From 2 to 20 WildFire appliances that you want to group
and manage as a cluster. At a minimum, a cluster must have two WildFire
appliances configured in a high-availability (HA) pair.
Firewalls that forward samples to the cluster for traffic
analysis and signature generation.
) One or two Panorama appliances for centralized
cluster management if you choose not to manage the cluster locally.
To provide HA, use two Panorama appliances configured as an HA pair.
Each WildFire appliance you add to a WildFire appliance cluster
becomes a node in that cluster (as opposed to a standalone WildFire
appliance). Panorama can manage up to 10 WildFire appliance clusters
with a total of 200 WildFire cluster nodes (10 clusters,
each with the maximum of 20 nodes).
Panorama can manage standalone WildFire appliances as
well as WildFire appliance clusters. The combined total of standalone
WildFire appliances and WildFire appliance cluster nodes that Panorama
can manage is 200. For example, if Panorama manages three clusters
with a total of 15 WildFire cluster nodes and eight standalone WildFire
appliances, then Panorama manages a total of 23 WildFire appliances and
can manage up to 177 more WildFire appliances.
—Two controller nodes manage the
queuing service and database, generate signatures, and manage the
cluster locally if you don’t manage the cluster with a Panorama
M-Series or virtual appliance. Each cluster can have a maximum of
two controller nodes. For fault tolerance, each WildFire appliance
cluster should have a minimum of two nodes configured as a primary
controller node and a controller backup node HA pair. Except during
normal maintenance or failure conditions, each cluster should have
two controller nodes.
Worker Node (cluster client)
—Cluster nodes that are
not controller nodes are worker nodes. Worker nodes increase the analysis
capacity, storage capacity, and data resiliency of the cluster.
Server Node (cluster server)
—The third node in a WildFire
cluster is automatically configured as a server node, a special type
of worker node that provides database and infrastructure redundancy
features in addition to standard worker node capabilities.
When a firewall registers with a cluster node, or when you add
a WildFire appliance that already has registered firewalls to a
cluster, the cluster pushes a registration list to the connected
firewalls. The registration list contains every node in the cluster.
If a cluster node fails, the firewalls connected to that node reregister
with another cluster node. This type of resiliency is one of the
benefits of creating WildFire appliance clusters.
A WildFire appliance cluster increases the
analysis throughput and storage capacity available on a single network
so that you can serve a larger network of firewalls without segmenting
If a cluster node goes down, HA configuration
provides fault tolerance to prevent the loss of critical data and
services. If you manage clusters centrally using Panorama, Panorama
HA configuration provides central management fault tolerance.
Single signature package distribution
All firewalls connected to a cluster receive
the same signature package, regardless of the cluster node that
received or analyzed the data. The signature package is based on
the activity and results of all cluster members, which means that
each connected firewall benefits from the combined cluster knowledge.
Centralized management (
You save time and simplify the management
process when you use Panorama to manage WildFire appliance clusters.
Instead of using the CLI and scripting to manage a WildFire appliance
or cluster, Panorama provides a single-pane-of-glass view of your
network devices. You can also push common configurations, configuration
updates, and software upgrades to multiple WildFire appliance clusters,
and you can do all of this using the Panorama web interface instead
of the WildFire appliance CLI.
When a cluster has two or more active nodes,
the cluster automatically distributes and load balances analysis,
report generation, signature creation, storage, and WildFire content
distribution among the nodes.