Fortinet black logo

Administration Guide

HA-Cluster

Copy Link
Copy Doc ID 8db33d2a-14ec-11ed-9eba-fa163e15d75b:735542
Download PDF

HA-Cluster

A single FortiSandbox device can scan a limited number of files in a given time period. To handle heavier loads, you can use multiple FortiSandbox devices in a load-balancing high availability (HA) cluster.

There are three types of nodes in a cluster: primary, secondary, and worker.

Primary

The primary node (Unit 1 in the diagram) manages the cluster, distributes jobs and gathers the results, and interacts with clients. It can also perform normal file scans. All scan-related configuration should be done on the primary node and they will be broadcasted from the primary node to the other nodes. Any scan-related configuration that has been set on a worker node will be overwritten.

On the primary node, users can:

  • Change a worker node's role (secondary and worker)
  • Configure a worker node's network settings
  • Upgrade worker nodes
  • View VM status page of worker nodes
  • Configure FortiGuard settings of worker nodes
  • Configure VM images of worker nodes, such as setting clone numbers of each VM image
  • Configure a ping server to frequently check unit's network condition and downgrade itself as a secondary node when necessary to trigger a failover

Although all FortiSandbox models can work as a primary node, we recommend using a more powerful model.

When the primary and secondary nodes are using a FortiSandbox VM model, you have the option of deploying without VM Clones. See, Deploying primary and secondary nodes without VM Clones.

Secondary

The secondary node (Unit 2 in the diagram) is for HA support and normal file scans. It monitors the primary node's condition and, if the primary fails, the secondary will assume the role of primary. The former primary will then become a secondary when it is back up.

To support failover, ensure both the primary and secondary nodes are configured correctly:

  • Both the primary and secondary nodes must be the same model.
  • Both nodes must have the same network interface configuration, including:
    • The same subnet for port1.
    • The same subnet for port2.
    • The same subnet for port3.
    • The same routing table.

The secondary node is not required to set up a HA-Cluster but is recommended.

When the primary and secondary nodes are using a FortiSandbox VM model, you have the option of deploying without VM Clones. See, Deploying primary and secondary nodes without VM Clones.

Worker

The worker nodes (Units 3–5 in the diagram) perform normal file scans and report results back to the primary and secondary nodes. They can also store detailed job information. Workers should have their own network settings and VM image settings.

Workers can be any FortiSandbox model including FortiSandbox VM. Workers in a cluster do not need to be the same model.

The total number of worker nodes, including the secondary node, cannot exceed 100.

For heavy job loads, use more powerful FortiSandbox models.

Deploying primary and secondary nodes without VM Clones

When the primary and secondary node are using a FortiSandbox VM00 model, you have the option of deploying without VM Clones (i.e. dispatcher). That VM00 deployment dedicates its full VM resources for HA support, receiving incoming files and distribution of files to the worker nodes. There is no scan performed on the VM00. On this type of VM00 deployment, only the FortiCare Premium Support subscription is necessary as all the scans are performed on the worker nodes.

HA-Cluster

A single FortiSandbox device can scan a limited number of files in a given time period. To handle heavier loads, you can use multiple FortiSandbox devices in a load-balancing high availability (HA) cluster.

There are three types of nodes in a cluster: primary, secondary, and worker.

Primary

The primary node (Unit 1 in the diagram) manages the cluster, distributes jobs and gathers the results, and interacts with clients. It can also perform normal file scans. All scan-related configuration should be done on the primary node and they will be broadcasted from the primary node to the other nodes. Any scan-related configuration that has been set on a worker node will be overwritten.

On the primary node, users can:

  • Change a worker node's role (secondary and worker)
  • Configure a worker node's network settings
  • Upgrade worker nodes
  • View VM status page of worker nodes
  • Configure FortiGuard settings of worker nodes
  • Configure VM images of worker nodes, such as setting clone numbers of each VM image
  • Configure a ping server to frequently check unit's network condition and downgrade itself as a secondary node when necessary to trigger a failover

Although all FortiSandbox models can work as a primary node, we recommend using a more powerful model.

When the primary and secondary nodes are using a FortiSandbox VM model, you have the option of deploying without VM Clones. See, Deploying primary and secondary nodes without VM Clones.

Secondary

The secondary node (Unit 2 in the diagram) is for HA support and normal file scans. It monitors the primary node's condition and, if the primary fails, the secondary will assume the role of primary. The former primary will then become a secondary when it is back up.

To support failover, ensure both the primary and secondary nodes are configured correctly:

  • Both the primary and secondary nodes must be the same model.
  • Both nodes must have the same network interface configuration, including:
    • The same subnet for port1.
    • The same subnet for port2.
    • The same subnet for port3.
    • The same routing table.

The secondary node is not required to set up a HA-Cluster but is recommended.

When the primary and secondary nodes are using a FortiSandbox VM model, you have the option of deploying without VM Clones. See, Deploying primary and secondary nodes without VM Clones.

Worker

The worker nodes (Units 3–5 in the diagram) perform normal file scans and report results back to the primary and secondary nodes. They can also store detailed job information. Workers should have their own network settings and VM image settings.

Workers can be any FortiSandbox model including FortiSandbox VM. Workers in a cluster do not need to be the same model.

The total number of worker nodes, including the secondary node, cannot exceed 100.

For heavy job loads, use more powerful FortiSandbox models.

Deploying primary and secondary nodes without VM Clones

When the primary and secondary node are using a FortiSandbox VM00 model, you have the option of deploying without VM Clones (i.e. dispatcher). That VM00 deployment dedicates its full VM resources for HA support, receiving incoming files and distribution of files to the worker nodes. There is no scan performed on the VM00. On this type of VM00 deployment, only the FortiCare Premium Support subscription is necessary as all the scans are performed on the worker nodes.