Puppet device

With Puppet device, you can manage network devices, such as routers, switches, firewalls, and Internet of Things (IOT) devices, without installing a Puppet agent on them. Devices that cannot run Puppet applications require a Puppet agent to act as a proxy. The proxy manages certificates, collects facts, retrieves and applies catalogs, and stores reports on behalf of a device.

Puppet device runs on both *nix and Windows. The Puppet device application combines some of the functionality of the Puppet apply and Puppet resource applications. For details about running the Puppet device application, see the puppet device man page.

Note: If you are writing a module for a remote resource, we recommend using transports instead of devices. Transports have extended functionality and can be used with other workflows, such as with Bolt . For more information on transports and how to port your existing code, see Resource API Transports.

The Puppet device model

In a typical deployment model, a Puppet agent is installed on each system managed by Puppet. However, not all systems can have agents installed on them.

For these devices, you can configure a Puppet agent on another system which connects to the API or CLI of the device, and acts as a proxy between the device and the primaryPuppet server.

In the diagram below, Puppet device is on a proxy Puppet agent (agent.example.com) and is being used to manage an F5 load balancer (f5.example.com) and a Cisco switch (cisco.example.com).

Puppet device’s run environment

Puppet device runs as a single process in the foreground that manages devices, rather than as a daemon or service like a Puppet agent.


The puppet device command runs with the privileges of the user who runs it.

Run Puppet device as:

  • Root on *nix

  • Either LocalService or a member of the Administrators group on Windows


By default, Puppet device outputs directly to the terminal, which is valuable for interactive use. When you run it as a cron job or scheduled task, use the logdest option to direct the output to a file.

On *nix, run Puppet device with the --logdest syslog option to log to the *nix syslog service:
puppet device --verbose --logdest syslog
Your syslog configuration determines where these messages are saved, but the default location is /var/log/messages on Linux, and /var/log/system.log on Mac OS X. For example, to view these logs on Linux, run:
tail /var/log/messages
On Windows, run Puppet device with the --logdest eventlog option, which logs to the Windows Event Log, for example:
puppet device --verbose --logdest eventlog

To view these logs on Windows, click Control PanelSystem and SecurityAdministrative ToolsEvent Viewer.

To specify a particular file to send Puppet device log messages to, use the --logdest <FILE> option, which logs to the file specified by <FILE>, for example:
puppet device --verbose --logdest /var/log/puppetlabs/puppet/device.log

You can increase the logging level with the --debug and --verbose options.

In addition to local logging, Puppet device submits reports to the primary Puppet server after each run. These reports contain standard data from the Puppet run, including any corrective changes.

Network access

Puppet device creates outbound network connections to the devices it manages. It requires network connectivity to the devices via their API or CLI. It never accepts inbound network connections.

Installing device modules

You need to install the device module for each device you want to manage on the primary Puppet server.

For example, to install the f5 and cisco_ios device modules on the primary server, run the following commands:

$ sudo puppet module install f5-f5
$ sudo puppet module install puppetlabs-cisco_ios

Configuring Puppet device on the proxy Puppet agent

You can specify multiple devices in device.conf, which is configurable with the deviceconfig setting on the proxy agent.

For example, to configure an F5 and a Cisco IOS device, add the following lines to the device.conf file:

type f5
url https://username:password@f5.example.com

type cisco_ios
url file:///etc/puppetlabs/puppet/devices/cisco.example.com.yaml
The string in the square brackets is the device’s certificate name — usually the hostname or FQDN. The certificate name is how Puppet identifies the device.

For the url, specify the device’s connection string. The connection string varies by device module. In the first example above, the F5 device connection credentials are included in the url device.conf file, because that is how the F5 module stores credentials. However, the Cisco IOS module uses the Puppet Resource API, which stores that information in a separate credentials file. So, Cisco IOS devices would also have a /etc/puppetlabs/puppet/devices/<device cert name>.conf file similar to the following content:

"address": "cisco.example.com"
"port": 22
"username": "username"
"password": "password"
"enable_password": "password"

For more information, see device.conf.

Classify the proxy Puppet agent for the device

Some device modules require the proxy Puppet agent to be classified with the base class of the device module to install or configure resources required by the module. Refer to the specific device module README for details.

