| Forge/Official | Development |
|---|---|
 |
 |
The Pluribus Puppet module allows network administrators using Pluribus switches to easily configure and manage Pluribus and Netvisor specific resources. By implementing Puppet, the workflow for creating a new setup is streamlined, and tasks that previously may have taken days or weeks can be accomplished in a few hours. With Puppet, configurations can be re-used, or applied to many different setups.
If you are familiar with the Puppet master-agent setup, you may be weary to setup Puppet on a full size deployment, but with Pluribus' fabric capabilities, only one switch in your network needs to host a Puppet agent. From that one switch you can control the entire configuration of any switch on your fabric, without having to keep track of node groups and individual switches. Pluribus Puppet also supports hosting agents on all or some of the switches in your network, and can be configured to meet your specific needs as a system administrator.
If you are a Pluribus employee needing to quickly configure large networks for testing or demos, Puppet offers a quick and repeatable way to create network configurations. The nature of Puppet allows us to completely build, tear-down, and re-build the same setup over and over without worrying about missing a step or debugging a script.
If you are a developer looking to contribute, or a new development team member, please visit the developer page.
This module allows for the management of Pluribus switches through the Puppet DSL and Puppet manifest files by adding new Types and Providers to support Netvisor functionality. Users of this module can remotely configure and manage the setup and maintenance of Pluribus switches, and the added functionality makes the deployment of Pluribus switches in your data center quick and painless.
To use Pluribus Puppet you must first install and configure a Puppet master server and a Pupept agent on one, all, or some of the switches in your fabric. A guide for installing the Puppet Master and one for installing the Puppet Agent.
After installing the master and agent and getting them configured install the Pluribus Puppet module from Forge. Once installed, you will have access to the functionality provided by this module, including all of the types listed in the type catalog. Once installed, these types can be used in your Puppet manifests to configure Pluribus Network switches.
Puppet utilized manifest file declerations to manage resources. If you have used the Netvisor CLI before, puppet essentially automates the CLI commands and allows you to quickly manage resources.
For examples, creating vLANs on the cli is done with the following command:
CLI (...) > vlan-create id 101 scope fabric ports none
With Puppet, this same vLAN can be not only created, but managed with the following deceleration:
pn_vlan { '101':
ensure => present,
scope => 'fabric
}This may seem complicated and contain more typing than the traditional commands, especially taking into account compiling the Puppet catalog and pushing it to the agents, but the real power comes from the management capabilities and range namevars.
On the CLI, to delete the vLAN, you need to type the following:
CLI (...) > vlan-delete id 101
With Puppet, you change one line:
pn_vlan { '101':
- ensure => present,
+ ensure => absent,
scope => 'fabric
}The other powerful thing you can do with Puppet is using ranged namevars. If we want to create 5 vLANS from the CLI we must issue 5 separate commands like this:
CLI (...) > vlan-create id 101 scope fabric ports none
Vlan 101 created
CLI (...) > vlan-create id 102 scope fabric ports none
Vlan 102 created
CLI (...) > vlan-create id 103 scope fabric ports none
Vlan 103 created
CLI (...) > vlan-create id 104 scope fabric ports none
Vlan 104 created
CLI (...) > vlan-create id 105 scope fabric ports none
Vlan 105 created
With Puppet we can accomplish the same thing with the following manifest declaration:
pn_vlan { '101-105':
ensure => present,
scope => fabric,
}And instead of deleting one by one, we can once again change only one line:
pn_vlan { '101-105':
- ensure => present,
+ ensure => absent,
scope => 'fabric
}Hopefully this example has convinced you that Puppet is a powerful tool and can be used to streamline CLI setups. Keep in mind too that once you have written a manifest you can apply it to any number of switch setups and obtain the same end setup every time.
The following is a reference guide for using the various types provided by pn-puppet. These types are provided by the pn-puppet.
