Differences

This shows you the differences between two versions of the page.

--- devops:puppet [2016/04/18 09:16] – [Classes Metaparameter] skipidar
+++ devops:puppet [2023/11/01 07:15] (current) – ↷ Page moved from camunda:devops:puppet to devops:puppet skipidar
@@ Line 1: / Line 1: @@
 ===== Puppet =====
+The Learning VM TUtorial is here: https://kjhenner.gitbooks.io/puppet-quest-guide/content/quests/application_orchestrator.html
+Te learning VM is available here: https://puppet.com/download-learning-vm
 ==== Glossary ====
@@ Line 36: / Line 41: @@
   * describe entire system role as "database-server", "web application worker"
   * some abstract aspect as "cloud-tools", which is composed of many tools
+To **define** a class - means to describe it in a module, so that it can be used on nodes. \\
+To **declare** a class - means to apply the class to a node. So it will execute there and do its modifications, defined in the .pp file.
+<fc #FF0000>Achtung: you only can apply a clas once to a node. So if you wish to use the same peace of code with ifferent parameters (e.g. create multiple users, folders...) use defined resource types</fc>
 </WRAP>|
+| defined resource types | like classes, but appliable to nodes multiple times  |
 |node definition | The task of configuring which classes will be applied to a given node - is called **node classification**. \\ Node definitions are a puppet concept to write node classification down. |
 |.pp manifests|<WRAP>
@@ Line 50: / Line 62: @@
 |Puppet master | The puppet server which configures the puppet agents|
 |Puppet agent| Installed on machines, which should be controlled by puppet (by puppet master)|
@@ Line 172: / Line 183: @@
 </code>
+==== Defined Resource types ====
+  * Use **define** keyword instead of **class**
+  * **$title** is a special variable, which must be unique on a node and is mandatory
+  * Binding of parameter variables ($content, $password) to values happens in parallel, meaning that you cannot use the value of one parameter to set another. The exception is the $title variable.
+# /etc/puppetlabs/code/environments/production/modules/web_user/manifests/user.pp
+<code>
+define web_user::user (
+  $content  = "<h1>Welcome to ${title}'s home page!</h1>",
+  $password = undef,
+) {
+  $home_dir    = "/home/${title}"
+  $public_html = "${home_dir}/public_html"
+  user { $title:
+    ensure   => present,
+    password => $password,
+  }
+  file { [$home_dir, $public_html]:
+    ensure => directory,
+    owner  => $title,
+    group  => $title,
+    mode    => '0755',
+  }
+  file { "${public_html}/index.html":
+    ensure  => file,
+    owner   => $title,
+    group   => $title,
+    replace => false,
+    content => $content,
+    mode    => '0644',
+  }
+}
+</code>
+**Declaration** of users on a node. \\
+'shelob', 'frodo' are the titles of the users, which may be references via the **$title** variable
+# /etc/puppetlabs/code/environments/production/modules/web_user/examples/user.pp
+<code>
+web_user::user { 'shelob': }
+web_user::user { 'frodo':
+  content  => 'Custom Content!',
+  password => pw_hash('sting', 'SHA-512', 'mysalt'),
+}
+</code>
 ==== Classes Metaparameter ====
@@ Line 265: / Line 324: @@
 defines the default node configuration.
+==== Node setup ====
+Puppet master provides a bash script for setting up nodes:
+<code>
+curl -k https://<master.example.com>:8140/packages/current/install.bash | sudo bash
+</code>
+Puppet MASTER keeps signed certificates of each Node, which is a part of the infrastructure.
+To involve a node - sighn its certificate.
+List all unsigned certificates of Nodes. Executable on master.
+<code>
+puppet cert list
+</code>
+Sign a certificate of node named **webserver.learning.puppetlabs.vm **
+<code>
+puppet cert sign webserver.learning.puppetlabs.vm
+</code>
 ==== Variables ====
@@ Line 595: / Line 673: @@
 yamldir = /opt/puppetlabs/puppet/cache/yaml
 </code>
+===== Orchestration =====
+Details are here: https://kjhenner.gitbooks.io/puppet-quest-guide/content/quests/application_orchestrator.html
+Orchestration allows to install dependent applications in the right order.
+It works as following:
+  - setup the node to know the orchestrator
+  - create an Application, which may combine multiple nodes to a single semantical unit (MySQL Server, Apache with PHP application )
+    - setup orchestrator user with the rights
+    - setup an **artificial, public resource** to share the data between nodes (**in ruby**)
+    - define, which node **produces** data to fill artificial, public resource. Which node consumes data from artificial, public resource.
+    - declare the resource
+    - in site.pp declare - which part of the application will be installed on which node.
+{{http://i520.photobucket.com/albums/w327/schajtan/2016-04-25_14-27-49_zpsphhbuy9w.png}}