To classify proxy Puppet agent:

  1. Classify the agent with the base class of the device module, for each device it manages in the manifest. For example:
    node 'agent.example.com' {
      include cisco_ios
      include f5
  2. Apply the classification by running puppet agent -t on the proxy Puppet agent.

Classify the device

Classify the device with resources to manage its configuration.

The examples below manage DNS settings on an F5 and a Cisco IOS device.

  1. In the site.pp manifest, declare DNS resources for the devices. For example:
    node 'f5.example.com' {
     f5_dns{ '/Common/dns':
      name_servers => ['', '"],
      same     => ['localhost",' example.com'],
    node 'cisco.example.com' {
     network_dns { 'default':
      servers => [', ''],
      search => ['localhost",'example.com'],
  2. Apply the manifest by running puppet device -v on the proxy Puppet agent.
Note: Resources vary by device module. Refer to the specific device module README for details.

Get and set data using Puppet device

The traditional Puppet apply and Puppet resource applications cannot target device resources: running puppet resource --target <DEVICE> does not return data from the target device. Instead, use Puppet device to get data from devices, and to set data on devices. The following are optional parameters.

Get device data with the resource parameter

puppet device --resource <RESOURCE> --target <DEVICE>
Use the resource parameter to retrieve resources from the target device. For example, to return the DNS values for example F5 and Cisco IOS devices:
sudo puppet device --resource f5_dns --target f5.example.com
sudo puppet device --resource network_dns --target cisco.example.com

Set device data with the apply parameter

puppet device --verbose --apply <FILE> --target <DEVICE>
Use the --apply parameter to set a local manifest to manage resources on a remote device. For example, to apply a Puppet manifest to the F5 and Cisco devices:
sudo puppet device --verbose --apply manifest.pp --target f5.example.com
sudo puppet device --verbose --apply manifest.pp --target cisco.example.com

View device facts with the facts parameter

puppet device --verbose --facts --target <DEVICE>
Use the --facts parameter to display the facts of a remote target. For example, to display facts on a device:
sudo puppet device --verbose --facts --target f5.example.com

Managing devices using Puppet device

Running the puppet device or puppet-device command (without --resource or --apply options) tells the proxy agent to retrieve catalogs from the primary server and apply them to the remote devices listed in the device.conf file.

To run Puppet device on demand and for all of the devices in device.conf , run:

sudo puppet device --verbose

To run Puppet device for only one of the multiple devices in the device.conf file, specify a --target option:

$ sudo puppet device -verbose --target f5.example.com
To run Puppet device on a specific group of devices, as opposed to all devices in the device.conf file, create a separate configuration file containing the devices you want to manage, and specify the file with the --deviceconfig option:
$ sudo puppet device --verbose --deviceconfig /path/to/custom-device.conf

To set up a cron job to run Puppet device on a recurring schedule, run:

$ sudo puppet resource cron puppet-device ensure=present user=root minute=30 command='/opt/puppetlabs/bin/puppet device --verbose --logdest syslog'


Follow the steps below to run Puppet device in a production environment, using cisco_ios as an example.

  1. Install the module on the primary Puppet server: sudo puppet module install puppetlabs-cisco_ios.

  2. Include the module on the proxy Puppet agent by adding the following line to the primary server’s site.pp file:
    include cisco_ios
  3. Edit device.conf on the proxy Puppet agent:
    type cisco_ios
    url file:///etc/puppetlabs/puppet/devices/cisco.example.com.yaml
  4. Create the cisco.example.com credentials file required by modules that use the Puppet Resource API:
      "address": "cisco.example.com"
      "port": 22
      "username": "username"
      "password": "password"
      "enable_password": "password"
  5. Request a certificate on the proxy Puppet agent: sudo puppet device --verbose --waitforcert 0 --target cisco.example.com

  6. Sign the certificate on the primary server: sudo puppetserver ca sign cisco.example.com

  7. Run puppet device on the proxy Puppet agent to test the credentials: sudo puppet device --target cisco.example.com

Automating device management using the puppetlabs device_manager module

The puppetlabs-device_manager module manages the configuration files used by the Puppet device application, applies the base class of configured device modules, and provides additional resources for scheduling and orchestrating Puppet device runs on proxy Puppet agents.

For more information, see the module README.

Troubleshooting Puppet device

These options are useful for troubleshooting Puppet device command results.

--debug or -d Enables debugging
--trace or -t Enables stack tracing if Ruby fails
--verbose or -v Enables detailed reporting