Allows for the management of clustered node. This type can create and destroy and manage node clusters.
name sets the cluster name. This is the type's namevar and is required. This can be any string as long as it only contains letters, numbers, _, ., :, and -.
ensure tells Puppet how to manage the cluster. Ensuring present will mean that the cluster will be created and on the switch after a completed catalog run. Setting this to absent will ensure that the cluster is not present on the system after the catalog run.
nodes specifies the two nodes which will be clustered together. This property takes an array of nodes that are present on the fabric. If the nodes are not present Puppet will throw an error and not complete the catalog run. If you pass more than two valid switches in the array only the first two will be used.
force_clustering forces a cluster between the specified nodes. It is not recommended you use this but an example use case would be if two switches were previously clustered incorrectly you could use force clustering to force the correct switches into the desired cluster. The default value for this property is no.
This example implementation creates a simple cluster between two switches, creatively named switch-01 and switch-02. Their cluster name, also creatively named, is switch-cluster.
The following shows both the traditional CLI implementation and below it the same command created in a Puppet manifest file.
CLI:
CLI (...) > cluster-create name switch-cluster cluster-node-1 switch-01 cluster-node-2 switch-02
Puppet:
node your-pluribus-switch {
pn_cluster { 'switch-cluster':
ensure => present,
nodes => ['switch-01', 'switch-02']
}
}Allows for the management of trunks or LAGs.
name sets the trunk name. This is the type's namevar and is required. This can be any string as long as it only contains letters, numbers, _, ., :, and -.
ensure tells Puppet how to manage the trunk. Ensuring present will mean that the trunk will be created and present on the switch after a completed catalog run. Setting this to absent will ensure that the trunk is not present on the system after the catalog run.
switch is the name of the switch where the link aggregation will occur. This should be a switch that is both on the same network as the Puppet agent and the same fabric. If Puppet cannot find the specified switch it will throw an error during the catalog run.
ports are the ports to be aggregated. This should be passed as a comma separated list, no whitespace, and port ranges are allowed.
The following example shows how to create trunks between two clusters. The first cluster is called spine-cluster and contains the two nodes spine-01 and spine-02. The second cluster is called leaf-cluster and contains the two nodes leaf-01 and leaf-02. spine-01 is connected to leaf-01 on ports 11 and 12, and connected to leaf-02 on 13 and 14. spine-02 is connected to leaf-01 on ports 15 and 16, and connected to leaf-02 on 17 and 18. The leaf to spine ports are the same numbers for the leaves.
CLI:
CLI (...) > switch spine-01 trunk-create name spine01-to-leaf ports 11,12,13,14
Created trunk spine01-to-leaf, id <id>
CLI (...) > switch spine-02 trunk-create name spine02-to-leaf ports 15,16,17,18
Created trunk spine02-to-leaf, id <id>
CLI (...) > switch leaf-01 trunk-create name leaf01-to-spine ports 11,12,15,16
Created trunk leaf01-to-spine, id <id>
CLI (...) > switch leaf-02 trunk-create name leaf02-to-spine ports 13,14,17,18
Created trunk leaf02-to-spine, id <id>
Puppet:
node your-pluribus-switch {
pn_lag { 'spine01-to-leaf':
ensure => present,
switch => 'spine-01',
ports => '11-14',
}
pn_lag { 'spine02-to-leaf':
ensure => present,
switch => 'spine-02',
ports => '15-18',
}
pn_lag { 'leaf01-to-spine':
ensure => present,
switch => 'leaf-01',
ports => '11,12,15,16',
}
pn_lag { 'leaf02-to-spine':
ensure => present,
switch => 'leaf-02',
ports => '13,14,17,18',
}
}Allow management of vLAGs. You must have LAGs/trunks in place on the switches in a vLAG prior to declaring the vLAG.
name sets the vLAG name. This is the type's namevar and is required. This can be any string as long as it only contains letters, numbers, _, ., :, and -.
