppetlabs/puppet/ssl-k8s', +) { + + # never touch the puppet masters themselves (same guard as client.pp) + if $enabled and $facts['enc_role'] != 'roles::infra::puppet::master' { + + # Ensure the fresh ssldir exists with agent-correct ownership/perms. The + # puppet agent will create it on demand, but pre-creating it keeps the very + # first post-switch run from racing directory creation and makes the + # migration state auditable. Contents (keys, certs, crl) are populated by + # the agent against the new CA -- we only own the directory itself. + file { $new_ssldir: + ensure => 'directory', + owner => 'root', + group => 'root', + mode => '0771', + } + } +} diff --git a/site/profiles/templates/puppet/client/puppet.conf.erb b/site/profiles/templates/puppet/client/puppet.conf.erb index d9bc058..b3d31d5 100644 --- a/site/profiles/templates/puppet/client/puppet.conf.erb +++ b/site/profiles/templates/puppet/client/puppet.conf.erb @@ -1,12 +1,15 @@ [main] dns_alt_names = <%= @dns_alt_names_string %> +<% unless @effective_ssldir.nil? -%> +ssldir = <%= @effective_ssldir %> +<% end -%> [agent] -server = <%= @server %> -ca_server = <%= @ca_server %> +server = <%= @effective_server %> +ca_server = <%= @effective_ca_server %> environment = <%= @environment %> report = true -report_server = <%= @server %> +report_server = <%= @effective_report_server.nil? ? @effective_server : @effective_report_server %> runinterval = <%= @runinterval %> runtimeout = <%= @runtimeout %> show_diff = <%= @show_diff %>