1. Foreman 1.2 Manual

Foreman Architecture

A Foreman installation will always contain a central foreman instance that is responsible for providing the Web based GUI, node configurations, initial host configuration files, etc. However, if the foreman installation supports unattended installations then other operations need to be performed to fully automate this process. The smart proxy manages remote services and is generally installed with all Foreman installations to allow for TFTP, DHCP, DNS, and Puppet, and the Puppet CA.

Smart-Proxy

A Smart-Proxy is located on or near a machine that performs a specific function and helps foreman orchestrate the process of commissioning a new host. Placing the proxy on or near to the actual service will also help reduce latencies in large distributed organizations.

Foreman Architecture

Release notes for 1.2

Headline features

Discovery plugin support

foreman_discovery is supported with Foreman 1.2 and provides Metal-as-a-Service hardware discovery. New hosts boot into a discovery image which registers with Foreman, then can be converted and built from the web interface. See the README file for a step by step guide to configuring Foreman and the plugin.

BMC and IPMI support

Power control and boot device controls are now available from the web interface for bare metal hosts with a BMC network interface configured. The BMC interface can be added under the host network tab, and a smart proxy with the BMC feature configured is also required to perform actions over IPMI.

Internationalization

The web interface has been internationalized and now supports German, Spanish and French. The interface will switch to the language specified by the browser by default and can be overridden from the user settings within Foreman. Further translations are welcomed through the Foreman project on Transifex.

SPICE HTML5 support

Virtual machine consoles on oVirt hosts can now be viewed from any HTML5-capable browser, instead of requiring the Linux-only XPI plugin. The XPI view is still available via the “New Window” link.

SELinux targeted support

An SELinux module for the standard targeted policy is now shipped in the foreman-selinux subpackage, supporting all standard configurations and options of Foreman. Please file any AVCs caused by Foreman or common plugins as bugs.

PostgreSQL by default

The Foreman installer default database has changed from SQLite to PostgreSQL, providing increased performance and reducing concurrency errors seen even in very small environments. MySQL is also fully supported and can be selected as an alternative. Database management in the installer modules can also be disabled, see this announcement for details.

Puppet runs via MCollective

The smart proxy can now initiate Puppet runs via the MCollective “puppet” agent, when configured with :puppet_provider: mcollective (documentation).

API version 2

A new “experimental” API version (v2) has been added, providing new features while retaining API v1 compatibility. API v1 still remains the default version. To use API v2, users must pass “version=2” in the header. For example:

curl -u admin:secret -H ‘Accept:application/json,version=2’ http://example.com/api/locations

New API v2 features include adding and removing puppet classes to hosts and host groups, CRUD for locations and organizations, and managing regular parameters for hosts, host groups, domains and operating systems (not smart parameters).

This also marks the official deprecation of the pre-1.1 API (non /api paths), see notes below.

Security fixes

  • Fix remote code execution in Foreman via bookmarks storage (CVE-2013-2121)
  • Fix role privilege escalation by users with create/edit user privileges (CVE-2013-2113)
  • Fix arbitrary remote execution risk in Puppet run smart proxy API (CVE-2013-0210)

Upgrade notes

Changes to production deployment for non-package users

For users running Foreman from git instead of RPMs or Debian packages, two additional deployment steps are now required for production usage:

  • rake locale:pack to compile locale .mo files
  • rake assets:precompile to compile images, JavaScript etc.

Software collections in RPMs

The RPMs now provide a full Ruby 1.9 stack via software collections (SCL), complete with the version of Ruby on Rails and other dependencies used by Foreman. This allows for a single stack of packages across each supported RPM-based distro, plus easier support of future dependencies as Ruby 1.8 goes EOL upstream. The software collection is installed under /opt/rh/ruby193.

This changes any rake or ruby commands used for managing Foreman, which now become ruby193-rake and ruby193-ruby respectively. Commands can also be run with scl enable ruby193 'original command' to run in the context of the collection (note the quotes).

More about software collections is available in the Software Collections manual section. Please note the upgrade instructions below, particularly for Passenger configuration.

Host group matcher inheritance

Matchers used in smart variables or class parameters to match host groups are now inherited by children of those matching host groups too (e.g. a matcher for hostgroup=Base will also apply to Base/Web). This behaviour can be reverted by the host_group_matchers_inheritance setting (More > Settings > Puppet).

ActiveRecord-based storeconfigs / database sharing now unsupported

Previously it was suggested that a single database could be shared between Puppet ActiveRecord-based storeconfigs and Foreman, which would be used to import data from Puppet - this configuration was deprecated in Foreman 1.1 and is now unsupported in 1.2. Foreman must be configured with a standalone database, then configure either an ENC or fact and report uploads, which provides most of the same functionality as database sharing without the associated issues.

Puppet 3 also deprecated the use of ActiveRecord-based storeconfigs and it is now recommended to use PuppetDB instead for storeconfigs and exported resources. This can be used alongside Foreman as the two do not need to interact.

Pre-1.1 API (non-/api) is deprecated

The older API is now deprecated and users should migrate to the versioned API at /api, available since Foreman 1.1. If you use paths such as /hosts rather than /api/hosts then you are using the pre-1.1 API and should change your program to use the new one.

Error after upgrade: ArgumentError (dump format error (user class))

If you see the following error in production.log, clear cookies in your browser:

ArgumentError (dump format error (user class)):
  /usr/lib/ruby/gems/1.8/gems/passenger-4.0.5/lib/phusion_passenger/rack/thread_handler_extension.rb:77:in `process_request'

VMware compute resource public keys need updating

When the Ruby version changes during an upgrade, an issue will be hit with VMware compute resources as the hash of the public key used to identify vSphere will change. This will happen when upgrading an RPM installation, since it switches from Ruby 1.8 to 1.9.

To resolve, an admin user should browse to the compute resource in Foreman and click “Test Connection”, which will update the public key on record.

Package upgrade instructions

Please see the end of the package sections for each platform for specific upgrade instructions:

If you have problems with Passenger after upgrading the RPMs, please see the SCL section for information about the expected configuration.

Release notes for 1.2.3

Foreman 1.2.3 is a security and bug fix release for the 1.2 series. Only the core Foreman component has been updated - the installer, smart proxy and SELinux packages remain at version 1.2.1.

Infrastructure

  • Update ancestry version to 2.0 due to Fedora 19 repo incompatibility (#3142)
  • Pin gettext_i18n_rails version for Ruby 1.8 support (#3167)

Security

  • Fix SQL injection via host/host group parameter override associations (CVE-2013-4386, #3160)

A full list of changes in 1.2.3 is available via the Roadmap

Release notes for 1.2.2

Foreman 1.2.2 is a security and bug fix release for the 1.2 series. Only the core Foreman component has been updated - the installer, smart proxy and SELinux packages remain at version 1.2.1.

Infrastructure

  • Pin gettext version for Ruby 1.8 support (#3005)

Security

  • Fix potential DoS via hosts controller due to symbol conversion (CVE-2013-4180, #2860)
  • Fix privilege escalation as hosts API wasn’t restricted to user’s permitted hosts (CVE-2013-4182, #2863)

UI

  • Error fixed on organization mismatches report when locations are disabled (#3010)

A full list of changes in 1.2.2 is available via the Roadmap

Release notes for 1.2.1

Foreman 1.2.1 is a bug fix release for the 1.2 series. It also adds Fedora 19 packages.

API

  • Fix accepted parameters to partition table API to include layout and OS family (#2791)
  • Fix rundecks API when using ‘rundeckfacts’ parameter (#1770)

Authentication and authorization

  • Fix permissions problem for non-admins changing puppet classes on hosts or host groups (#2779)

Compute resources

  • Retry SSH authentication failures due to delayed cloud-init(#2839)

Host management

  • Fix location/organization scope issue on host and host group form (#2734)

Infrastructure

  • Pin nokogiri version for Ruby 1.8 support (#2808)

Installer

  • modules/ subdirectory now created beneath each static environment (#2770)

Packaging

  • Fedora 19 packages built
  • ruby193-rubygem-rubyipmi and rubygem-rubyipmi RPMs available for proxy BMC support
  • rubygem-ldap updated to resolve protocol errors (#2737)
  • Default root password not stored in foreman-debug output (#2872)

Provisioning templates

  • Add repos hook for repositories associated with a host and OS (#2837)

Puppet integration

  • Fix handle_ca error when no Puppet CA is assigned to a host (#2693)
  • Prevent Puppet classes being assigned to a host before pressing “Submit” (#2788)

Smart Proxy

  • Dynamic module path parsing error fixed (#2738)

UI

  • Fix host group edit page failure when unattended mode was disabled (#2373)
  • Fix 404 error on mismatches report if a location or organization is selected (#2732)
  • Fix failure when autocompleting right hand side of search expression (#2747)
  • Fix cloning of locations and organizations (#2772)
  • Fix LDAP password being overwritten as blank (#2782)
  • Fix diffs being saved in the provisioning template editor (#2797)
  • Fix search errors on PostgreSQL due to order clause (#2801)
  • Fix Puppet class parameters refresh on host form on environment change (#2851)

A full list of changes in 1.2.1 is available via the Roadmap

Release notes for 1.2.0

API

  • Experimental APIv2 added, providing:
    • Puppet class management nested on hosts and host groups (#2250)
    • Locations and organizations CRUD API (#2239)
    • Parameters on hosts, domains, host groups and operating systems can now be managed via nested calls, e.g. GET /api/domains/6/parameters (#1988)
    • Template combinations (host groups and environments) management, nested under /config_templates
    • puppetrun API added (#2539)
    • Resource list/show methods can now specify location_id and organization_id to filter (#2239)
    • Full hash parameters (rather than using _ids) now supported in config templates for operating systems
    • Separate APIv2 documentation linked via “v2” in top-right of docs web page
  • All API routes now follow existing delegated roles and permissions (#2248, #2266)
  • Usergroups can be searched and ordered when listed
  • Permissions error when managing environments fixed (#2273)
  • Ensure documentation is loaded to show links on /api
  • No longer returns all classes for non-existent host (#2550)
  • Documentation for host group parent_id fixed (#2728)

Authentication and authorization

  • All pages and actions now attached to permissions and can be delegated, including “Run Puppet” (#2202, #2221, #985)
  • Host groups ownership is hierarchical, users now have access to child host groups (#2219)
  • trusted_puppetmaster_hosts setting added to permit specified hosts to ENC/reports/facts (#2153)
  • Signo support added for SSO with Katello, using OpenID protocol (#2417)
  • Users can no longer select “Any Organization” and view hosts outside of their organizations (#2298)
  • Admin flag can no longer be removed from mandatory builtin admin user

Compute resources

  • Amazon EC2: IAM roles can be specified on images (#2229)
  • Amazon EC2: fix empty security group error (#2490)
  • libvirt: support for NAT named networks (#1951)
  • libvirt: VNC listening address error fixed, default set with libvirt_default_console_address setting (#2562)
  • libvirt: fix NIC “id” error under Ruby 1.9 (#2002)
  • OpenStack: available operating system images shown (#2251)
  • OpenStack: text console log can be viewed for hosts (#2252)
  • OpenStack: floating/public IP address can now been selected (#2253)
  • oVirt/RHEV: selecting templates fixed (#2224)
  • oVirt/RHEV: fix host installation due to disk locking (#2163)
  • oVirt/RHEV: fix error when adding disk to existing VM (#2316)
  • oVirt/RHEV: boot order now set to network, disk on new VMs (#2610)
  • oVirt/RHEV: creation of VMs with foreman_####### names fixed (#2317)
  • SPICE console now uses HTML5 instead of requiring XPI extension, which is still accessible via “New Window” button (#2277)
  • VMware: network label now set correctly on NICs (#2188)
  • VMware: public key hashes now updated via Test Connection after upgrading Ruby (#2656)
  • Image based CRs: A record is now created in domain (#2325)
  • Fix test connection button for existing compute resources
  • Connections are now tested with improved error messages when creating hosts (#2606)
  • SSH disconnects when connecting to new VMs is now retried (#2712)

Host management

  • BMC tab for power control added to hosts with a BMC NIC configured, requires a BMC-enabled smart proxy (#426)
  • Environment can be updated to the host group selection via the multi-select change environment window (#686)
  • Add Gentoo operating system, add SLES and SLED to SUSE family
  • Kickstart and preseed templates have been updated from community-templates repo (#2038)
  • Installation media URLs have been updated (#2695)
  • Domain no longer appended to FQDN on host update when FQDN domain differs (#2130)
  • Fix OS selection when scoped by both organization and location (#2222)
  • Behaviour of new host form fixed when “Any Location/Organization” selected (#2197)
  • Free IP address correctly retrieved from smart proxy on Ruby 1.9 (#2398)
  • Non-digit characters stripped from OS versions, unusual OS versions can now ben imported (#2372)
  • Issues updating BMC and host root passwords resolved (#2484, #2489)
  • Host group provisioning falls back to root_pass setting (#2538)
  • Windows newlines now stripped during host group template rendering (#2527)

Infrastructure

  • Ruby on Rails updated to 3.2
  • Ruby 2.0 support, requires Puppet 3.2.0 or higher (#2367)
  • audited 3.0.0 supported (#2393)
  • foreman-debug script added to collect data for support purposes (#2434)
  • Load libvirt group when bundler_ext used (#2428)
  • ruby_parser warnings generated in rake cronjobs have been silenced (#2217)
  • Plugins can now load settings from config/settings.plugins.d/*.yaml (#2389)
  • Puppet can be loaded from a gem by setting :puppetgem: true in settings.yml (#2268)
  • Show error in rake tasks when Facter domain is nil (#2408)
  • Add –dry-run option to foreman-config
  • Restrictive column lengths for installation media and others extended (#2587)
  • Permission failure when overriding class parameters fixed (#2386)
  • Settings in settings.yaml now override UI defaults without errors (#2343)
  • User given names can now contain i18n characters (#2669)
  • Hang while validating installation media names fixed (#2685)

Installer

  • PostgreSQL and MySQL can be configured, PostgreSQL is now the default database (#2013, #2014)
  • Oracle Linux is now supported (#2342)
  • Answers file correctly written before running Puppet (#2442)
  • Apache HTTP virtual host IP configured correctly on Debian and Ubuntu (#2346)
  • Concat issue on Ruby 1.9 fixed (#2447)

Packaging

  • RPMs now use Software Collections (SCL) and supply a complete Ruby 1.9.3 stack
  • SELinux targeted policy is now provided (#2125)
  • RPMs now use bundler_ext, so Gemfile.lock no longer requires updating after package update (#2204)
  • Fedora 18 packages built
  • RPMs are now GPG signed by 0xE775FF07 (#2106)
  • RPMs: foreman-ec2 has been renamed to foreman-compute for consistency (#2123)
  • Secret token rake task can now run from any directory, to prevent zero byte files (#2633)

Provisioning templates

  • “regexp buffer overflow” when rendering templates on Ruby 1.8 fixed (#2100)
  • Templates no longer deleted when associated environment is deleted (#2246)
  • Duplicate epel and HTTP Proxy snippets after install fixed (#2358)
  • Refreshed EPEL release RPM version number (#2359)