ensure tells Puppet how to manage the vLAG. Ensuring present will mean that the vLAG will be created and present on the switch after a completed catalog run. Setting this to absent will ensure that the vLAG is not present on the system after the catalog run.
cluster tells Puppet which cluster the vLAG should be applied to.
port is the vLAG port on switch.
peer-port is the vLAG port on peer-switch.
mode the vLAG mode. Can either be set to active or standby, corresponding to active-active and active-standby vLAG modes respectively. This property defaults to active.
failover is how L2 failover will be handled by the vLAG. This can either be specified as move or ignore. The default value for this property is move.
lacp_mode controls the link aggregation control protocol mode. This can be either active, passive or off. The default value is active.
lacp_timeout sets the type of LACP timeout. This can be set to either fast or slow. The default setting is fast.
lacp_fallback sets the fallback type of the LACP connection. This can be set to either bundle or individual. By default, this value is set to bundle.
lacp_fallback_timeout sets the fallback timeout in seconds. This can be any integer between 30 and 60. By default the fallback timeout is set to 50 seconds.
The following example shows how to create trunks between two clusters. The first cluster is called spine-cluster and contains the two nodes spine-01 and spine-02. The second cluster is called leaf-cluster and contains the two nodes leaf-01 and leaf-02. spine-01 is connected to leaf-01 on ports 11 and 12, and connected to leaf-02 on 13 and 14. spine-02 is connected to leaf-01 on ports 15 and 16, and connected to leaf-02 on 17 and 18. The leaf to spine ports are the same numbers for the leaves.
CLI:
CLI (...) > cluster-create name spine-cluster ...
CLI (...) > trunk-create name spine01-to-leaf ...
Created trunk spine02-to-leaf, id <#>
CLI (...) > trunk-create name spine02-to-leaf ...
Created trunk spine02-to-leaf, id <#>
CLI (...) > switch spine-01 vlag-create name spine-to-leaf port spine01-to-leaf peer-switch spine02 peer-port leaf2-to-spine mode active-active failover-ignore-L2 lacp-mode slow lacp-fallback bundle lacp-fallback-timeout 45
Puppet:
pn_cluster { 'spine-cluster':
...
}
pn_lag { 'spine01-to-leaf':
...
require => Pn_cluster['spine-cluster']
}
pn_lag { 'spine02-to-leaf':
...
require => Pn_cluster['spine-cluster']
}
pn_vlag { 'spine-to-leafs':
ensure => present,
cluster => 'spine-cluster',
port => 'spine01-to-leaf',
peer-port => 'spine02-to-leaf',
mode => active,
failover => ignore,
lacp_mode => active,
lacp_timeout => slow,
lacp_fallback => bundle,
lacp_fallback_timeout => 45,
require => Pn_lag['spine01-to-leaf',
'spine02-to-leaf'],
}Manage vLANs.
id is the vLAN id, this can be any number between 2 and 4092. Comma separated or whitespace separated is allowed. Ranges are allowed.
ensure tells Puppet how to manage the vLAN. Ensuring present will mean that the vLAN will be created and present on the switch after a completed catalog run. Setting this to absent will ensure that the vLAN is not present on the system after the catalog run.
scope is the name of the vNET assigned to the vRouter.
description is the description of the vLAN. Can only contain letters, numbers, _, ., :, and -. The default value is '-'.
stats enables or disables vLAN statistics. This can either be enable or disable. The default value is enable.
ports is a comma separated list of ports that the vLAN will use. There cannot be any whitespace separating the ports, ranges are allowed. The default value is 'none'
untagged_ports is a comma separated list of untagged ports that the vLAN will use. There cannot be any whitespace separating the ports, ranges are allowed. The default value is 'none'
switch the switch where the vRouter will live, this can be the name of any switch on the fabric. By deafult this value is set to local and creates a vRouter on whatever node is specified in the manifest.
