AOS Cluster

Demand for compute resources can grow beyond the capacity of a single virtual machine (VM). Multiple VMs can be clustered, specifically for scaling off-box agents and IBA probe processing units (as of version 3.0).

When a worker VM is added to the main Apstra controller VM, it registers with the Apstra server VM through sysdb, collects facts about the VM (such as core/memory/disk configuration and usage), and launches a container on the local VM. The Apstra controller VM reacts to REST API requests, configures the worker VM for joining or leaving the cluster, and keeps track of cluster-wide runtime information. It also reacts to container configuration entities and schedules them to the worker VM.

REST API and Clusters

In addition to managing Apstra server VMs using the web interface (as described below) you can also use REST API. Navigate to Platform > Developers for REST API Documentation and tools. See the cluster section.

Apstra server VMs include the following details:

Static Configuration
Address
IP address or Fully-qualified Domain Name (FQDN) of the VM
Name
Apstra VM name, such as controller (the main Apstra controller) or node1 (a worker node)
State
ACTIVE, MISSING, or FAILED
Roles
Controller or worker
Tags
The function of the node: iba and/or offbox
Capacity Score
Calculated by Apstra
CPU
Number of CPUs
Errors

If an error exists, a message appears here. For example, the following image shows the error message when an agent process has been restarted because an agent has crashed.

_images/aos_cluster_error_1.png
Usage
Container Service Usage
Current VM container server usage (Percentage)
Containers Count
Number of containers
Memory Usage
Current VM memory usage (percentage)
CPU Usage
Current VM CPU usage (percentage)
Disk Usage
Current VM disk usage per logical volume (GB and percentage)
Containers
The containers that are running on the node and the resources used by each container
Username
Web interface username login credential, such as admin
Password
Web interface password login credential

From the web interface, navigate to Platform > AOS Cluster. You can health check the VMs from here. To see more information, click an address.

_images/aos_cluster_330.png

Cloning Apstra VM

Instead of entering all details for a new VM, you can clone an existing one, give it a new name and customize it.

Creating Apstra VM

  1. Install Apstra software on the VMs to be clustered, making sure they are all the same version as the main controller (which acts as the cluster manager). If they are not the same version, the controller will not accept them as part of the cluster.
  2. From the list view (Platform > AOS Cluster) click Add Node, then enter a name, tags (optional), address (IP or FQDN), username and password.
  3. Click Create. As the main Apstra controller connects to the new Apstra VM worker node, the state of the new Apstra VM changes from INIT to ACTIVE.

Editing Apstra VM

  1. Either from the list view (Platform > AOS Cluster) or the details view, click the Edit button for the VM to edit.
  2. Make your changes.
  3. Click Update to update the Apstra VM worker node.

Deleting Apstra VM

  1. Either from the list view (Platform > AOS Cluster) or the details view, click the Delete button for the Apstra VM to delete.
  2. Click Delete to delete the Apstra VM.