Puppet integration

  • Parameters can be overriden via the host group UI (#2132)
  • Matchers for host groups are now inherited by child host groups, controlled by host_group_matchers_inheritance setting (#2220)
  • Puppet master and CA proxies are now optional when creating hosts (#2461)
  • Facts importer no longer recreates unchanged values in database (#2486)
  • FQDN is now stored lower cased when imported through facts, preventing UI issues (#2351)
  • SQL error when importing modules (e.g. Example42) fixed (#2485)
  • FQDN fact no longer used to identify nodes, certname used instead (#1938)
  • Location and organization matchers now match correctly (#2334)
  • Non-admin users can create and manage smart variables with the external_variables permissions (#2645)

Security

  • Fix remote code execution in Foreman via bookmarks storage (CVE-2013-2121, #2631)
  • Fix role privilege escalation by users with create/edit user privileges (CVE-2013-2113, #2630)

Smart Proxy

  • MCollective can be used to trigger Puppet runs using the “puppet” agent (#2116, documentation)
  • DNS GSS-TSIG support added for Windows AD DNS and FreeIPA (#1685, documentation)
  • Shell provider added to BMC API, runs shutdown commands locally (#2387)
  • Fix arbitrary remote execution risk in Puppet run API (CVE-2013-0210)
  • Fix Sinatra 1.4.2 incompatibility, causing HTML to be returned instead of JSON
  • Facts API now refreshes data on each call (#2255)
  • Environments list fixed when ActiveRecord storeconfigs is enabled (#2209)
  • Proxy not started at boot if :daemon set to false (#2152)
  • Puppet class parameters with an “undef” default no longer marked as required (#2191)
  • Puppet manifests with import statements are no longer ignored (#2067)
  • Path to puppetca in Puppet Enterprise fixed
  • “master” no longer detected as an environment (#2654)
  • Sinatra 1.4.3 excluded due to “Accept” header handling bug (#2673)

UI

  • New top bar UI design
  • User interface has been internationalized (#2269, #2368, #2420, #2426)
  • Users can select their preferred language (#2444)
  • Plugins can add tabs and new settings to the Settings UI (#2407)
  • Latest reports and metrics now shown on dashboard (#2406)
  • Replace Highcharts with Flot
  • Footer moved to separate About page, showing proxy and compute resource statuses (#2511)
  • Hosts can be searched by comment field (#2475)
  • Hosts and host groups can be searched by the full host group name / label (#2263)
  • Subnet edit page allows editing of location/organization (#2203)
  • Gravatar can be disabled with use_gravatar setting (#2247)
  • Enable host searches on OS minor and major version numbers (#2499)
  • Absolute URLs fixed in Puppet class import (#2416)
  • Escaped HTML in help popups fixed (#2396)
  • Fix display of multiple form errors under Ruby 1.9 (#2242)
  • Favicon now shown when Foreman hosted at non-root context (#2526)
  • Link to host added on some audit entries
  • Filter and sorting added to multi-select components
  • Organizations created via API now shown in UI menu immediately (#2513)
  • Audit history listing fixes (#2537, #2546)
  • After updating a Puppet class, it redirects to the referring page (#2335)
  • Progress indicators added across the host form UI (#2582)
  • Search errors for domains or compute resources on PostgreSQL fixed (#2602)
  • Searching for hosts with digits no longer matches dates and timestamps (#2604)
  • Subnet addresses are now validated when saving, preventing later errors (#2285)
  • Password fields now have autocomplete disabled to prevent browser interference (#2607)
  • Error messages from smart proxies now HTML escaped (#2622)
  • Host submit progress bar now present when hosted at non-root URL (#2636)
  • Prevent 404 error when renaming an object with an invalid name (#2404)

A full list of changes in 1.2.0 is available via the Roadmap

2. Quickstart

The Foreman installer is a collection of Puppet modules that installs everything required for a full working Foreman setup. It uses native OS packaging (e.g. RPM and .deb packages) and adds necessary configuration for the complete installation.

Components include the Foreman web UI, Smart Proxy, Passenger (for the puppet master and Foreman itself), and optionally TFTP, DNS and DHCP servers. It is configurable and the Puppet modules can be read or run in “no-op” mode to see what changes it will make.

Supported platforms

  • RHEL derivatives (CentOS), version 6 (EPEL required, plus subscribe to “optional” channel on RHEL 6)
  • Fedora 18, 19
  • Debian 7 (Wheezy)
  • Debian 6 (Squeeze) (update Puppet from backports)
  • Ubuntu 12.04 (Precise)

Other operating systems will need to use alternative installation methods (see the manual).

2.1 Installation

The Foreman installer uses Puppet to install Foreman. This guide assumes that you’ve already installed Puppet (without a master), but the installer will setup Foreman and the puppet master with Passenger and the Smart Proxy by default.

Downloading the installer

For Red Hat variants, run this (replace ‘el6’ with ‘f18’ or ‘f19’ as appropriate):

yum -y install http://yum.theforeman.org/releases/1.2/el6/x86_64/foreman-release.rpm
yum -y install foreman-installer

For Debian variants, run this (replace ‘wheezy’ with ‘precise’ if on Ubuntu 12.04, or ‘squeeze’ for Debian 6):

echo "deb http://deb.theforeman.org/ wheezy stable" > /etc/apt/sources.list.d/foreman.list
wget -q http://deb.theforeman.org/pubkey.gpg -O- | apt-key add -
apt-get update && apt-get install foreman-installer

Running the installer

The procedure is initially interactive, allowing changes to the configuration and then the option is given to run Puppet to perform the installation:

ruby /usr/share/foreman-installer/generate_answers.rb

After Puppet runs, Foreman will be accessible at https://fqdn/ with a default username/password of “admin” and “changeme”. To complete the installation, click on More > Configuration > Smart Proxies and add a new proxy with the URL https://fqdn:8443/.

2.2 Puppet Management

After installation, the Foreman installer will have set up a puppet master on the host, fully integrated with Foreman. First run the Puppet agent on the Foreman host which will send the first Puppet report to Foreman, automatically creating the host in Foreman’s database.

puppet agent --test
Puppet 3+ will show a warning the first time that the node can't be found, this can be ignored.

In Foreman, click on the Hosts tab and your Foreman host should be visible in the list with an “O” status. This indicates its status is OK, with no changes made on the last Puppet run.

Downloading a Puppet module

Next, we’ll install a Puppet module for managing the NTP service. If you have Puppet 2.7.14 or higher, install the module automatically from Puppet Forge to our “production” environment (the default):

puppet module install -i /etc/puppet/environments/production/modules/ saz/ntp

On older versions, download the tar.gz and unpack to /etc/puppet/environments/production/modules/. Rename the directory to “ntp”, removing the author and version number.

In Foreman, go to More > Configuration > Puppet Classes and click Import from hostname (top right) to read the available Puppet classes from the puppet master and populate Foreman’s database. The “ntp” class will appear in the Puppet class list if installed correctly.

Using the Puppet module

Click on the “ntp” class in the list, change to the Smart Class Parameters tab and select the server_list parameter on the left hand side. Tick the Override checkbox so Foreman manages the “server_list” parameter of the class and change the default value if desired, before submitting the page.

Change back to the Hosts tab and click Edit on the Foreman host. On the Puppet Classes tab, expand the ntp module and click the + icon to add the ntp class to the host, then save the host.

Managed parameters can be overridden when editing an individual host from its Parameters tab.

Clicking the YAML button when back on the host page will show the ntp class and the server_list parameter, as passed to Puppet via the ENC (external node classifier) interface. Re-run puppet agent --test on the Foreman host to see the NTP service automatically reconfigured by Puppet and the NTP module.

Adding more Puppet-managed hosts

Other hosts with Puppet agents installed can use this puppet master by setting server = foreman.example.com in puppet.conf. Sign their certificates in Foreman by going to More > Configuration > Smart Proxies > Certificates or using puppet cert list and puppet cert sign on the puppet master.

Puppet classes can be added to host groups in Foreman instead of individual hosts, enabling a standard configuration of many hosts simultaneously. Host groups are typically used to represent server roles.

3. Installing Foreman

There are several different methods of installing Foreman. The recommended way is with the puppet based Foreman Installer but you may also use your distribution’s package manager or install directly from source.

3.1 Supported Platforms

The following operating systems are supported by the installer, have packages and are tested for deploying Foreman:

  • RHEL derivatives (CentOS) version 6
    • EPEL is required
    • On RHEL 6, subscribe to the “optional” child channel
  • Fedora 18, 19
  • Debian 7 (Wheeze)
  • Debian 6 (Squeeze)
    • Ensure Puppet is updated from backports
  • Ubuntu 12.04 (Precise)

All platforms will require Puppet 2.6.5 or higher. Puppet 3.x is supported and may be installed from the Puppet Labs repositories.

Other operating systems will need to use alternative installation methods, such as from source.

The following operating systems are known to install successfully from Foreman:

  • RHEL derivatives (CentOS) 3+
  • Fedora
  • Ubuntu
  • Debian
  • Solaris 8, 10
  • OpenSUSE 11.4

3.2 Foreman Installer

The Foreman installer is a collection of Puppet modules that installs everything required for a full working Foreman setup. It uses native OS packaging (e.g. RPM and .deb packages) and adds necessary configuration for the complete installation.

Components include the Foreman web UI, Smart Proxy, Passenger (for the puppet master and Foreman itself), and optionally TFTP, DNS and DHCP servers. It is configurable and the Puppet modules can be read or run in “no-op” mode to see what changes it will make.

It’s strongly recommended to use the installer instead of only installing packages, as the installer uses OS packages and it saves a lot of time otherwise spent replicating configuration by hand.

By default it will configure:

  • Apache HTTP with SSL (using a Puppet-signed certificate)
  • Foreman running under mod_passenger
  • Smart Proxy configured for Puppet, TFTP and SSL
  • Puppet master running under mod_passenger
  • Puppet agent configured
  • TFTP server (under xinetd on Red Hat platforms)

Other modules can be enabled, which will also configure:

  • ISC DHCP server
  • BIND DNS server

3.2.1 Installation

The Foreman installer uses Puppet to install Foreman. This guide assumes that you’ve already installed Puppet (without a master), but the installer will setup Foreman and the puppet master with Passenger and the Smart Proxy by default.

Downloading the installer

For Red Hat variants, run this (replace ‘el6’ with ‘f18’ or ‘f19’ as appropriate):

yum -y install http://yum.theforeman.org/releases/1.2/el6/x86_64/foreman-release.rpm
yum -y install foreman-installer

For Debian variants, run this (replace ‘wheezy’ with ‘precise’ if on Ubuntu 12.04, or ‘squeeze’ for Debian 6):

echo "deb http://deb.theforeman.org/ wheezy stable" > /etc/apt/sources.list.d/foreman.list
wget -q http://deb.theforeman.org/pubkey.gpg -O- | apt-key add -
apt-get update && apt-get install foreman-installer

Running the installer

The procedure is initially interactive, allowing changes to the configuration and then the option is given to run Puppet to perform the installation:

ruby /usr/share/foreman-installer/generate_answers.rb

After Puppet runs, Foreman will be accessible at https://fqdn/ with a default username/password of “admin” and “changeme”.

To complete the installation, log into Foreman, click on More > Configuration > Smart Proxies and add a new proxy with the URL https://fqdn:8443/. This will enable Puppet class imports from your new puppet master.

3.2.2 Installer Options

The installer is a collection of Puppet modules that can be operated using an answers file distributed inside the foreman_installer module. This is located at /usr/share/foreman-installer/foreman_installer/answers.yaml in the foreman-installer packages.

The answers file describes the classes that will be applied to the host to install Foreman. By default, the all-in-one setup will include Foreman, a puppet master and the Smart Proxy:

---
foreman: true
puppet: true
puppetmaster: true
foreman_proxy: true

This can be customized for your site by copying /usr/share/foreman-installer/foreman_installer/answers.yaml to /etc/foreman-installer/answers.yaml and editing. For example, a standalone puppet master and Smart Proxy can use:

---
foreman: false
puppet: true
puppetmaster: true
foreman_proxy: true

Other examples are given in /usr/share/foreman-installer/README.md. The installer itself provides a text driven interface to generate this answers file, and can be run by executing:

ruby /usr/share/foreman-installer/generate_answers.rb

3.3 Install From Packages

Packages are available for Red Hat and Debian-based distributions. This will install a standalone Foreman service running under WEBrick, which has limited scalability.

The Puppet-based Foreman installer is recommended for most environments, instead of installing only the packages as it will perform full configuration too.

3.3.1 RPM Packages

Foreman is packaged for the following RPM based distributions:

  • RHEL and derivatives, version 6
  • Fedora 18, 19

For most users, it’s highly recommended to use the installer as the packages only provide the software and a standalone Foreman service. The installer installs these packages, then additionally configures Foreman to run under Apache and Passenger with PostgreSQL, plus can configure a complete Puppet setup integrated with Foreman.

Pre-requisites

All RHEL and derivatives must be subscribed to EPEL to provide additional dependencies. Install epel-release from here to automatically configure it.

RHEL 6 hosts must also be subscribed to the “Optional” channel in RHN.

Optionally, the Puppet Labs repository can be configured to obtain the latest version of Puppet available, instead of the version on EPEL. See the Puppet Labs Package Repositories documentation on how to configure these.

Available repositories

Three main repos are provided at http://yum.theforeman.org:

  • /releases/latest or /releases/VERSION (e.g. /releases/1.2) carries the stable releases and updates of Foreman and its dependencies.
  • /rc carries release candidates only in the few weeks prior to a release. Ensure after a release is made that you use the main releases repo instead.
  • /nightly carries the latest development builds and as such, may be unstable or occasionally broken.

Under each repo are directories for each distribution, then each architecture.

Release packages

To set up the repository, foreman-release packages are provided which add a repo definition to /etc/yum.repos.d. Install the appropriate release RPM from these lists:

yum localinstall http://yum.theforeman.org/releases/1.2/el6/x86_64/foreman-release.rpm
yum localinstall http://yum.theforeman.org/releases/1.2/f18/x86_64/foreman-release.rpm
yum localinstall http://yum.theforeman.org/releases/1.2/f19/x86_64/foreman-release.rpm

For release candidate or nightly RPMs, change the URL as appropriate based on the above list of available repositories.

Signing

Release and release candidate packages are signed by the “Foreman Archive Signing Key packages@theforeman.org” (0xE775FF07). Nightly packages are not signed.

A copy of the key is available from public keyservers or here.

Available packages

Install foreman and other foreman-* packages to add functionality:

foreman               Foreman server
foreman-proxy         Foreman Smart Proxy
foreman-compute       EC2, OpenStack and Rackspace provisioning support
foreman-libvirt       libvirt provisioning support
foreman-ovirt         oVirt/RHEV provisioning support
foreman-vmware        VMware provisioning support
foreman-cli           Foreman CLI utility
foreman-console       Console additions
foreman-selinux       SELinux targeted policy
foreman-mysql         MySQL database support
foreman-mysql2        MySQL database support
foreman-postgresql    PostgreSQL database support
foreman-sqlite        SQLite database support

Setup

  1. Configure by editing /etc/foreman/settings.yaml and /etc/foreman/database.yml
  2. After changing the database, migrate it: sudo -u foreman /usr/share/foreman/extras/dbmigrate
  3. Start the foreman service or set up passenger: service foreman start

Upgrading from 1.1

Please take a backup of your database prior to attempting an upgrade, in case you need to rollback later.

To upgrade an existing Foreman 1.1 installation, first update with the appropriate foreman-release package from the above list of release packages, e.g.

yum upgrade http://yum.theforeman.org/releases/1.2/el6/x86_64/foreman-release.rpm
cat /etc/yum.repos.d/foreman.repo.rpmnew > /etc/yum.repos.d/foreman.repo

Next upgrade all Foreman packages and migrate the database:

yum upgrade foreman\*
sudo -u foreman /usr/share/foreman/extras/dbmigrate

If you use Apache and Passenger (the default installer configuration), additional changes are required to support SCL. First run rpm -q passenger-release to work out if the system has a Passenger package from upstream or EPEL.

If you have passenger-release installed:

rpm -e passenger-release
yum downgrade mod_passenger-4.0.5 rubygem-passenger-4.0.5 rubygem-passenger-native-4.0.5 rubygem-passenger-native-libs-4.0.5

or if you don’t have passenger-release:

yum upgrade \*passenger\*

Next, configure the new version of Passenger:

yum install ruby193-rubygem-passenger-native
sed -i.bak '/\(Rack\|Rails\)AutoDetect/d' /etc/httpd/conf.d/{foreman,puppet}.conf
sed -i.bak2 '/PassengerAppRoot/a\  PassengerRuby /usr/bin/ruby193-ruby' /etc/httpd/conf.d/foreman.conf
service httpd restart

It’s recommended to clear your browser’s cookies after the upgrade. If you wish to install the SELinux policy, also install the foreman-selinux package.

If Foreman doesn’t load due to a Passenger error (particularly referencing bundler_ext), verify the configuration against the description in the Software Collections section below.

3.3.2 Software Collections

The RPMs use a packaging technique called Software Collections, or SCL. This provides Ruby and all dependencies required to run Foreman separately from the version of Ruby provided by the distribution.

The current stack is “ruby193”, which provides Ruby 1.9.3 and Ruby on Rails 3.2. All packages will have a “ruby193-“ prefix, allowing them to be easily identified, and will install entirely underneath /opt/rh/ruby193.

The system Ruby version is left alone and will still be used for packages provided both by the distribution, and other third parties who target the system Ruby (e.g. Puppet packages).

Foreman currently uses SCL only on RHEL and EL clones where a newer version of Ruby is desired. Fedora only uses the distro version of Ruby.

In order to run rake commands for Foreman, or scripts that run in the same environment, ruby193-rake and ruby193-ruby wrappers are provided as alternatives for the usual rake or ruby. In order to run a series of commands or a script directly within the software collection, scl enable ruby193 'other command' can be used. It is also possible to run scl enable ruby193 bash to get a shell within the context of the SCL.

More general information about software collections is available from these links:

Passenger under the SCL

When running Foreman under Passenger (the default installer setup), a specific configuration is needed for SCL (on EL), since Foreman operates under the SCL Ruby and other apps such as the puppetmaster will use the system Ruby. Passenger 4 is shipped in the Foreman repos as it can be configured with separate Ruby binaries per VirtualHost. The full configuration is described below.

The following packages must be installed:

  • mod_passenger
  • ruby193-rubygem-passenger
  • ruby193-rubygem-passenger-native
  • ruby193-rubygem-passenger-native-libs
  • rubygem-passenger
  • rubygem-passenger-native
  • rubygem-passenger-native-libs

Ensure all version numbers match and are at least 4.0. mod_passenger provides the Apache module, while there are two copies of the Ruby components, one for the SCL Ruby (ruby193-rubygem-*) and one for the system Ruby (rubygem-*).

The /etc/httpd/conf.d/passenger.conf file is supplied by mod_passenger and should contain:

LoadModule passenger_module modules/mod_passenger.so
<IfModule mod_passenger.c>
   PassengerRoot /usr/lib/ruby/gems/1.8/gems/passenger-4.0.5
   PassengerRuby /usr/bin/ruby
</IfModule>

Check for .rpmsave or .rpmnew config backup files if this isn’t correct. Note that this refers to the system Ruby paths by default, allowing everything except Foreman (i.e. the puppetmaster) to run under the system Ruby.

Next, the Foreman config file at /etc/httpd/conf.d/foreman.conf must contain this entry in both HTTP and HTTPS VirtualHost sections:

PassengerRuby /usr/bin/ruby193-ruby

The full foreman.conf template from the installer is available here for comparison.

Ensure both RailsAutoDetect and RakeAutoDetect config entries are removed from foreman.conf and puppet.conf when using Passenger 4, since they have been deprecated.

When successfully configured, two Passenger RackApp processes will be visible and by inspecting the open files, the Ruby version being loaded can be confirmed:

# ps auxww | grep RackApp
foreman  16627  0.1 15.4 466980 157196 ?       Sl   07:35   0:09 Passenger RackApp: /usr/share/foreman
puppet   16697  0.8 11.3 253080 115720 ?       Sl   07:35   1:13 Passenger RackApp: /etc/puppet/rack
# lsof -p 16697 | grep libruby
ruby    16697 puppet  mem    REG              253,0   951224  272286 /usr/lib64/libruby.so.1.8.7
# lsof -p 16627 | grep libruby
ruby    16627 foreman  mem    REG              253,0  2041096  171190 /opt/rh/ruby193/root/usr/lib64/libruby.so.1.9.1

3.3.3 Debian Packages

The Foreman packages should work on the following Debian-based Linux distributions:

Distributions

  • Debian Linux 7.0 (Wheezy)
  • Debian Linux 6.0 (Squeeze)
  • Ubuntu Linux 12.04 LTS

Foreman might still work on these distros, but it has not been tested (let us know if it works for you):

  • Debian Linux 5.0 (Lenny)
  • Ubuntu Linux 11.04
  • Ubuntu Linux 10.10

If you encounter any errors during the installation, please file a bug report!

Apt Configuration

Add one of the following lines to your /etc/apt/sources.list (alternatively in a separate file in /etc/apt/sources.list.d/foreman.list):

# Stable packages

# Debian Wheezy
deb http://deb.theforeman.org/ wheezy stable
# Debian Squeeze
deb http://deb.theforeman.org/ squeeze stable
# Ubuntu Precise
deb http://deb.theforeman.org/ precise stable

# RC packages - these will only exist during an Release Candidate phase leading up to a release

# Debian Wheezy
deb http://deb.theforeman.org/ wheezy rc
# Debian Squeeze
deb http://deb.theforeman.org/ squeeze rc
# Ubuntu Precise
deb http://deb.theforeman.org/ precise rc

# Nightly builds. Beware: HERE BE DRAGONS

# Debian Wheezy
deb http://deb.theforeman.org/ wheezy nightly
# Debian Squeeze
deb http://deb.theforeman.org/ squeeze nightly
# Ubuntu Precise
deb http://deb.theforeman.org/ precise nightly

The public key for secure APT can be downloaded here

You can add this key with

apt-key add pubkey.gpg

or combine downloading and registering:

wget -q http://deb.theforeman.org/pubkey.gpg -O- | apt-key add -

The key fingerprint is

7059 542D 5AEA 367F 7873  2D02 B348 4CB7 1AA0 43B8
Foreman Automatic Signing Key (2014) <packages@theforeman.org>

Remember to update your package lists!

apt-get update

Install packages

The packages are now split by gem dependencies - there are 11 packages to choose from. These are:

Main package:

  • foreman

Database gems - you need at least one of these:

  • foreman-sqlite3
  • foreman-mysql2
  • foreman-mysql
  • foreman-pgsql

Optional functionality:

  • foreman-console
  • foreman-compute
  • foreman-libvirt
  • foreman-ovirt
  • foreman-test
  • foreman-vmware

Installation instructions are:

# Install packages  (adjust additional packages as needed)
apt-get install foreman foreman-sqlite3 foreman-libvirt

# Copy sample db config to /etc
cp /usr/share/foreman/config/database.yml.example /etc/foreman/database.yml

# Review settings and DB config
vi /etc/foreman/settings.yaml /etc/foreman/database.yml

# Perform initial DB setup
su - foreman -s /bin/bash -c /usr/share/foreman/extras/dbmigrate

The packages should auto-run db:migrate if /etc/foreman/database.yml exists. So the initial db:migrate is only needed during first install, upgrades should just work.

Upgrading from 0.4.2

Upgrading from 0.4.2 to 1.0 has been tested on Debian 6 (32 and 64 bit) and we expect that simply upgrading should work fine

apt-get update
apt-get dist-upgrade

The packages include a db:migrate, so your database will be migrated too. However, please note the new split-package architecture - you may want check and see if any of the new dependencies fit your usecase(s).

apt-cache search foreman

Upgrading from 1.0

Upgrading from 1.0 to 1.1 has been tested on Debian 6 (32 and 64 bit) and on Ubuntu 12.04 (32 and 64 bit). The only notable packaging change is that the packages are now self-sufficient - they no longer call out to Github/Rubygems during the postinst. This should be upgrade-safe (and has been tested) but odd environments may affect the upgrade. If in doubt, ask in IRC.

Please note foreman-fog has been renamed to foreman-compute - this should happen transparently through Apt

Upgrading from 1.1

Upgrading from 1.1 to 1.2 has been tested on Debian 6 and Ubuntu 12.04. Updating the packages will upgrade the application and automatically migrate the database. Please take a backup of your database prior to attempting an upgrade, in case you need to rollback later.

3.4 Install From Source

Installing the latest development code: Foreman has now moved to using Rails 3 and Bundler to get up and running. This is the preferred way to get Foreman if you want to benefit from the latest improvements. By using the git repository you can also upgrade more easily. You will need to have Bundler installed manually for now (check your distribution repositories, or install it via rubygems).

Foreman will run with the following requirements (aside from rubygem dependencies):

  • Ruby 1.8.7 or 1.9
  • rubygems
  • Puppet >= 0.24.4

The installation has been successfully tested on RHEL[5,6], Fedora[13,14,15,16,17], Debian Linux 6.0 (Squeeze) and Ubuntu Linux 12.04 (the community has reported varying success with other Debian/Ubuntu versions - 12.10 seems fine for example). For older operating systems you might need additional packages (e.g. sqlite). It is also known to work on Solaris and Mac.

to get latest “development” version do:

You will want to make sure that you have one of the mysql-devel, postgresql-devel, and sqlite-devel libraries installed so the database gems can install properly. also, you would also need gcc, ruby-devel, libxml-devel, libxslt-devel, and libvirt-devel packages

Additionally, it is important that config/database.yml is set to use the correct database in the “production” block. Rails (and subsequently Foreman) will use these connection settings under “production” to manage the database it uses and setup the necessary schema.

git clone https://github.com/theforeman/foreman.git -b develop
cd foreman
cp config/settings.yaml.example config/settings.yaml
cp config/database.yml.example config/database.yml
gem install bundler
# depending on database configured in yml file you can skip others
# (we are assuming sqlite to be configured)
bundle install --without mysql mysql2 postgresql test --path vendor
# set up database schema, precompile assets and locales
RAILS_ENV=production bundle exec rake db:migrate assets:precompile locale:pack

You can run Foreman with the command ”./script/rails s -e production”

Latest stable version

to get latest “stable” version do:

git clone https://github.com/theforeman/foreman.git -b 1.0-stable
cd foreman
git submodule init
git submodule update

If you are behind a proxy or firewall and don’t have access to Github using the git protocol, use http protocol instead:

git clone https://github.com/theforeman/foreman.git foreman
cd foreman
sed -i 's/git:\/\//http:\/\//g' .gitmodules
git submodule init
git submodule update

3.5 Configuration

The following sections detail the configuration steps required to get Foreman working in your environment. Lets get started!

3.5.1 Initial Setup

Configuration

Foreman configuration is managed from two places; a configuration file config/settings.yaml and from the SETTINGS/Foreman Settings page. A full description of the configuration options is given at foreman_configuration

Database

Foreman requires a database of its own to operate - database sharing is unsupported. By default, the installer uses PostgreSQL, while a package or source installation will use SQLite. If you want to use other database (e.g. MySQL) please modify the configuration file under config/database.yml.

In all cases, please use the production settings.

to initialize the database schema type:

cd (foreman installation path) && RAILS_ENV=production rake db:migrate

On foreman 1.0 use bundle:

cd (foreman installation path) && RAILS_ENV=production bundle exec rake db:migrate

For more information please see the database configuration page here

Import Data from Puppet

At this point, you might want to go through the [[FAQ]] to see how can you import your data into Foreman.

Start The Web Server

if you installed via rpm, just start the foreman service, or start the builtin web server by typing:

RAILS_ENV=production rails server

and point your browser to http://foreman:3000

If you would like to keep the server running, its recommend to setup passenger or use the RPM. Example usage with passenger can be found on GitHub.

Getting your Puppet Reports into Foreman

Read Puppet_Reports to learn how to get your nodes to report to Foreman.

3.5.2 Configuration Options

Configuration is broken into two parts. The config/settings.yaml file and the SETTINGS/Foreman Settings page. The configuration file contains a few low-level options that need to be set before Foreman starts but the majority of Foreman customization is managed from within the program on the Foreman Settings page.

The config/settings.yaml file

YAML start

The first non-comment line of this file must be three dashes.

---
login

This boolean option configures whether Foreman requires users to to login. If it is set then each user will be expected to authenticate themselves and all operations will occur, and be audited, under their identity. When this option is false then all activity will be executed under the admin account.

:login: true
require_ssl

This boolean option configures whether Foreman insists on using only https/ssl encrypted communication channels in the web interface. This does not configure the channels used to contact the smart-proxies. Note that certain operations will still accept a http connection even if this is set, for example, the downloading of a finish script.

:require_ssl: true
unattended

This boolean option configures whether Foreman will act as a simple node classifier for puppet, or support the full spectrum of operations required for managing a host’s lifecycle. When set to true then foreman will provide full host building facilities for various operating systems.