CLI:
CLI (...) > vlan-create id 101 scope fabric description puppet-vlan ports none untagged-ports none
Puppet:
pn_vlan { '101':
ensure => present,
scope => fabric,
description => 'puppet-vlan',
ports => 'none',
untagged_ports => 'none',
}Manage vRouters. On systems that only allow one vRouter the latest executed vRouter deceleration will be created.
name is the name of the vRouter to be managed. Name can be any string as long as it only contains letters, numbers, _, ., :, and -.
ensure tells Puppet how to manage the vRouter. Ensuring present will mean that the vRouter will be created and present on the switch after a completed catalog run. Setting this to absent will ensure that the vRouter is not present on the system after the catalog run.
vnet is the name of the vNET assigned to the vRouter.
hw_vrrp_id is a hardware id for VRRP interfaces that may live on this vRouter.
service simply enables or diables the vRouter. This can be set to either enable or disable. By default this is set to enable.
bgp_as is the AS number for any BGP interfaces that you will create later. Can be any integer. By default this property is set to '' and tells Puppet not to set up BGP on the vRouter. (This can always be changed in the manifest later.)
router_id is the IP address assigned to the vRouter, both router_id and bgp_as must be specified to create a vRouter that can host a BGP interface.
switch the switch where the vRouter will live, this can be the name of any switch on the fabric. By deafult this value is set to local and creates a vRouter on whatever node is specified in the manifest.
CLI:
CLI (...) > vrouter-create name demo-vrouter vnet demo-vnet-global hw-vrrp-id 18 enable
Puppet:
pn_vrouter { 'demo-vrouter':
ensure => present,
vnet => 'demo-vnet-global',
hw_vrrp_id => 18,
service => enable,
}Manage vRouter BGP interfaces. To create a BGP interface you must first create a pn_vrouter_ip so that the BGP interface has an established ip interface to live on.
name is a combination of the vRouter name and the BGP neighbor IP address, separated by a space.
ensure tells Puppet how to manage the BGP interface. Ensuring present will mean that the BGP interface will be created and present on the switch after a completed catalog run. Setting this to absent will ensure that the BGP interface is not present on the system after the catalog run.
bgp_as is the AS ID for the BGP interface.
increment is how much the address will be incremented by in a range.
switch is the name of the switch where the vRouter BGP interface will be hosted. This can be any switch on the fabric. The default value is local which creates a BGP interface on the node where the resource was declared.
CLI:
CLI (...) > vrouter-create name demo-vrouter vnet demo-vnet-global hw-vrrp-id 18 enable bgp-as 65001
CLI (...) > vlan-create id 101 scope fabric
CLI (...) > vrouter-interface-add vrouter-name demo-vrouter ip 101.101.101.2/24 vlan 101 if data
CLI (...) > vrouter-bgp-add vrouter-name demo-vrouter neighbor 101.101.101.1 remote_as 65001 bfd
Puppet:
pn_vrouter { 'demo-vrouter':
ensure => present,
vnet => 'demo-vnet-global',
hw-vrrp-id => 18,
service => enable,
bgp_as => '65001',
router_id => '172.168.85.8',
}
pn_vlan { '101':
require => Pn_vrouter['demo-vrouter'],
ensure => present,
scope => 'fabric',
description => 'bgp',
}
pn_vrouter_ip { '101':
require => Pn_vlan['101'],
ensure => present,
vrouter => 'demo-vrouter',
ip => 'x.x.x.2',
mask => '24',
}
pn_vrouter_bgp { 'demo-vrouter 101.101.101.1':
require => Pn_vrouter_ip['101'],
ensure => present,
bgp_as => '65001',
}
pn_vrouter_bgp { 'demo-vrouter 101.101.101.2-10':
require => Pn_vrouter_ip['101'],
ensure => present,
bgp_as => '65001',
increment => '2',
}Manage vRouter IP interfaces and vRouter VRRP interfaces. If you are creating a VRRP interface you must specify both vrrp_ip and vrrp_priority, otherwise and IP interface will be created. When you create a VRRP interface, pn_vrouter_if creates an IP interface AND a VRRP interface in one resource deceleration. If you don't create a vRouter prior to creating a vRouter interface the interface will make one. If you create a vRouter after a vRouter interface with a diffrent name the vRouter interface will be removed. Be careful and use Before and Require statements.
