Cluster

What is an Atlas cluster?

An Atlas cluster is a group of BeyondTrust appliances configured to work together as part of a distributed system, providing redundancy and load balancing for enhanced performance and reliability.

How is an Atlas cluster useful?

An Atlas cluster ensures high availability, fault tolerance, and scalability for your BeyondTrust environment, making it easier to manage large deployments and ensuring continuous service even in the event of hardware failures.

How do I access the Cluster page?

1. Use a Chromium-based browser to sign in to your Remote Support URL.
   This URL is provided in the BeyondTrust welcome email and includes your site URL followed by /login.
2. From the left menu, click Management.
   The Software page opens and displays by default.
3. At the top of the page, click Cluster.
   The Cluster page displays.

How to configure an Atlas cluster

Status

Large-scale geographic deployments benefit from BeyondTrust Atlas Cluster technology, establishing a single BeyondTrust site across multiple B Series Appliances, which are termed nodes in a cluster. The primary B Series Appliance/primary node is the site of most administration tasks. The traffic node is a BeyondTrust Appliance B Series that participates in effectively routing your support traffic.

On the primary node, you will configure both the primary itself and the traffic nodes.

ℹ️

Note

Find more information about Atlas in the BeyondTrust Atlas Technology Guide.

Current status

Confirms the role of the site instance from which you accessed the page.

Primary node(s)

Displays a list of the primary nodes available.

Sync now

Synchronize the clustered B Series Appliances.

Disband cluster

Disband the cluster, effectively removing each B Series Appliance from its role in the cluster.

Status history

Show or hide the log of clustered B Series Appliance messages.

Traffic nodes

Method for choosing traffic nodes

This selector is used to define how a traffic node is chosen for a representative or customer client connection. The available methods for defining the connection are Random, A Record Lookup, SRV Record Lookup, IP Anycast, and Timezone Offset. Your choice of connection method is highly dependent upon your network infrastructure, among other complex considerations.

Add, edit, remove

Create a new node, modify an existing node, or remove an existing node.

Accepting new client connections

Be sure this is checked; otherwise, clients will not be able to use the traffic node.

Add traffic node

Accepting new client connections

If enabled, then new client connections will be allowed to use this node. If you disable new client connections for this node, then existing client connections will not be affected.

ℹ️

Note

If this is not checked, clients will not be able to use the traffic node.

Name

Create a unique name to help identify this node.

Public address

Enter the hostname you set up in DNS for this node, and enter the port over which clients will communicate with the node.

Timezone offset

Used only if Method for Choosing Traffic Nodes is set to Timezone Offset. This process involves detecting the time zone setting of the host machine and using that setting to match the appropriate traffic node that has the closest time zone offset. The time zone offset is derived from the customer time zone setting relative to Coordinated Universal Time (UTC).

Internal address

This can be the same as the public address. Advanced configurations can optionally set this to a different hostname for inter-appliance communication.

Network address prefixes

You may leave this blank.

For advanced configurations, enter network address prefixes, one per line, in the form of ip.add.re.ss[/netmask]. Netmask is optional and can be given in either dotted-decimal format or as an integer bitmask. If netmask is omitted, as single IP address is assumed.

When this field is populated, the primary node attempts to assign a client to this traffic node if the client's IP address matches one of the network address prefixes. If the client's IP address matches more than one traffic node's network address prefixes, the client is assigned to the traffic node with the longest matching prefix. If the matching prefixes are of equal length, one of the matching traffic nodes is chosen at random. If a client's IP address does not match any network address prefixes, the client is assigned using the method configured.

Primary node configuration

Primary node

Name

Create a unique name to help identify this node.

Public address

Enter the hostname you set up in DNS for this node, and enter the port over which clients will communicate with the node.

Internal address

This can be the same as the public address. Advanced configurations can optionally set this to a different hostname for inter-appliance communication.

Backup primary node

The current site instance is the primary site instance in a failover relationship. If the failover roles are swapped, then the backup site instance will become the new primary node in the cluster. You can configure the cluster node configuration for the backup site instance below. The backup site instance will be automatically removed from the cluster if the failover relationship is broken.

Name

Create a unique name to help identify this node.

Public address

Enter the hostname you set up in DNS for this node, and enter the port over which clients will communicate with the node.

Internal address

This can be the same as the public address. Advanced configurations can optionally set this to a different hostname for inter-appliance communication.

Maximum client fallback to primary

Allows the number of clients you set to fall back to using the primary for traffic control if necessary.

When a client is unable to successfully choose a viable default traffic node (due to temporary mis-configurations, DNS problems, software-mismatches, etc.), it will have to fall back to using the primary node as its default traffic node. To avoid overloading the primary node with too much network traffic, this value can be used to limit the concurrent number of clients that will fall back to using the primary node in this capacity.