:unattended: true

The ‘More/Settings’ page

administrator

When Foreman needs to mail the administrator then this is the email address that it will contact. Default: root@.

default_puppet_environment

When Foreman receives a fact upload from a machine that it has not previously come across it will create a host in its database. If the facts from that host did not contain information about the puppet environment then it will assign the default_puppet_environment environment to this host. Default: production

Default_variables_Lookup_Path

A Smart-variable’s match criteria are evaluated in a specific order and if this search order is not provided then Default_variables_Lookup_Path is used. Default: [“fqdn”, “hostgroup”, “os”, “domain”]

document_root

This is the directory into which the rake puppet:rdoc:generate task will place the web-based on-line documentation for your puppet classes. It should be under a web server’s control and the default location is served by Foreman’s own web server. Default: public/puppet/rdoc

email_reply_address

The return address applied to outgoing emails. Default: Foreman-noreply@<your domain>

Enable_Smart_Variables_in_ENC

Whether Smart-variables should be included in the yaml node information provided to puppet. Default: true

entries_per_page

The number of entries that will be shown in the web interface for list operations. Default: 20

failed_report_email_notification

If this option is set to true then an email will be sent to the host’s owner whenever a report is received that contains errors. If the host is not owned or Foreman is not configured to use logins then send the email to the administrator. Default: false

foreman_url

Emails may contain embedded references to Foreman’s web interface. This option allows the hostname component of the URL to be configured. Default: foreman.

ignore_puppet_facts_for_provisioning

If this option is set to true then Foreman will not update a host’s IP and MAC with the values that it receives in a host’s facts and it will also include Foreman’s values for IP and MAC to puppet in its node information. Default: false

manage_puppetca

If this option is set to true then Foreman will manage a host’s Puppet certificate signing. If it is set to false then some external mechanism is required to ensure that the host’s certificate request is signed. Default: true

modulepath

This it the modulepath that foreman uses when processing puppet modules. It is usually able to determine this itself at runtime but if it is not able to find a value then modulepath is used. Default: /etc/puppet/modules

puppet_interval

This is the number of minutes between each run of puppet. Default: 30

puppet_server

The default puppet server hostname. For larger organizations this is often a non fqdn so that a name like puppet can be a different host within each DNS domain. Default: puppet

puppetconfdir

Path to puppet.conf. Default: /etc/puppet/puppet.conf

puppetrun

If this option is set to true then Foreman will be able to trigger a puppet run on any host that it manages. Default: false

root_pass

If a root password is not provided whilst configuring a host then this encrypted password is used when building the machine. The plain text password “123123” has been encrypted to produce the default value. Default: xybxa6JUkz63w (To generate a new one you should use: openssl passwd -1 “your_password” )

safemode_render

The default templating system used within Foreman allows unlimited interpolated variables and expressions. This could obviously be abused so a evaluation environment is provided that restricts the template variables and expressions to a whitelist. When this option is true then only known helper methods and instance variables will be available in template expansion. Default: true

ssl_ca_file

The SSL Certificate Authority file that Foreman will use when connecting to its smart-proxies. Default: The CA file used by puppet

ssl_certificate

The SSL certificate that Foreman will use when connecting to its smart-proxies. Default: The host certificate used by puppet

ssl_priv_key

The SSL private key file that Foreman will use when connecting to its smart-proxies. Default: The private key file used by puppet

update_environment_from_facts

If Foreman receives an environment fact from one of its hosts and if this option is true, it will update the host’s environment with the new value. By default this is not the case as Foreman should manage the host’s environment. Default: false

using_storeconfigs

Must be false, a shared storeconfigs database (which this setting refers to) is now unsupported.

3.5.3 Database Setup

Foreman is a rails application. Therefore, anything that is supported under RAILS (sqlite, mysql, postgresql, …) can be used. See 3.3 Install From Packages for a list of packages for connecting foreman to the databse of your choice. At this time, Oracle DB is known to not work. Patches are welcome!

The database configuration file can be found at:

/etc/foreman/database.yml

SQLite (default)

When using SQLite, you can install the foreman-sqlite package. See 3.3 Install From Packages.

By default, the database will be created in the db subdirectory.

production:
  adapter: sqlite3
  database: db/production.sqlite3
  pool: 5
  timeout: 5000

MySQL

When using MySQL, you can install the foreman-mysql package. See 3.3 Install From Packages.

Edit your config/database.yml and modify:

production:
  adapter: mysql
  database: puppet
  username: puppet
  password: password
  host: localhost
  socket: "/var/run/mysqld/mysqld.sock"

If you use foreman 1.0 with foreman-mysql2 use ‘adapter: mysql2’ instead

Afterwards you would need to re-populate your database, simply execute extras/dbmigrate script.

PostgreSQL

When using PostgreSQL, you can install the foreman-postgresql package. See 3.3 Install From Packages.

Edit your config/database.yml and modify:

production:
  adapter: postgresql
  database: foreman
  username: foreman
  password: password
  host: localhost

Switching from SQLite to MySQL/PostgreSQL while maintaining existing data

We have a rake task for this. First setup your database.yml to have the sqlite db as production and the mysql/psql db as dev:

production:
  adapter: sqlite3
  database: db/production.sqlite3
  pool: 5
  timeout: 5000

development:
  adapter: postgresql
  database: foreman
  username: foreman
  password: password
  host: localhost

Now migrate both dbs so they’re consistent:

bundle exec rake db:migrate RAILS_ENV=production
bundle exec rake db:migrate RAILS_ENV=development

Now move the data to the new db

bundle exec rake db:convert:prod2dev

(On RPM distros, remove “bundle exec” from the start of these commands.)

Once you’ve migrated to your new database using prod2dev, you should update your database.yml file to point your ‘production’ environment at the new database. You should also update the ‘development’ environment to point to an alternative location (for example, at SQLite) to ensure you don’t accidentally overwrite your production database.

Special note for migrating to Postgres

The psql sequence numbers will be wrong after the prod2dev execution. You can fix them like this:

cat <<EOF > reset.sql
SELECT  'SELECT SETVAL(' ||quote_literal(S.relname)|| ', MAX(' ||quote_ident(C.attname)|| ') ) FROM ' ||quote_ident(T.relname)|| ';'
FROM pg_class AS S, pg_depend AS D, pg_class AS T, pg_attribute AS C
WHERE S.relkind = 'S'
    AND S.oid = D.objid
    AND D.refobjid = T.oid
    AND D.refobjid = C.attrelid
    AND D.refobjsubid = C.attnum
ORDER BY S.relname;
EOF
psql -Atq -f reset.sql -o temp foreman
psql -f temp foreman
rm temp reset.sql

(big thanks to Fixing Sequences for the fix)

3.5.4 Puppet Reports

Foreman uses a custom puppet reports address (similar to tagmail or store) which Puppet will use to upload it’s report into Foreman. This enables you to see the reports through the web interface as soon as the client finish its run.

Configuration

Client

Ensure that the puppet clients has the following option in their puppet.conf:

report = true

Without it, no reports will be sent.

puppetmaster
  1. copy https://raw.github.com/theforeman/puppet-foreman/master/templates/foreman-report_v1.rb.erb to your report directory
  2. e.g. /usr/lib/ruby/1.8/puppet/reports/foreman.rb, /usr/lib/ruby/site_ruby/1.8/puppet/reports/foreman.rb, /var/lib/gems/1.8/gems/puppet-2.6.4/lib/puppet/reports/foreman.rb or /usr/lib/ruby/vendor_ruby/puppet/reports.
  3. make sure you copied the foreman-report_v1.rb.erb to foreman.rb so puppet can find it!
  4. open the new file with your favorite editor
  5. edit the URL to point to foreman.
  6. add this report in your puppetmaster reports - e.g, in your master puppet.conf under the main section add:
reports=log, foreman

and restart your puppetmaster

You should start seeing reports coming in under the reports link.

Expire Reports automatically

You will probably want to delete your reports after some time to limit database growth. To do so, you can set a cronjob:

Available conditions:

  • days => number of days to keep reports (defaults to 7)
  • status => status of the report (defaults to 0 –> “reports with no errors”)

Example:

  1. Expires all reports regardless of their status
    rake reports:expire days=7 RAILS_ENV="production"
  2. expires all non interesting reports after one day
    rake reports:expire days=1 status=0 RAILS_ENV="production"

Note: on Debian systems, these commands must be run as the foreman user from its default home directory.

3.5.5 Importing Facts

There are several ways you can get your facts visible in Foreman.

If you are planning using Foreman as your Puppet External Nodes Classifier, the recommended approach would be using the ENC script, which both updates your facts, and fetch the information at the same time. If you are not using the ENC features, the ENC script will still serve as an example of how to upload facts via an HTTP POST request. See Pushing facts through HTTP for another example.

Running Foreman on the puppetmaster

If you run Foreman on the same machine as your puppetmaster and you have only one puppet master, you probably want to create a simple cronjob:

rake puppet:import:hosts_and_facts RAILS_ENV=production

This will import your facts (only new facts) every time you run the script.

Multiple puppet masters or Foreman on another server

Pushing facts through HTTP directly to Foreman

Foreman accepts remote puppetmaster sending their nodes facts through http, making it effective way to send facts over WAN.

This script is meant to run on each of your puppetmasters. The script will scan over all of the new fact files (from its last run) and will send them to foreman.

You will need to edit the script and set the Foreman URL (based on where Foreman is located in your setup).

You should run this script in a cronjob, if you have many puppet masters, you might consider adding a simple sleep argument (as mentioned inside the script header), this will avoid all of your puppetmasters hammering your database at the same time.

local file system access to fact yaml files

The previously mentioned rake task accepts a dir parameter, which means you can tell it to import facts from any directory. if you store your fact yaml on some network share, or if you want to mount your puppetmaster you can:

rake puppet:import:hosts_and_facts dir=/my/dir/with/yaml/files RAILS_ENV=production

However, I would not recommend this option if you have any network latency or if it requires you to run an NFS (or sshfs) etc

4. General Foreman

This section covers general information on using Foreman to manage your infrastructure. It covers the features of the web interface, managing puppet, provisioning systems and the installation and configuration of Foreman Smart Proxies.

4.1 Web Interface

4.1.1 LDAP Authentication

Foreman natively supports LDAP authentication using one or multiple LDAP directories.

Enabling LDAP

Enable LDAP and User/Group settings menus

Edit your config/setting.yml

:login: true

and restart Foreman

Setting up

Go to More > Users > LDAP Authentication

Click on New LDAP Source and enter the following

  • Name: an arbitrary name for the directory
  • Host: the LDAP host name
  • Port: the LDAP port (default is 389)
  • LDAPS: check this if you want or need to use LDAPS to access the directory
  • Account: leave this field empty if your LDAP can be read anonymously, otherwise enter a user name that has read access to the LDAP or use $login (which will be replaced with the actual user credentials upon login)
  • Account Password: password for the account (if defined above and its not using the $login)
  • baseDN: the top level DN of your LDAP directory tree

On the fly user creation

By checking On-the-fly user creation, any LDAP user will have his Foreman account automatically created the first time he logs into Foreman. For that, you have to specify the LDAP attributes name (firstname, lastname, email) that will be used to create their Foreman accounts.

Examples

Active Directory
Name              = My Directory
Host              = host.domain.org
Port              = 636
TLS               = yes
Onthefly register = yes
Account           = MyDomain\$login
Password          = (leave blank)
Base DN           = CN=users,DC=host,DC=domain,DC=org
attr login        = sAMAccountName
attr firstname    = givenName
attr lastname     = sN
mail              = mail
OpenLDAP
Name              = My Directory
Host              = host.domain.org
Port              = 389
TLS               = no
Onthefly register = yes
Account           = (leave blank if anonymous access is enabled)
Password          = (leave blank)
Base DN           = ou=Users,dc=domain,dc=co,dc=il
attr login        = uid
attr firstname    = givenName
attr lastname     = sn
mail              = mail

Note that LDAP attribute names are case sensitive.

Troubleshooting

If you want to use on-the-fly user creation, make sure that Foreman can fetch from your LDAP all the required information to create a valid user. For example, on-the-fly user creation won’t work if you don’t have valid email addresses in your directory (you will get an ‘Invalid username/password’ error message when trying to log in).

4.1.2 Roles and Permissions

A user’s access to the features of Foreman are constrained by the roles and permissions that they are granted. These permissions are also used to restrict the set of hosts, host groups and domains that a user is able to access and modify.

Note: a user with global admin enabled is not restricted by the authorization system. This is the default for installations that do not have :login:true in config/settings.yml.

A logged in user will be granted the Anonymous role plus one or more additional roles. The permissions associated with these roles are aggregated and determine the final permission set.

Roles may be administered only by a user with global admin privileges.

Roles

These may be created, deleted and edited on the Roles page. Each role will be associates with one or more base privileges.

There are two builtin system roles

  1. Anonymous: This is a set of permissions that every user at your installation will be granted, irrespective of any other roles that they have.
  2. Default user: When a new Role is created this set of permissions are used as the template for the Role. The name is somewhat misleading but basically an ordinary default user who was assigned this Role would have these permissions set.

Permissions

These determine the operations that are allowed to be performed upon the items to which they refer. For simple items, like an architecture, this operates as expected but for more complex items, such as, the hosts a user is able to operate on, there is an additional layer of security called filtering.

When editing a user account there is a section at the bottom that narrows the scope of the permissions granted to a subset of the hosts, domains and host groups. See filtering below.

Permission Description
Permissions for Architectures, Authentication providers, environments, External variables, Common parameters, Medias, Models, Operating systems, Partition tables, Puppet classes and User groups
view The user is allowed to see this type of object when listing them on the index page
create The user is allowed to create this type of object
edit The user is allowed to edit this type of object
destroy The user is allowed to destroy this type of object
Permissions for Domains
view The user is allowed to see a list of domains when viewing the index page
create The user is allowed to create a new domain and will also be able to create domain parameters
edit The user is allowed to edit a domain and will also be able to edit a domain's parameters. If they have domain filtering active in their profile then only these domains will be editable
destroy The user is allowed to destroy a domain and will also be able to destroy domain parameters. If they have domain filtering active in their profile then only these domains will be deletable
Permissions for Host groups
view The user is allowed to see a list of host groups when viewing the index page
create The user is allowed to create a new host group and will also be able to create host group parameters
edit The user is allowed to edit a host group and will also be able to edit a host group's parameters. If they have host group filtering active in their profile then only these host groups will be editable
destroy The user is allowed to destroy a host group and will also be able to destroy host group parameters. If they have host group filtering active in their profile then only these host groups will be deletable
Permissions for Hosts
view The user is allowed to see a list of hosts when viewing the index page. This list may be constrained by the user's host filters
create The user is allowed to create a new host. This operation may be constrained by the user's host filters
edit The user is allowed to edit a host. This operation may be constrained by the user's host filters
destroy The user is allowed to destroy a host. This operation may be constrained by the user's host filters
Permissions for Users
view The user is allowed to see a list of users when viewing the index page. A user will always be able to see their own account even if they do not have this permission
create The user is allowed to create a new user
edit The user is allowed to edit existing users. A user will always be able to edit their own basic account settings and password
destroy The user is allowed to delete users from the system

Filtering

If the filtering section at the bottom of the user’s profile page has no content then the permissions that the user has been granted will apply to all hosts within the system.

However, if the filtering section is in use then the permissions will apply only to those items selected in the filters and the user will have no access to anything not selected by the filters.

This is primarily a mechanism for restricting access to hosts. However if one or more domains or host groups are selected then this also restricts where parameters can be created, edited and deleted.

Filtering operates by generating a list of hosts on which actions can be performed. The list may be built out of four components

  1. Ownership: The hosts that a user owns directly or hosts that are owned by a user group of which the user is a member.

  2. Domain membership: The hosts that exist within one or more indicated domains.

  3. Compute resource membership: The hosts that are deployed on one or more compute resources.

  4. Host group membership: The hosts that are defined as being of one or more host group types.

  5. Fact filtering: These restrict the hosts to those machines that have this fact associated with them. As a fact is only generated during a puppet run, this filter will only refer to machines that have been built and therefore cannot be used to restrict the creation of machines.

These four pools of hosts can be combined by adding them together or the filters can be used to restrict the selected hosts to a smaller and smaller subset of the total. Think of it as set operations.

Note: If the “Administrator” check box is checked for a user, filtering will not take effect.

Example

A user’s filter section has the select checkbox unticked, indicating that the user’s owned hosts are not included in the final set of hosts, (unless they are selected some other way.)

The domain selection is prefixed with plus all and the two domains a.com and b.com are ticked. This implies that the hosts selected are the user’s hosts, (of which there are none,) plus all the hosts in a.com and b.com.

The host group section is prefixed by must be and the hostgroup web server is selected. This implies that the hosts selected are the user’s hosts, (of which there are none,) plus all the hosts in a.com and b.com but they must be of type web server.

The fact filter section is prefixed by must match and there are two filters virtual = vmware and architecture = i386. This implies that the hosts selected are the user’s hosts, (of which there are none,) plus all the hosts in a.com and b.com but they must be of type web server and must match an i386 vmware host.

In any of these sections above, if the prefix had been plus all then every host in the system that matched the selections would have been added to the final host list, (though possibly removed by a later filter.)

Trends in Foreman allow you to track changes in your infrastructure over time. It allows you to track both Foreman related information and any puppet facts. The Trend pages give a graph of how the number of hosts with that value have changed over time, and the current hosts list.

There are two pieces to the Trends area, the Trends to track and their corresponding counters. To define trend counters, use the “Add Trend Counter” button on the ‘/trends’ page. Optionally set the “Name” field to over-ride odd puppet fact names to be more readable. Once created you can optionally ‘Edit’ the Trend to change the display names of the underlying values.

Next, to start collecting trend data, set a cron job to execute ‘rake trends:counter’. Each time the rake task executes it will create 1 tick on the graphs, so you can fine tune the granularity with your cron job. We recommend using the same as your puppet run interval (30 minutes). Here’s an example to run once per hour:

0 * * * * cd /usr/share/foreman/ && /usr/bin/rake trends:counter

Finally note that these trends are the same for all users. So if you delete a Trend category you will loose all history for that trend and the trackers will start all over again. So please be careful when deleting.

4.1.4 Auditing

Foreman supports auditing of almost all changes that happen within Foreman, from both the UI and from the API. Auditing is done at a user level, and is thus ineffective if :login: is set to false, as all audits will be done as the ‘admin’ user.

Basic View

Got to the Audit tab to see a view of what has changed. This view can be filtered by the type of change or by the object that was altered (e.g. search for a hostname to see all changes relating to that host). You also get the parent object - so if a parameter was modified, you can see what host/group that parameter belongs to. The timestamp of the change and the user who performed it will be listed.

Extended Audits for Templates

Template changes also store a diff of the changes, and the ability to roll back to a previous version of the template.

4.2 Managing Puppet

In this section we’ll look at the various ways we can control and interact with Puppet.

4.2.1 Environments

Puppet environments are mapped directly into Foreman. They can be used at various levels throughout the Foreman interface. Puppet environments are generally used to separate classes from different types of Host, typically allowing changes to a module to tested in one environment (e.g. development) before being pushed to another (e.g production).

Defining Environments

There are several ways to create Puppet environments within Foreman.

Manual creation

To create an environment by hand, simply go to More > Configuration > Environments and click New Environment. Give the new environment a name and save.

Importing from Puppet

Foreman can detect all the environments and classes contained on a PuppetMaster, and import them automatically. To do this, go to More > Configuration > Environments and click on *Import from *. Foreman will scan the PuppetMaster, and display a confirmation of the detected changes. Select the changes you wish to apply and confirm.

Assigning Environments to Hosts

This is done from the Host Edit page, on the Host tab. Selecting an environment will filter the classes visible on the Puppet Classes tab to just the classes in the selected environment.

You can also also mass-assign an environment to a group of hosts - tick the checkboxes of the required Hosts in the Hosts index, and then select Change Environment from the buttons at top of the page.

Environments with Hostgroups

You can assign an environment to a hostgroup as well. This functions as a form of default - a user creating a new host and selecting the hostgroup will automatically have the environment pre-selected. The user is not prevented from changing the environment of the new host, it simply saves a few clicks if they are happy with it.

4.2.2 Classes

Puppet classes are generally imported from the Puppet Master(s) via the Import button on the Puppet Classes page. They can also be created by hand, and manually associated with a set of environments (for filtering purposes).

Importing Classes

Go to More > Configuration > Puppet Classes and click the Import button. This will not be visible unless you have at least one Puppet Master with a puppet-enabled Smart Proxy

Ignoring classes on import

It’s often to have a module structure like this:

$ tree git/
git/
└── manifests
    ├── init.pp
    ├── install.pp
    ├── params.pp
    └── repo.pp

In this situation, Foreman would offer to create:

git
git::install
git::params
git::repo

However, if we know that the subclasses are not intended for direct consumption, but are only really part of the internal structure of the module, then we would want to exclude those from the import mechanism, so that Foreman only offers to import git. We can achieve this via the file config/ignored_environments.yml. This file takes a set of regular expressions - any class which matches one of them will not be imported. So, for this example, we might configure:

:filters:
  - !ruby/regexp '/install$/'
  - !ruby/regexp '/params$/'
  - !ruby/regexp '/repo$/'

Assigning Classes to Hosts

To cause Puppet to apply your classes, you will need to assign them to your Hosts. This can be done at either an individual host level, or at a group level. The process is the same; edit the Host(group), select an Environment, and then go to the Puppet Classes tab and select what classes you want in this Host(Group).

Checking the results

To see how Foreman is passing the classes to Puppet, go to a Host and click the YAML button. You will be shown the exact YAML data sent to the PuppetMaster - the classes will be in the “classes” hash.

4.2.3 Parameters

Foreman’s concept of parameters maps onto Puppet’s idea of default-scope parameters. Foreman allows us to define a hierarchy of parameter inheritance.

Global Parameters

These are defined in More > Configuration > Global Parameters and will apply to every host in Foreman.

Domain Parameters

These are defined for all Hosts in a given domain. Edit the domain from More > Provisioning > Domains and switch to the Parameters tab and specify a parameter. If it has the same name as a Global Parameter, it will override the Global one.

Hostgroup Parameters

These are defined for all Hosts in the Group. Edit the Hostgroup from More > Configuration > Host Groups and switch to the Parameters tab and specify a parameter. If it has the same name as a Global or Domain Parameter, it will override it.

Host Parameters

The final (most-specific) level of Parameters applies only to a single Host. Edit a Host and switch to Parameters, and you will see all it’s inherited parameters from the other three layers (note: they will all be marked as “Scope: Global” as this refers to the Puppet scope, not the Foreman scope). You can override higher-level parameters or define new ones here.

Checking the results

To see how Foreman is passing the parameters to Puppet, go to a Host and click the YAML button. You will be shown the exact YAML data sent to the PuppetMaster -the parameters will be in the “parameters” hash.

4.2.4 Smart Variables

Smart variables are a tool to provide data (Key / Value), normally to your puppet ENC, depending on a set of rules. They are intended to be a stepping stone to full parameterized classes.

Smart variable is usually associated with a puppet class, and may have multiple values, all depending on hierarchical context or various conditions a user can wish to apply.

for example:

Example

Creating a Smart Variable

Start by going to Foreman > More > Puppetclasses and then click one of your classes to edit it.

Click on the Smart Variables tab. If you have any existing Smart Variables, they will be listed at the left side of the tab.

Click New Variable, and you will be presented with a set of input fields:

Name What the parameter will be called in the ENC data
Description A free form text box for your own information
Default Value What the ENC will use if no other criteria is matched
Type Validator A combo-box of data types. The type applies to the next field, the validator
Validator Constraint Used to enforce certain values for the Smart Variable. See below for examples
Order A list of variables which Foreman will search (in order) for the validator

Once you’ve created the defaults for your Smart variable, you then need to add at least one criterion to match against - click the “plus” under your variable, and two more input fields will appear:

Match Should state a name = value relationship that Foreman use to match against the entries in searchlist
Value What the parameter should be in the ENC, if this rule is matched

Validators

The fourth and fifth fields in the Smart Variable combine to produce a validation criteria for the final value of the Smart Variable.

String validators

At present, the string type cannot be validated - leave the validator field blank, and all strings in the variable will be considered acceptable

Regexp / List validators

By entering a list (comma-separated, no spaces) or a regex (no delimiter required), the value to be assigned to the Smart Variable will be checked against this list. If the value does not match the validator, and error will be raised.

Ordering

All the ordering information is stored in the sixth field of the Smart Variable (order). The order matters and is used to find the first match.

Examples

Example 1 - Simple change to a single host

All our hosts use server.foo for something, except bob.domain.com which uses server2.bar:

Name target
Description The target server to talk to
Default Value server.foo
Type Validator string
Validator Constraint
Order fqdn
hostgroup
os
domain
Match fqdn = bob.domain.com
Value server2.bar
Example 2 - Change for a group of hosts (via custom fact) with validation and ordering

Most hosts need to use a port of 80 but all machines with a fact region and value europe need to use 8080. To do this, you have to add the factname (in this example region) to the searchlist:

Name port
Description The port to use
Default Value 80
Type Validator list
Validator Constraint 80,443,8080
Order fqdn
region
hostgroup
os
domain
Match region = europe
Value 8080
Match fqdn = foo.domain
Value 67

Note that all machines will get either 80 or 8080 as required, except foo.domain which will generate an error, since 67 is not in the list validator. Note also that foo.domain will match before region, since it is higher in the searchlist. The rule ordering does not matter.

It is also possible to mix conditions, e.g.

Name port
Description The port to use
Default Value 80
Type Validator list
Validator Constraint 80,443,8080
Order fqdn
region, hostgroup, environment
hostgroup
environment
domain
Match fqdn = foo.domain
Value 67
Match region = europe, hostgroup = "web servers", environment = production
Value 8080

API usage

It’s also possible to retrieve these values if you’re not using a ENC, via a custom Puppet function or a http request.. You’ll need to retrieve the value from

https://foreman/hosts/<fqdn>/lookup_keys/<key>

You can find a ready-made function for your puppet module here

4.2.5 Parameterized Classes

Parameterized Class Support (PCS) permits detecting, importing, and supplying parameters direct to classes which support it, via the ENC.

Requirements

  • Foreman 1.1+
  • Foreman-Proxy 1.1+
  • Puppet 2.6.5+

Setup

Firstly, you’ll need to enable the PCS support. Go to More > Settings, select the Puppet tab, and ensure Parameterized_Classes_in_ENC is set to true.

Settings

Now you’ll need to import some parameterized classes from your puppet master. If you don’t have any parameterized classes in your modules dir, the foreman-installer has several, you can download a few modules from the Puppet Forge. Once you have some parameterized modules, import your classes (see 4.2.2 Classes)

Import

Configure a class

This example will work with the foreman class from the installer. Click on the class, and you should get a page with 3 tabs, like so:

3 Tabs

The middle tab, “Smart Class Parameter”, is the important one. Click onto that, and you should see something like this:

Edit

On the left, we have a list of possible parameters that the class supports. On the right, we have the configuration options for the parameter selected.

Lets configure the foreman class to change the user the foreman processes run as. Select the user parameter, at the end of the list. Now lets go through the options:

  • Puppet Environments / Name
    • These can’t be edited, they’re just for information.
  • Description
    • Purely information textbox for making notes in. Not passed to puppet, or reused anywhere else
  • Override (important)
    • If this is unchecked, Foreman will not attempt to control this variable, and it will not be passed to Puppet via the ENC.
  • Type
    • The type of data we want to pass. Most commonly a string, but many other data types are supported. There’s no easy way to tell what type of data puppet is expecting, so you will need to read through the code/documentation that comes with a particular module to find out. Changing the type field requires an appropriately set “Default Value” field.
  • Default Value
    • This will be imported from puppet initially, but if puppet is using any class inheritance, you’ll get something unhelpful like “${$foreman::params::user}”. This is because Foreman won’t follow the inheritance, so you’ll need to set a sensible default value

Ok, so let’s configure our user parameter. We want to tick Override, set type to “String” and set the default value to “foreman”, like so:

User Param

Setting up Matchers

We’ve configured the default, but that’s not very useful. We need to be able to override the default for hosts or groups of hosts. To do that we need the “Override Value For Specific Hosts” section at the bottom of the page.

Let’s say that any machine in the “development” puppet environment should use a value of “foremandev” instead of “foreman” for the “user” parameter. Add “environment” to the end of the matchers list, then click the “New Matcher-Value” button, and fill it out like this:

Matcher

This is a basic configuration - for more complex examples of using matchers, see the 4.2.4 Smart Variables page.

Complex Data

Here’s an example of adding an array parameter. Note the use of YAML in the editbox:

Array

This will be converted to the normal [“a”,”b”] syntax when you save. You can also use Hashes, YAML or JSON as data types too.

Editing Param from within a Host

If Foreman manages the value of a class parameter (“override = true”), it’s also possible to update a host-specific override from the host itself. That way you don’t have to grant access to the Puppet Classes page to everyone. From a Host, click Edit, go to the Parameters tab, and you’ll see the variable, the class-scope, and the current value. You can then override the value for that host:

Host Edit

If you go back and look at the Puppet class, you’ll see Foreman has added a matcher for that host:

Host Matcher

Currently this only works for Hosts, not Hostgroups. For more complex logic, like matching on facts, use the Puppet Class page

Input Validation

The “Optional Input Validator” section can be used to restrict the allowed values for the parameter. This functions in the same way as for Smart Variables, but it is important to note that the validation applies to changes made from the Host edit page as well as the Puppet Classes edit page.

For example, to restrict the “user” field to either “foreman” or “foremandev”, tick the Required checkbox, and then set:

  • Type: List
  • Rule: foreman,foremandev

4.3 Smart Proxies

The Smart Proxy is a project which provides a restful API to various sub-systems.

Its goal is to provide an API for a higher level orchestration tools (such as Foreman). The Smart proxy provides an easy way to add or extended existing subsystems and API’s.

Currently supported (Click on the links below for more details).

  • DHCP - ISC DHCP and MS DHCP Servers
  • DNS - Bind and MS DNS Servers
  • TFTP - any UNIX based tftp server
  • Puppet - Any Puppet server from 0.24.x
  • Puppet CA - Manage certificate signing, cleaning and autosign on a Puppet CA server

If you require another sub system type or implementation, Please add a new feature request.

After you got it running, You would need to configure each one of the sub systems via the settings.yml file in the config directory.

4.3.1 Smart Proxy Installation

A smart proxy is an autonomous web-based foreman component that is placed on a host performing a specific function in the host commissioning phase. It receives requests from Foreman to perform operations that are required during the commissioning process and executes them on its behalf. More details can be found on the Foreman Architecture page.

To fully manage the commissioning process then a smart proxy will have to manipulate these services, DHCP, DNS, Puppet CA, Puppet and TFTP. These services may exist on separate machines or several of them may be hosted on the same machine. As each smart proxy instance is capable of managing all the of these services, there is only need for one proxy per host. In the special case of a smart proxy managing a windows DHCP server, the host machine must be running Windows and support the netsh dhcp utility, it does not need to be the Microsoft DHCP server itself.

Packages

RPM and Debian packages are available, see the Install from Packages section for configuration and install the foreman-proxy package.

Source code

You can get the latest stable code from GitHub (via git).

git clone git://github.com/theforeman/smart-proxy.git

Configuration file

Usually can be found at /etc/foreman-proxy/settings.yml or on the config/settings.yml subdirectory. You can use the settings.yml.example file inside the config directory as a template for your own settings.yml.

If you don’t plan to use one of the subsystems, please disable them in this configuration file. For more information see Smartproxy Configuration

Start the daemon

  bin/smart-proxy.rb

Or if you installed it via a package simply start the foreman-proxy service.