name is the id of the vLan that the vRouter interface will live on. The name consists of a comma or whitespace seperated list of vLANs, followed by an IP pattern including netmask. The IP follows pattern matching.
ensure tells Puppet how to manage the vRouter interface. Ensuring present will mean that the vRouter interface will be created and present on the switch after a completed catalog run. Setting this to absent will ensure that the vRouter interface is not present on the system after the catalog run.
vrrp_ip is the ip of the VRRP interface. This also obeys IP pattern matching, and the only criteria is that this ip cannot be the same as the IP of the IP interface. Default is none.
vrrp_priority The VRRP interface priority, this can be a number between 0 and 255. Default is none.
switch is the name of the switch where the IP interface will be created. This can be any switch on the fabric. The default value is local, which creates an IP interface on the node where the resource was declared.
CLI:
CLI (...) > vrouter-create name demo-vrouter vnet demo-vnet-global hw-vrrp-id 18 enable
CLI (...) > vlan-create id 101 scope fabric
CLI (...) > vrouter-interface-add vrouter-name demo-vrouter ip 101.101.101.2/24 vlan 101 if data
CLI (...) > vrouter-interface-add vrouter-name demo-vrouter ip 101.101.101.1/24 vlan 101 if data vrrp-id 18 vrrp-primary eth0.101 vrrp-priority 110
CLI (...) > vlan-create id 102 scope fabric
CLI (...) > vrouter-interface-add vrouter-name demo-vrouter ip 102.102.102.2/24 vlan 101 if data
CLI (...) > vrouter-interface-add vrouter-name demo-vrouter ip 102.102.102.1/24 vlan 101 if data vrrp-id 18 vrrp-primary eth0.101 vrrp-priority 110
Puppet:
pn_vrouter { 'demo-vrouter':
ensure => present,
vnet => 'demo-vnet-global',
hw-vrrp-id => 18,
service => enable,
}
pn_vlan { '101-102':
require => Pn_vrouter['demo-vrouter'],
ensure => present,
scope => 'fabric',
}
pn_vrouter_if { '101-102 x.x.x.2/24':
require => Pn_vlan['101'],
ensure => present,
vrrp_ip => 'x.x.x.1',
vrrp_priority => '110',
}Creates a vRouter loopback interface on the destination switch.
name is a combination of the vRouter name and the loopback IP address, separated by a space.
ensure tells Puppet how to manage the loopback interface. Ensuring present will mean that the loopback interface will be created and present on the switch after a completed catalog run. Setting this to absent will ensure that the loopback interface is not present on the system after the catalog run.
switch is the name of the switch where the IP interface will be created. This can be any switch on the fabric. The default value is local, which creates an IP interface on the node where the resource was declared.
CLI:
CLI (...) > vrouter-loopback-interface-add vrouter-name spine1vrouter ip 172.16.1.1
Puppet:
pn_vrouter_loopback { 'spine1vrouter 172.16.1.1':
ensure => present,
}- LAG
- vLAG
- VLAN
- Port Configuration
- vRouters
- VRRP
- BGP
- OSPF
- Pluribus CLI as a parameter (allows to configure any new feature)
Pluribus Puppet currently only runs on ONVL distributions of Netvisor. This is because the Puppet does not support all distributions of Solaris and cannot be run on the nvOS distribution. This problem is being looked into and hopefully nvOS will be supported alongside ONVL in the near future.
There are currently no additional resources for the Pluribus Puppet module.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent