Installation overview
This chapter describes the Fabric Services System components, the requirements for these components, and provides an overview of the installation process.
Supported deployment models
The Fabric Services System is deployed on one, three, or six Fabric Services System nodes. The Fabric Services System is deployed as an application on virtual machine servers. In a deployment with six nodes, the Fabric Services System can be deployed with the Digital Sandbox.
Networking for the Fabric Services System nodes
- Fabric Services System OAM interface
This interface is used to access the UI and the API of the Fabric Services System. This is also the network through which the deployer VM reaches the node.
- Fabric management interface
This interface is used to communicate with the management interfaces of the fabric (for example, SR Linux devices). This interface is where the Fabric Services System exposes its DHCP and ZTP services.
DHCP and ZTP
- all the management interfaces of the fabric devices are connected to the same Layer 2 network as the DHCP interface of the Fabric Services System
- as the DHCP relay
Fabric Services System deployer
The Fabric Services System deployer is a virtual machine (VM) that is used to provision a Kubernetes environment and install Fabric Services System microservices on the Fabric Services System nodes.
It contains all the software packages, container images, and Helm charts that are used by the deployer script to install the entire Fabric Services System environment.
The Fabric Services System deployer VM:
- installs a Kubernetes cluster on the Fabric Services System nodes in your environment
- configures the nodes on the cluster for local storage
- labels the nodes to prepare them for Fabric Services System software deployment
- installs Fabric Services System microservices using Helm charts and container repositories hosted in the deployer
- provides technical support tools and backup and restore tools
- initializes the self-signed certificates used by the system used for the northbound services
- initializes the self-signed certificates use with managed nodes
Fabric Services System nodes
The Fabric Services System nodes are the VMware vSphere-based or KVM-based VMs that host the Kubernetes environment on which the Fabric Services System application and Digital Sandbox are run.
The following deployment models are supported:
- an environment with one node, which hosts only the Fabric Services System application for very small scale deployments and does not support the use of Digital Sandbox
- an environment with three nodes, which hosts only the Fabric Services System application and does not support the use of Digital Sandbox
- an environment with a minimum of six nodes, which hosts both the Fabric Services System application and supports the use of Digital Sandbox
Requirements
This section details the specific requirements for the components and their connectivity.
For geo-redundant systems, see Configuring geo-redundancy
in the Fabric
Services System User Guide.
Fabric Services System deployer requirements
- QCOW2: the image that provides the deployer VM and which includes the operating system and the deployer software for the KVM hypervisor
-
OVA: the VM image for the VMware vSphere 7.0 or 8.0 hypervisor
- CPU: 1 virtual CPU (vCPU) on a VMware vSphere or KVM hypervisor with a modern
x86-64 CPU that supports virtualization
Recommended: 2 vCPU
- memory: 8 GB RAM
- storage: 50 GB of storage (used by the deployer QCOW2 image)
- networking:
- 1 GbE network interface card (NIC)
- the configured DNS servers must be reachable, functional, and able to resolve the hostnames used for the Fabric Services System nodes
Fabric Services System node requirements
The Fabric Services System nodes are deployed as virtual machine servers. The following are the requirements of the Fabric Services System nodes in KVM and VMware hypervisor:
-
CPU: 32 vCPUs on a modern x86-64 CPU
- memory: 64 GB
- storage: at least 500 GB of available SSD-based storage
- networking:
- at least one 10 Gbps NIC
- the configured DNS servers must be reachable, functional, and able to resolve the hostnames used for the Fabric Services System nodes
-
operating system: VMware vSphere 7.0 or 8.0 or RHEL/Rocky 8.9
- hypervisor: ESXi 7.0 or 8.0 or KVM
- resource reservation for CPU, memory, and disks must be set to 100% for the Fabric Services System node virtual machines
Geo-redundancy
To bring up geo-redundant setup, ensure that the setup is similar in the active and
backup sites and that network connectivity is stable in both sites. For additional
requirements, see Geo-redundancy configuration
in the Fabric Services
System User Guide.
Installation process overview
The installation consists of the following high-level tasks:
- Downloading the Fabric Services System deployer image
This task describes how to access the Fabric Services System deployer image so it can be used in the next task.
- Fabric Services System deployer VM creation
This task describes how to create the VM in a VMware vSphere or KVM environment and how to configure the Fabric Services System deployer VM.
- Preparing the Fabric Services System virtual machine nodes
This task describes how to prepare and configure the nodes for deployment.
- Editing the installation configuration file
This task describes how to modify the variables and enter the expected data in the JSON-based configuration file, which is used to specify the configuration and deployment of the Fabric Services System environment.
- Installing the Fabric Services System environment
This task describes the deployment of the Fabric Services System environment using the sample-input.json file.