Add the smart-proxy to Foreman

  • Go to [FOREMAN_URL]/smart_proxies and click New Proxy
  • Type in the Name for your Proxy and the URL of your Proxy, with the Port used.

For example:

  Name: Puppet-Proxy
  URL: http://puppet.your-domain.com:8443 

4.3.2 Smart Proxy Settings

The configuration for Smart-Proxy is held in the /etc/foreman-proxy/settings.yml or config/settings.yml file.

YAML start

The first non-comment line of this file must be three dashes.

---

SSL configuration

The existence of all the three ssl key entries below enables the use of an SSL connections.

NOTE that both client certificates need to be signed by the same CA, which must be in the ssl_ca_file, in order for this to work see SSL for more information

  :ssl_certificate: ssl/certs/fqdn.pem
  :ssl_ca_file: ssl/certs/ca.pem
  :ssl_private_key: ssl/private_keys/fqdn.key

This is the list of hosts from which the smart proxy will accept connections. If this list is empty then every verified SSL connection is allowed to access the API.

:trusted_hosts:
- foreman.prod.domain
- foreman.dev.domain

Instance attributes

If this entry is present and not false then Smart-Proxy will attempt to disconnect itself from the controlling terminal and daemonize itself.

:daemon: true

The port listened to by the proxy. If this is not present then the default Sinatra port of 4567 is used.

:port: 8443

TFTP section

Activate the TFTP management module within the Smart-Proxy instance.

The tftproot value is directory into which tftp files are copied and then served from. The tftp daemon will also be expected to chroot to this location. This component is only supported in the Unix environment

:tftp: true
:tftproot: /var/lib/tftpboot
:tftp_servername: name of your tftp server (used for next server value in your dhcp reservation) - defaults to the host name of your proxy.

NOTE: the foreman proxy user must have read/write access to the tftpboot/pxelinux.cfg and tftpboot/boot directories.

DNS section

Activate the DNS management module within the Smart-Proxy instance.

The DNS module can manipulate any DNS server that complies with the ISC Dynamic DNS Update standard and can therefore be used to manage both Microsoft and Bind servers. Updates can also be done using GSS-TSIG, see the documentation.

The dns_key specifies a file containing a shared secret used to generate a signature for the update request (TSIG record). This option should not be used if you plan to use Kerberos/GSS-TSIG (for example for DNS servers shipped with FreeIPA or Microsoft AD).

If neither the dns_key or GSS-TSIG is used then the update request is sent without any signature. Unsigned update requests are considered insecure. Some DNS servers can be configured to accept only signed signatures.

The dns_server option is used if the Smart-Proxy is not located on the same physical host as the DNS server. If it is not specified then localhost is presumed.

:dns: true
:dns_key: /home/proxy/keys/Kapi.+157+47848.private
:dns_server: dnsserver.site.domain.com

NOTE: if you use a key, make sure that the foreman proxy account can read that file.

DHCP section

Activate the DHCP management module within the Smart-Proxy instance.

:dhcp: true

If the DHCP server is ISC compliant then set dhcp_vendor to isc. In this case Smart-Proxy must run on the same host as the DHCP server. If the proxy is managing a Microsoft DHCP server then set dhcp_vendor to native_ms. Smart-Proxy must then be run on an NT server so as to access the Microsoft native tools, though it does not have to be the same machine as the DHCP server. More details can be found at [[Foreman:Foreman Architecture]].

:dhcp_vendor: isc

The DHCP component needs access to the DHCP configuration file as well as the currently allocated leases. The section below shows these values for a RedHat client. In the case of a Smart-Proxy hosted on an Ubuntu machine then these values would be more appropriate: /etc/dhcp3/dhcpd.conf and /var/lib/dhcp3/dhcpd.leases

:dhcp_config: etc/dhcpd.conf
:dhcp_leases: etc/dhcpd.leases

NOTE: Make sure that the foreman proxy account can read both ISC configuration files.

If your native_ms implementation is slow then you can request that the smart proxy only operate on a subset of the subnets managed by the dhcp server.

:dhcp_subnets: [192.168.1.0/255.255.255.0, 192.168.11.0/255.255.255.0]

If you secured your DHCP with an “omapi_key”, add the entries:

:dhcp_key_name: omapi_key
:dhcp_key_secret: XXXXXXXX

Puppet Certificate Authority section

Activate the Puppet CA management module within the Smart-Proxy instance.

This should only be enabled in the Smart-Proxy that is hosted on the machine responsible for providing certificates to your puppet clients. You would expect to see a directory /var/lib/puppet/ssl/ca on such a host.

:puppetca: true

If your puppet SSL directory is located elsewhere, you’ll need to set ‘ssldir’ as well.

:ssldir: /etc/puppet/ssl
:puppetdir: /etc/puppet

The proxy requires write access to the puppet autosign.conf file, which is usually owner and group puppet, and has mode 0644 according to the puppet defaults.

Ensure the foreman-proxy user is added to the puppet group ( e.g. gpasswd -a foreman-proxy puppet or usermod -aG puppet foreman-proxy)

puppet.conf:

[master]
autosign = $confdir/autosign.conf {owner = service, group = service, mode = 664 }

Sudo access to the proxy is required - in your sudoers file ensure you have the following lines:

For older puppet (pre-3.0) with separate sub-commands available:

foreman-proxy ALL = NOPASSWD: /usr/sbin/puppetca *
Defaults:foreman-proxy !requiretty

For newer monolithic puppet without separate commands (3.0-onwards)

foreman-proxy ALL = NOPASSWD: /usr/bin/puppet cert *
Defaults:foreman-proxy !requiretty

Puppet section

Activate the puppet management module within the Smart-Proxy instance.

This should only be enabled in the Smart-Proxy that is hosted on the machine capable of executing puppetrun. This will be a puppetmaster. This can also be set to true if you need to import puppet classes from the puppetmaster. Without this the import will not be possible

:puppet: true
:puppet_conf: /etc/puppet/puppet.conf
# Defaults to %INSTALL_DIR%/.puppet/puppet.conf
puppet run/kick

Sudo access for the proxy is required - in your sudoers file ensure you have the following lines (use /opt/puppet/bin/puppet for Puppet Enterprise):

Defaults:foreman-proxy !requiretty
foreman-proxy ALL = NOPASSWD: /usr/sbin/puppetrun

If you are using Puppet 3.0 or higher, the puppetrun binary has been removed and so the Smart Proxy will use puppet kick. The sudoers entry should be:

Defaults:foreman-proxy !requiretty
foreman-proxy ALL = NOPASSWD: /usr/bin/puppet kick *
MCollective

The proxy can trigger Puppet runs using the MCollective “puppet” agent. To enable this, add this line to settings.yml:

:puppet_provider: mcollective

And then add a sudoers rule:

Defaults:foreman-proxy !requiretty
foreman-proxy ALL = NOPASSWD: /usr/bin/mco puppet runonce *

Logging

The proxy’s output is captured to the log_file and may be filtered via the usual unix syslog levels:

  • WARN
  • DEBUG
  • ERROR
  • FATAL
  • INFO
  • UNKNOWN

See Ruby’s Logger class for details.

:log_file: /tmp/proxy.log
:log_level: DEBUG

4.3.3 ISC DHCP

ISC implementation is based on the omapi interface, which means:

  • No need for root permissions on your DHCP server
  • No need to restart (or “sync”) your dhcp server after every modifications.

Configuration

  • dhcpd configuration file: ensure you have the following line in your dhcpd.conf file (somewhere in the top first lines):
omapi-port 7911;
  • configure the settings file to point to your dhcpd.conf and dhcpd.leases files (make sure they are readable by the smart-proxy user)
  • make sure the omshell command (/usr/bin/omshell) can be executed by the smart-proxy user.
  • make sure that /etc/dhcp and /etc/dhcp/dhcpd.conf has group foreman-proxy

Securing the dhcp API

The dhcpd api server will listen to any host. You might need to add a omapi_key to provide basic security.

Example generating a key (on CentOS):

yum install bind97
dnssec-keygen -r /dev/urandom -a HMAC-MD5 -b 512 -n HOST omapi_key
cat Komapi_key.+*.private |grep ^Key|cut -d ' ' -f2-
  1. Edit your “/etc/dhcpd.conf”:

     omapi-port 7911;
     key omapi_key {
     algorithm HMAC-MD5;
     secret "XXXXXXXXX"; #<-The output from the generated key above.
     };
     omapi-key omapi_key;
     
  2. Make sure you also add the omapi_key to your proxy’s [[Smart-Proxy:Settingsyml#DHCP-section settings.yml]]
  3. Restart the dhcpd and foreman-proxy services

Testing

If everything works, you could browse your dhcp server data if you goto http://proxy:8443/dhcp

The next step is to set up appropriate Subnets in Foreman from the settings menu.

Sample dhcpd.conf

ddns-update-style interim;
ignore client-updates;
authoritative;
allow booting;
allow bootp;

omapi-port 7911;
#Optional key:
key omapi_key {
        algorithm HMAC-MD5;
        secret "2wgoV3yukKdKMkmOzOn/hIsM97QgLTT4CLVzg9Zv0sWOSe1yxPxArmr7a/xb5DOJTm5e/9zGgtzL9FKna0NWis==";
}
omapi-key omapi_key;

subnet 10.1.1.0 netmask 255.255.255.0 {
# --- default gateway
  option routers      10.1.1.254;
  option subnet-mask  255.255.255.0;

  option domain-name    "domain.com";
  option domain-name-servers  10.1.1.1, 8.8.8.8;
  option log-servers    syslog;
  option ntp-servers    ntp;

  range dynamic-bootp 10.1.1.10 10.1.1.250;
  default-lease-time 21600;
  max-lease-time 43200;

}

4.3.4 MS DHCP

The Microsoft smart-proxy installation procedure is very basic compared to the RPM or APT based solution.

It is required that this procedure is executed as an administrator.

It is not required that the smart-proxy be on the same host as the MS dhcp server. The smart-proxy just needs to be on a windows host that has netsh commands available.

  1. Go to the smart-proxy repository at https://github.com/theforeman/smart-proxy
  2. Select download and choose the latest revision
  3. Extract this to a directory that does not have any spaces in its name.
  4. Go to the rubyinstaller webpage at http://rubyinstaller.org/downloads/
  5. Download and install the “ruby 1.8.7 release 334”:http://rubyforge.org/frs/download.php/74293/rubyinstaller-1.8.7-p334.exe (Allow the ruby associations to be installed.)
  6. Open a CMD window and, using gem install –version X.X.X –platform ?????, add these gems
  columnize (0.3.2)
  highline (1.6.1)
  json (1.4.6 x86-mswin32)
  linecache (0.43 mswin32)
  mime-types (1.16)
  mocha (0.9.11)
  net-ping (1.3.7)
  rack (1.2.0)
  rake (0.8.7)
  rest-client (1.6.1)
  sinatra (1.1.0)
  tilt (1.1)
  win32-api (1.4.6 x86-mswin32-60)
  win32-open3 (0.3.2 x86-mswin32-60)
  win32-service (0.7.1 x86-mswin32-60)
  windows-api (0.4.0)
  windows-pr (1.1.2)

Command to download them all:

wget http://rubygems.org/downloads/columnize-0.3.2.gem \
     http://rubygems.org/downloads/haml-3.0.24.gem \
     http://rubygems.org/downloads/highline-1.6.1.gem \
     http://rubygems.org/downloads/json-1.4.6-x86-mswin32.gem \
     http://rubygems.org/downloads/linecache-0.43-mswin32.gem \
     http://rubygems.org/downloads/mime-types-1.16.gem \
     http://rubygems.org/downloads/mocha-0.9.11.gem \
     http://rubygems.org/downloads/net-ping-1.3.7.gem \
     http://rubygems.org/downloads/rack-1.2.0.gem \
     http://rubygems.org/downloads/rake-0.8.7.gem \
     http://rubygems.org/downloads/rest-client-1.6.1.gem \
     http://rubygems.org/downloads/sinatra-1.1.0.gem \
     http://rubygems.org/downloads/tilt-1.1.gem \
     http://rubygems.org/downloads/win32-api-1.4.6-x86-mswin32-60.gem \
     http://rubygems.org/downloads/win32-open3-0.3.2-x86-mswin32-60.gem \
     http://rubygems.org/downloads/win32-service-0.7.1-x86-mswin32-60.gem \
     http://rubygems.org/downloads/windows-api-0.4.0.gem \
     http://rubygems.org/downloads/windows-pr-1.1.2.gem

To get it to work on Windows 2008 R2 some of the packages has to change

 columnize (0.3.2)
 highline (1.6.1)
 json (1.4.6 x86-mingw32)
 linecache (0.43 mswin32)
 mime-types (1.16)
 mocha (0.9.11)
 net-ping (1.3.7)
 rack (1.2.0)
 rake (0.8.7)
 rest-client (1.6.1)
 sinatra (1.1.0)
 tilt (1.1)
 win32-api (1.4.6 x86-mingw32)
 win32-open3 (0.3.2 x86-mingw32)
 win32-service (0.7.1 x86-mswin32-60)
 windows-api (0.4.0)
 windows-pr (1.1.2)

Easy copy and paste method (platform may be different for you. Please check gem environment to find out.

  gem install --version 0.3.2 --platform x86-mingw32 columnize
  gem install --version 1.6.1 --platform x86-mingw32 highline
  gem install --version 1.4.6 --platform x86-mingw32 json
  gem install --version 0.43 --platform x86-mingw32 linecache
  gem install --version 1.16 --platform x86-mingw32 mime-types
  gem install --version 0.9.11 --platform x86-mingw32 mocha
  gem install --version 1.3.7 --platform x86-mingw32 net-ping
  gem install --version 1.2.0 --platform x86-mingw32 rack
  gem install --version 0.8.7 --platform x86-mingw32 rake
  gem install --version 1.6.1 --platform x86-mingw32 rest-client
  gem install --version 1.1.0 --platform x86-mingw32 sinatra
  gem install --version 1.1 --platform x86-mingw32 tilt
  gem install --version 1.4.6 --platform x86-mingw32 win32-api
  gem install --version 0.3.2 --platform x86-mingw32 win32-open3
  gem install --version 0.7.1 --platform x86-mingw32-60 win32-service
  gem install --version 0.4.0 --platform x86-mingw32 windows-api
  gem install --version 1.1.2 --platform x86-mingw32 windows-pr

Command to download them all:

wget http://rubygems.org/downloads/columnize-0.3.2.gem \
     http://rubygems.org/downloads/haml-3.0.24.gem \
     http://rubygems.org/downloads/highline-1.6.1.gem \
     http://rubygems.org/downloads/json-1.4.6-x86-mingw32.gem \
     http://rubygems.org/downloads/linecache-0.43-mswin32.gem \
     http://rubygems.org/downloads/mime-types-1.16.gem \
     http://rubygems.org/downloads/mocha-0.9.11.gem \
     http://rubygems.org/downloads/net-ping-1.3.7.gem \
     http://rubygems.org/downloads/rack-1.2.0.gem \
     http://rubygems.org/downloads/rake-0.8.7.gem \
     http://rubygems.org/downloads/rest-client-1.6.1.gem \
     http://rubygems.org/downloads/sinatra-1.1.0.gem \
     http://rubygems.org/downloads/tilt-1.1.gem \
     http://rubygems.org/downloads/win32-api-1.4.6-x86-mingw32.gem \
     http://rubygems.org/downloads/win32-open3-0.3.2-x86-mingw32.gem \
     http://rubygems.org/downloads/win32-service-0.7.1-x86-mswin32-60.gem \
     http://rubygems.org/downloads/windows-api-0.4.0.gem \
     http://rubygems.org/downloads/windows-pr-1.1.2.gem

8) CD to the root of the smart-proxy install directory

9) Edit config/settings.yml so that it looks a bit like this

Sample config/settings.yml file

 ---
 # HTTPS settings
 :ssl_certificate: c:\documents\smart-proxy\config\signed.pem
 :ssl_private_key: c:\documents\smart-proxy\config\private.pem
 :ssl_ca_file:     c:\documents\smart-proxy\config\ca.pem
 
 :trusted_hosts: [ foreman.someware.com]
 
 :daemon: false
 
 
 # Enable DHCP management
 :dhcp: true
 # The vendor can be either isc or native_ms
 :dhcp_vendor: native_ms
 # The dhcp_server is only used by the native_ms implementation
 :dhcp_server: 172.29.90.240
 
 # Where our proxy log files are stored
 # filename or STDOUT
 # Unix setting
 #:log_file: log/proxy.log
 # Windows setting
 :log_file: c:\tmp\proxy.log
 # valid options are
 # Logger::WARN, Logger::DEBUG, Logger::Error, Logger::Fatal, Logger:INFO, LOGGER::UNKNOWN
 #:log_level: Logger::DEBUG

10) Create the SSL key

10.1) Login to your puppetmaster

10.2) puppetca –generate Smart-proxy FQDN. (Do not use an alias here.)

10.3) Copy the private key, the public certificate and the ca.pem from /var/lib/puppet/ssl over to the locations that you specified in the setting file.

11) Test the installation by running ruby bin\smart-proxy.rb

12) Install the program as a service

12.1) ruby extra\register-service.rb

12.2) This may install the service but not run it. If so then try to start the service from the Ordinary Microsoft services snapin utility.

13) You may test connectivity by running the extra\query.rb utility from your foreman host. (Note that this file comes from the extra directory in the smart-proxy release.)

4.3.5 BIND

Bind configuration manipulation is based on nsupdate, which means that in theory could also be used to manipulate other dns servers which support nsupdate (such as Microsoft DNS server).

Configuration

In order to communicate securely with your dns server, you would need a key which will be used by nsupdate and your named daemon using ddns-confgen or dnssec-keygen

example using ddns-confgen

execute ‘ddns-confgen -k foreman -a hmac-md5’ - this should output something like the following:

# To activate this key, place the following in named.conf, and
# in a separate keyfile on the system or systems from which nsupdate
# will be run:
key "foreman" {
        algorithm hmac-md5;
        secret "GGd1oNCxaKsh8HA84sP1Ug==";
};

# Then, in the "zone" statement for each zone you wish to dynamically
# update, place an "update-policy" statement granting update permission
# to this key.  For example, the following statement grants this key
# permission to update any name within the zone:
update-policy {
        grant foreman zonesub ANY;
};

# After the keyfile has been placed, the following command will
# execute nsupdate using this key:
nsupdate -k /path/to/keyfile

You should create a new file (such as /etc/rndc.key or other) and store the key “foreman {…} in it. in the proxy Settings file you should point to this file location - make sure that the proxy have read permissions to this file.

In your named file, you could add the update-policy statement or something like this named example file if you need more fine grained permissions.

4.3.6 GSS-TSIG DNS

Both BIND as configured in FreeIPA and Microsoft AD DNS servers can accept DNS updates using GSS-TSIG authentication. This uses Kerberos principals to authenticate to the DNS server. Under Microsoft AD, this is known as “Secure Dynamic Update”.

Pre-requisites

  • Kerberos principal in the realm/domain that Smart Proxy can use
  • Kerberos keytab for the above principal

FreeIPA configuration

A service principal is required for the Smart Proxy, e.g. foremanproxy/proxy.example.com@EXAMPLE.COM.

First of all, create a new principal (FreeIPA service) for Foreman, e.g. ipa service-add foremanproxy/proxy.example.com@EXAMPLE.COM.

Then you can fetch the keytab, e.g. ipa-getkeytab -p foremanproxy/proxy.example.com@EXAMPLE.COM -s ipa-server.example.com -k /etc/foreman-proxy/dns.keytab.

Store the keytab at /etc/foreman-proxy/dns.keytab, ensure permissions are 0600 and the owner is foreman-proxy.

The ACL on updates to the DNS zone then needs to permit the service principal. In the FreeIPA web UI, under the DNS zone, go to the Settings tab and add to the BIND update policy a new grant:

grant foremanproxy\047ipa.example.com@EXAMPLE.COM wildcard * ANY;

Note the \047 is written verbatim, and don’t forget the semicolon.

Proxy configuration

Next, update the proxy configuration file (/etc/foreman-proxy/settings.yml) with the following settings:

:dns_provider: nsupdate_gss
:dns_tsig_keytab: /etc/foreman-proxy/dns.keytab
:dns_tsig_principal: foremanproxy/ipa.example.com@EXAMPLE.COM

4.3.7 SSL

The smart proxy can work in SSL mode, where both sides verify and trust each other. Requests from Foreman will only be accepted if the SSL certificate can be verified. Since proxies abstract a high level of control over your infrastructure, the configuration and security of keys and certificates is important.

Using Puppet CA certificates

Since Foreman integrates with Puppet heavily, it is recommended to use the Puppet Certificate Authority (CA) to secure proxy access. See the Security Communciations with SSL section for more advanced installations (multiple or internal CAs).

If the smart proxy host is not managed by Puppet, you will need to generate a certificate - skip forward to the generate section.

When using Puppet’s certificates, the following lines will be required in puppet.conf to relax permissions to the puppet group. The foreman and/or foreman-proxy users should then be added to the puppet group.

[main]
privatekeydir = $ssldir/private_keys { group = service }
hostprivkey = $privatekeydir/$certname.pem { mode = 640 }
Configuring the proxy

Configure the locations to the SSL files in /etc/foreman-proxy/settings.yml, plus the list of trusted Foreman hosts:

:ssl_certificate: /var/lib/puppet/ssl/certs/FQDN.pem
:ssl_ca_file: /var/lib/puppet/ssl/certs/ca.pem
:ssl_private_key: /var/lib/puppet/ssl/private_keys/FQDN.pem

:trusted_hosts:
- foreman.corp.com
#- foreman.dev.domain
Generating a certificate

To generate a certificate for a proxy host that isn’t managed by Puppet, do the following:

  1. Generate a new certificate on your puppetmaster: puppet cert --generate <proxy-FQDN>
  2. Copy the certificates and key from the puppetmaster to the smart proxy in /etc/foreman-proxy:<ol>
  • /var/lib/puppet/ssl/certs/ca.pem
  • /var/lib/puppet/ssl/certs/proxy-FQDN.pem
  • /var/lib/puppet/ssl/private_keys/proxy-FQDN.pem
  • </ol>

    Follow the configuration section above, however use the /etc/foreman-proxy paths instead of the Puppet defaults.

    Configuring Foreman

    For Foreman to connect to an SSL-enabled smart proxy, it needs configuring with SSL certificates in the same way. If the Foreman system is managed by Puppet, it will already have these, else certificates can be generated following the above instructions.

    The locations of the certificates are managed in the Settings page, under Provisioning with the ssl_ca_file, ssl_certificate and ssl_priv_key settings. By default these will point to the Puppet locations - for manually generated certificates, or non-standard locations, they may have to be changed.

    Lastly, when adding the smart proxy in Foreman, ensure the URL begins with https:// rather than http://.

    4.3.8 TFTP

    An essential first step in netbooting a system is preparing the TFTP server with the PXE configuration file and boot images. This document assumes that you have already configured your DHCP infrastructure, either via manual configuration or through the DHCP smart proxy.

    Configuration Values

    Once enabled, there is currently only one valid setting to change, the default TFTP root. This is set with the :tftproot: parameter, which defaults to /var/lib/tftpboot.

    Foreman tries to guess the right server name that should put into the dhcp record, if this is not what you want, you can override it - see tftp_servername under Settings.yml.

    Setting Up the Proxy Server Host

    Regardless of the filesystem setup is performed, you must also make sure you have the wget utility installed and in the default path. wget is used to download OS specific installation when a given host is enabled for the build process.

    Automatic Setup

    Foreman includes a TFTP server module that will perform all of the basic setup. It defaults to TFTP root of /var/lib/tftpboot, which may change if necessary. You will still need to provide the basic TFTP load images in your TFTP root directory. For vanilla PXE booting, this includes pxelinux.0, menu.c32, and chain.c32.

    Manual Setup

    The setup is very simple, and may be performed manually if desired.

    1. The TFTP root directory must exist (we will use /var/lib/tftpboot in this example).
    2. Populate /var/lib/tftpboot with PXE booting prerequisites. At a minimum, this should include:
      • pxelinux.0
      • menu.c32
      • chain.c32
    3. Create the directory /var/lib/tftpboot/boot and make it writeable by the foreman proxy user (foreman-proxy, for instance, when installing through a rpm package).
    4. Create the directory /var/lib/tftpboot/pxelinux.cfg and make it writeable by the foreman proxy user (foreman-proxy).

    Setting Up Foreman

    In most cases, the default templates should work fine. You do, however, need to make sure that a PXELinux or gPXE template is associated with your hosts. See [[Foreman:Unattended_installations Unattended Installations]] for details. The template will be used to define the PXE configuration file when a host is enabled for build.

    Workflow

    This is a rough outline of the steps triggered on the TFTP smart proxy host when you click on the “Build” link for a host.

    1. Call mkdir -p /var/lib/tftpboot/pxelinux.cfg if it does not already exist.
    2. Create a host-specific TFTP configuration file in /var/lib/tftpboot/pxelinux.cfg/01-XX-XX-XX-XX-XX-XX, named based off of the MAC address, using the associated PXE template.
    3. Call mkdir -p /var/lib/tftpboot/boot if it does not already exist.
    4. Download the OS specific kernel and initrd files using wget.
      1. The download URLs are derived from the installation media path, and OS specific log (see app/models/redhat.rb and debian.rb in foreman for examples of the gory details).
      2. The debian.rb file tries to guess if you want Ubuntu or Debian, based on the Name you give to your OS in the UI. If the name does not contain ‘ubuntu’ or ‘debian’, it may default to debian, hence fail to fetch the kernel/initrd.
    5. The exact wget command is wget --no-check-certificate -nv -c <src> -O "<destination>"
    6. At this point, the TFTP state is ready for the installation process.
    7. Once the host has completed installation, the OS specific installation script should inform foreman by retrieving the built URL.
    8. The host-specific TFTP configuration file is deleted.
    9. The kernel and initrd are not deleted, but left in place for future installs of the same OS and architecture combination. Please note that in the unlikely case that these files are modified, the simplistic freshness check of wget will likely get confused, corrupting the downloaded versions of the files. If this happens, you should simply delete the files and let them be re-downloaded from scratch.

    Limitations

    At the moment, the proxy is not able to fetch boot files using NFS. As a workaround, expose your installation medium (or use a public mirror) over http/ftp to configure one machine with the require boot files. this would be resolved as part of #992.

    4.4 Provisioning

    This chapter details the configuration of the required UI components necessary to provision an OS onto a host.

    4.4.1 Operating Systems

    The Operating Systems page (More -> Provisioning -> Operating Systems) details the OSs known to Foreman, and is the central point that the other required components tie into.

    Creating an Operating System

    Simply click New Operatingsystem on the main page. You will be taken to a screen where you can create the bare essentials of a new OS. Not everything required for a successful provision is on this page (yet) - the remaining components will appear for selection as we create them.

    You will need to fill in the first few parts of the form (Name, Major, Minor, Family, and possibly some family-dependent information). In the case of OSs like Fedora, it is fine to leave Minor blank.

    If the default Partition Tables & Installation media are suitable, then you can assign them now. If not, return here after each step in this chapter to assign the newly created objects to your Operating System

    Auto-created Operating Systems

    Foreman does not come with any Operating Systems by default. However, Foreman will detect the Operating System of any host which reports in via Puppet - if the OS of that Host is supported, it will be created (with very basic settings) and assigned to the Host. Thus you may find some OSs already created for you.

    4.4.2 Installation Media

    The Installation Media represents the web URL from where the installation packages can be retrieved (i.e. the OS mirror). Some OS Media is pre-created for you when Foreman is first installed. However, it is best to edit the media you are going to use and ensure the Family is set.

    New Installation Media

    If your OS of choice does not have a mirror pre-created for you, click the New Medium button to create one. There are a few variables which can be used to pad out the URL. For example:

    http://mirror.averse.net/centos/$major.$minor/os/$arch  
    

    Be sure to set the Family for the Media

    Assign to Operating System

    If you have not already done so, return to the Operating System page for your OS and assign the Media to it now.

    4.4.3 Provisioning Templates

    The Provision Templates is the core of Foreman’s flexibility to deploy the right OS with the right options. There are several types of template, along with a flexible matching system to deliver different templates to different Hosts.

    Foreman comes with pre-created templates for the more common OSs, but you will need to review these.

    Template Kinds

    There are 5 template kinds:

    • PXELinux - Deployed to the TFTP server to ensure the Host boots the correct installer with the correct kernel options
    • Provision - The main unattended installation file; e.g. Kickstart or Preseed
    • Finish - A post-install script used to take custom actions after the main provisioning is complete
    • Script - An arbitrary script, not used by default, useful for certain custom tasks
    • gPXE - USed in {g,i}PXE environments in place of PXELinux

    In practice, most environments only make use of the first 3. The Create Host action deploys the PXELinux template to the TFTP server. The PXELinux template directs the host to retrieve the Provision template. The Provision template will direct the installer to retrieve and run the Finish template at the end of the install, and the Finish template will notify Foreman the build is complete just before reboot.

    Editing Templates

    Clicking a template will take you to the edit page. All templates are ERB-enabled, and you can access many variables about the to-be-installed-host within the template. See Template Writing for ideas.

    As noted in 4.1.4 Auditing, changes to the templates are logged as diffs - you can browse the history of changes to the templates from the History tab within the Edit Template page. You can also revert changes here.

    Template Association

    When editing a Template, you must assign a list of Operating Systems which this Template can be used with. Optionally, you can restrict a template to a list of Hostgroups and/or Environments

    When a Host requests a template (e.g. during provisioning), Foreman will select the best match from the available templates of that type, in the following order:

    • Host-group and Environment
    • Host-group only
    • Environment only
    • Operating system default

    The final entry, Operating System default, can be set by editing the Operating System page.

    Associating an Operating System default template

    You will need to associate at least one PXELinux, Provision, and Finish template to your Operating System, and this must be done in two steps. First edit each of the templates, switch to the Association tab, and ensure the appropriate OSs are checked. Then edit the Operating System, switch to the Templates tab, and choose a default template for each template kind.

    4.4.4 Partition Tables

    Partition templates are a subset of normal Provisioning Templates. They are handled separately because it is frequently the case that an admin wants to deploy the same host template (packages, services, etc) with just a different harddisk layout to account for different servers’ capabilities.

    Foreman comes with pre-created templates for common Operating Systems, but it is good to edit the template, check it’s content and it’s Family setting. If the Family is wrong, be sure to go back to Operating Systems and associate it with your Operating System.

    Dynamic Partition tables

    When creating a new Host, you will be given the option to create a Dynamic Partition table. This is essentially a ‘one-off’ partition table that is stored with the host and used only for that host. It replaces the choice of Partition Table from the normal list of those associated with the selected OS.

    4.4.5 Architectures

    Architectures are simple objects, usually created by Foreman automatically when Hosts check in via Puppet. However, none are created by default, so you may need to create them if you’re not using Foreman for reporting.

    Creating a new Architecture

    Simply click New Architecture to create a new one. This should match the Facter fact :architecture e.g. “x86_64”. If you’ve already created some Operating Systems, you can associate the Architecture with the OS now; if not, the list of Architectures will be present when you create an OS.

    5. Advanced Foreman

    Missing content. Consider contributing, you kind soul!

    5.1 API

    Missing content. Consider contributing, you kind soul!

    5.2 Compute Resources

    Foreman supports creating and managing hosts on a number of virtualization and cloud services - referred to as “compute resources” - as well as bare metal hosts.

    The capabilities vary between implementations, depending on how the compute resource provider deploys new hosts and what features are available to manage currently running hosts. Some providers are able to support unattended installation using PXE, while others are image-based. Some providers have graphical consoles that Foreman interfaces to, and most have power management features. A summary of all providers and their support features is given below, and more detailed sections follow with specific notes.

    Provider Package Unattended installation Image-based Console Power management
    EC2 foreman-compute no yes read-only yes
    Libvirt foreman-libvirt yes no VNC or SPICE yes
    OpenStack Nova foreman-compute no yes no no
    oVirt / RHEV foreman-ovirt yes yes VNC or SPICE yes
    Rackspace foreman-compute no yes no yes
    VMware foreman-vmware yes no VNC yes

    Support for these features is aimed at being as transparent as possible, allowing the same configuration to be applied to hosts irrespective of the provider in use (compute resource or not). The selection of compute resource is made when creating a new host and the host in Foreman’s database remains associated to the VM that’s created, allowing it to be managed throughout the lifetime of the host.

    5.2.1 Using Compute Resources

    The following steps describe how to configure a compute resource and provision new hosts on it.

    1. Ensure the necessary package for the provider (from the above table) is installed, e.g. yum -y install foreman-ovirt. Restart the Foreman application to complete installation.

    2. Add a compute resource under More > Provisioning > Compute Resources > New Compute Resource. Select the provider type from the menu and appropriate configuration options will be displayed. Check the notes sections below for any provider-specific setup instructions.

    3. Click the Test Connection button after entering the configuration. If no error is displayed, the test was successful.

    4. After saving the compute resource, existing virtual machines can be browsed by clicking on the compute resource and the Virtual Machines tab.

    5. For providers that use images, click on the compute resource, then the Images tab, where known images are listed. To register images that Foreman can use, click New Image and enter the details.

    6. To provision a new host on this compute resource, from Hosts, click New Host and select the compute resource from the Deploy to menu.

    Also note the following features:

    1. When viewing a host, power management controls and the console access button are in the top right hand corner of the page.

    2. If a host provisioned on a compute resource is deleted, the VM and associated storage on the compute resource will also be deleted.

    3. Users in Foreman can have access restricted to hosts present on certain compute resources. For more information, see Filtering in 4.1.2 Roles and Permissions.

    5.2.2 EC2 Notes

    • Ensure AMIs are added under the Images tab on the compute resource with the correct username for it to SSH into the image.
    • Add a provisioning template of type finish which will be executed over SSH on the new image.
    • Ensure the finish template is associated to the OS (on the Associations tab) and is set as the default for the operating system too.
    • Enabling use_uuid_for_certificates in More > Settings is recommended for consistent Puppet certificate IDs instead of hostnames.
    • Currently doesn’t support VPC, see issue #1871.

    A full example for configuring EC2 provisioning is given on the Foreman blog: EC2 provisioning using Foreman.

    5.2.3 Libvirt Notes

    • Currently only supports KVM hypervisors.
    • VM consoles will be configured by default to listen on 0.0.0.0, change this via libvirt_default_console_address in More > Settings > Provisioning.
    • libvirt’s DNS and DHCP server (dnsmasq) can be disabled and replaced by BIND and ISC DHCPD (managed by Foreman) by creating a new virtual network and disabling DHCP support.

    To connect to the hypervisor using SSH:

    1. Configure SSH keys (ssh-keygen) for the ‘foreman’ user on the Foreman host to connect fully automatically to the remote hypervisor host.
    2. Change to the ‘foreman’ user, test the connection and ensure the remote host has been trusted.
    3. If connecting to the hypervisor as a non-root user, set up PolicyKit to permit access to libvirt. Note that different versions of PolicyKit have different configuration formats. 1, 2.
    4. Add the compute resource with a URL following one of these examples:
      • qemu+ssh://root@hypervisor.example.com/system to use the remote ‘root’ account
      • qemu+ssh://hypervisor.example.com/system to use the remote ‘foreman’ account

    To connect to the hypervisor over TCP without authentication or encryption (not recommended):

    1. Set the following options in libvirtd.conf:
      • listen_tls = 0
      • listen_tcp = 1
      • auth_tcp = "none"
    2. Enable libvirtd listening, e.g. set LIBVIRTD_ARGS="--listen" in /etc/sysconfig/libvirtd
    3. Add the compute resource with a URL following this example:
      • qemu+tcp://hypervisor.example.com:16509/system

    If you have difficulty connecting, test access using the virsh command under the ‘foreman’ account on the Foreman host first, e.g. virsh -c qemu+ssh://hypervisor.example.com/system list.

    5.2.4 OpenStack Notes

    • Supports OpenStack Nova for creating new compute instances.
    • A floating IP address group is required as Foreman will use the public IP address of the instance to complete provisioning over SSH (#2270).

    A full example for configuring image-based provisioning is given on the Foreman blog, also applicable to OpenStack: EC2 provisioning using Foreman.

    5.2.5 oVirt / RHEV Notes

    • SPICE consoles are displayed using an HTML5 client, so no native XPI extension is necessary.

    5.2.6 Rackspace Notes

    • The compute resource URL refers to the identity API URL, e.g. https://identity.api.rackspacecloud.com/v2.0

    A full example for configuring image-based provisioning is given on the Foreman blog, also applicable to Rackspace: EC2 provisioning using Foreman.

    5.2.7 VMware Notes

    • Only VMware clusters using vSphere are supported, not standalone ESX or ESXi servers.

    Consoles are provided using VNC connections from Foreman to the ESX server, which requires a firewall change to open the respective ports (TCP 5910 to 5930)

    ssh root@esx-srv
    vi /etc/vmware/firewall/vnc.xml

    Add the following file content:

    <ConfigRoot>
    <service id='0032'>
     <id>VNC</id>
     <rule id = '0000'>
      <direction>inbound</direction>
      <protocol>tcp</protocol>
      <porttype>dst</porttype>
      <port>
       <begin>5910</begin>
       <end>5930</end>
      </port>
     </rule>
     <enabled>true</enabled>
    </service>
    </ConfigRoot>

    Apply and check the firewall rule:

    esxcli network firewall refresh
    esxcli network firewall ruleset list | grep VNC

    Lastly, make the rule persistent:

    cp /etc/vmware/firewall/vnc.xml /vmfs/volumes/datastore1/vnc.xml
    vi /etc/rc.local
    # At end of file:
    cp /vmfs/volumes/datastore1/vnc.xml /etc/vmware/firewall/
    esxcli network firewall refresh

    5.3 Install Locations

    Missing content. Consider contributing, you kind soul!

    5.4 Securing Communications with SSL

    The Foreman web application needs to communicate securely with associated smart proxies and puppet masters, plus users and applications connecting to the web interface. This section details recommended SSL configurations.

    5.4.1 Securing Puppet Master Requests

    In a typical ENC-based setup with reporting, puppet masters require access to Foreman for three tasks:

    1. Retrieval of external nodes information (classes, parameters)
    2. Uploading of host facts
    3. Uploading of host reports

    All traffic here is initiated by the puppet master itself. Other traffic from Foreman to the puppet master for certificate signing etc. is handled via smart proxies (SSL configuration covered in the next section).

    Configuration options

    The Foreman interface authorizes access to puppet master interfaces based on its list of registered smart proxies with the Puppet feature, and identifies hosts using client SSL certificates.

    Five main settings control the authentication, the first are in Foreman under Settings, Auth:

    • require_ssl_puppetmasters (default: true), requires a client SSL certificate on the puppet master requests, and will verify the CN of the certificate against the smart proxies. If false, it uses the reverse DNS of the IP address making the request.
    • restrict_registered_puppetmasters (default: true), only permits access to hosts that have a registered smart proxy with the Puppet feature.
    • trusted_puppetmaster_hosts, a whitelist of hosts that overrides the check for a registered smart proxy

    And two in config/settings.yaml:

    • login (default: true), must be enabled to prevent anonymous access to Foreman.
    • require_ssl (default: false), should be enabled to require SSL for all communications, which in turn will require client SSL certificates if require_ssl_puppetmasters is also enabled. If false, host-based access controls will be available for HTTP requests.

    Enabling full SSL communications

    Using Apache HTTP with mod_ssl and mod_passenger is recommended. For simple setups, the Puppet certificate authority (CA) can be used, with Foreman and other hosts using certificates generated by puppet cert.

    1. Set Foreman’s require_ssl_puppetmasters, restrict_registered_puppetmasters and require_ssl to true.
    2. The mod_ssl configuration must contain:
    1. *SSLCACertificateFile* set to the Puppet CA
    2. *SSLVerifyClient optional*
    3. *SSLOptions +StdEnvVars*
    1. ENC script (e.g. /etc/puppet/node.rb) should have these settings:
    1. *:ssl_ca* set to the Puppet CA
    2. *:ssl_cert* set to the puppet master's certificate
    3. *:ssl_key* set to the puppet master's private key
    1. Report processor (e.g. /usr/lib/ruby/site_ruby/1.8/puppet/reports/foreman.rb) should have these settings:
    1. *$foreman_ssl_ca* set to the Puppet CA
    2. *$foreman_ssl_cert* set to the puppet master's certificate
    3. *$foreman_ssl_key* set to the puppet master's private key
    Troubleshooting

    Warning messages will be printed to Foreman’s log file (typically /var/log/foreman/production.log) when SSL-based authentication fails.

    • No SSL cert with CN supplied indicates no client SSL certificate was supplied, or the CN wasn’t present on a certificate. Check the client script has the certificate and key configured and that mod_ssl has SSLVerifyClient set.
    • SSL cert has not been verified indicates the client SSL certificate didn’t validate with the SSL terminator’s certificate authority. Check the client SSL certificate is signed by the CA set in mod_ssl’s SSLCACertificateFile and is still valid. More information might be in error logs.
    • SSL is required indicates the client is using an HTTP URL instead of HTTPS.
    • No smart proxy server found on $HOST indicates Foreman has no smart proxy registered for the source host, add it to the Smart Proxies page in Foreman. A common cause of this issue is the hostname in the URL doesn’t match the hostname seen here in the log file - change the registered proxy URL to match. If no smart proxy is available or can be installed, use trusted_puppetmaster_hosts and add this hostname to the whitelist.
    Advanced SSL notes

    A typical small setup will use a single Puppet CA and certificates it provides for the Foreman host and puppet master hosts. In larger setups with multiple CAs or an internal CA, this will require more careful configuration to ensure all hosts can trust each other.

    • Ensure the Common Name (CN) is present in certificates used by Foreman (as clients will validate it) and puppet master clients (used to verify against smart proxies).
    • Foreman’s SSL terminator must be able to validate puppet master client SSL certificates. In Apache with mod_ssl, the SSLCACertificateFile option must point to the CA used to validate clients and SSLVerifyClient set to optional.
    • Environment variables from the SSL terminator are used to get the client certificate’s DN and verification status. mod_ssl’s SSLOptions +StdEnvVars setting enables this. Variable names are defined by ssl_client_dn_env and ssl_client_verify_env settings in Foreman.

    Reduced security: HTTP host-based authentication

    In non-SSL setups, host-based authentication can be performed, so any connection from a host running a puppet smart proxy is able to access the interfaces.

    1. Set restrict_registered_puppetmasters to true.
    2. Set require_ssl_puppetmasters and require_ssl to false.

    No security: disable authentication

    Entirely disabling authentication isn’t recommended, since it can lead to security exploits through YAML import interfaces and expose sensitive host information, however it may be useful for troubleshooting.

    1. Set require_ssl_puppetmasters, restrict_registered_puppetmasters and require_ssl to false.

    5.4.2 Securing Smart Proxy Requests

    Foreman makes HTTP requests to smart proxies for a variety of orchestration tasks. In a production setup, these should use SSL certificates so the smart proxy can verify the identity of the Foreman host.

    In a simple setup, a single Puppet Certificate Authority (CA) can be used for authentication between Foreman and proxies. In more advanced setups with multiple CAs or an internal CA, the services can be configured as follows.

    Proxy configuration options

    /etc/foreman-proxy/settings.yml contains the locations to the SSL certificates and keys:

    ---
    # SSL Setup
    
    # if enabled, all communication would be verified via SSL
    # NOTE that both certificates need to be signed by the same CA in order for this to work
    # see http://theforeman.org/projects/smart-proxy/wiki/SSL for more information
    :ssl_certificate: /var/lib/puppet/ssl/certs/FQDN.pem
    :ssl_ca_file: /var/lib/puppet/ssl/certs/ca.pem
    :ssl_private_key: /var/lib/puppet/ssl/private_keys/FQDN.pem
    

    In this example, the proxy is sharing Puppet’s certificates, but it could equally use its own.

    In addition it contains a list of hosts that connections will be accepted from, which should be the host(s) running Foreman:

    # the hosts which the proxy accepts connections from
    # commenting the following lines would mean every verified SSL connection allowed
    :trusted_hosts:
    - foreman.corp.com
    #- foreman.dev.domain
    
    Configuring Foreman

    For Foreman to connect to an SSL-enabled smart proxy, it needs configuring with SSL certificates in the same way.

    The locations of the certificates are managed in the Settings page, under Provisioning - the ssl_ca_file, ssl_certificate and ssl_priv_key settings. By default these will point to the Puppet locations - for manually generated certificates, or non-standard locations, they may have to be changed.

    Lastly, when adding the smart proxy in Foreman, ensure the URL begins with https:// rather than http://.

    Sharing Puppet certificates

    If using Puppet’s certificates, the following lines will be required in puppet.conf to relax permissions to the puppet group. The foreman and/or foreman-proxy users should then be added to the puppet group.

    [main]
    privatekeydir = $ssldir/private_keys { group = service }
    hostprivkey = $privatekeydir/$certname.pem { mode = 640 }
    

    6. Troubleshooting

    There are two primary methods of getting support for the Foreman: IRC and mailing lists.

    IRC

    We work on the libera.chat servers. You can get general support in #theforeman, while development chat takes place in #theforeman-dev.

    Mailing lists

    Mailing lists are available via Google Groups. Much like IRC, we have a general users (support, Q/A, etc) lists and a development list:

    Foreman 3.12.0 has been released! Follow the quick start to install it.

    Foreman 3.11.4 has been released! Follow the quick start to install it.