Adding Cluster Nodes

Adding a new node to an existing cluster results in its full synchronization with the initial node. This process includes clearing local data and restoring configuration and data based on information replicated from the cluster.

Warning

  • Data model objects: safes, users, servers, accounts and listeners are replicated within the cluster and object instances must not be added on each node. In case the replication mechanism fails to copy objects to other nodes, contact technical support department.

  • Session data and configuration parameters (servers, users, accounts, safes, listeners, external authentication servers) of the node being joined to the cluster are deleted. After joining, the node is initialized with data replicated from the cluster.

  • In environments containing a large amount of data, the initialization and data copying process to the joining node may take several hours.

Danger

  • The mechanism of joining a node to the cluster relies on completely wiping its data and copying data from the node on which the cluster was initialized.

  • Incorrect configuration of relationships between nodes or skipping any step of the procedure may lead to irreversible damage to the joining node. In such a case, the synchronization process will not be completed.

  • Some configuration elements (e.g., SMTP) are not replicated between cluster nodes and must be configured manually on each node.

Cluster Initialization and Adding Nodes

To add a node to the Fudo Enterprise cluster, follow the instructions below.

  1. Log in to the Fudo Enterprise Administration Panel where the cluster was initialized — the initial node.

  1. From the left menu, select Settings > Cluster.

  2. Click Add node.

../../_images/5-6-add-node-button.png

  1. Enter the node name and a description to help identify the node.

  2. Provide the IP address of the node being added (that is, the Fudo Enterprise instance to which data from the initial node will be replicated).

Note

The address of the joining node must have the General role set in the network settings. For more information about configuring network settings, see Network Interfaces Configuration.

../../_images/5-6-add-node-form.png

  1. Click Fetch to download the node’s SSH public key.

  2. Click Save and close.

  3. Repeat steps 3–7 if you want to add additional nodes.

Note

In scenarios where password changers are used, at least two active cluster nodes are required for them to operate. In a two-node environment, the failure of a single node causes password changers to stop working; therefore, a three-node configuration is recommended.

Setting Relationships Between Nodes

Note

  • The recommended cluster configuration is to define bidirectional relationships between nodes. It is not required to configure relationships between every pair of nodes; however, no node may remain without at least one defined relationship.

  • In a topology where nodes are not connected in a “full mesh” (each node connected to every other node), be aware of the limitations regarding session data replication. In such configurations, session data is replicated only between nodes that have a defined relationship.

Warning

Creating relationships is essential to ensure that data is replicated between nodes.

  1. Click the name of the first node to edit it.

  2. In the Relations section of the joining node, click Add Relation.

  3. From the Remote name list, select the node to which data from the node you are currently logged in (in this case, the primary node) will be replicated.

  4. From the Database list, select which data should be replicated to the selected cluster node.

  5. Select the OCR option to allow the selected node to process excess graphical sessions.

  6. Click Save and close.

  7. Edit the remaining nodes and, in the same way, add relationships to the other nodes in the cluster.

../../_images/5-6-relations.png

Note

Each Fudo Enterprise instance has a license-defined limit on the number of OCR processes used to process graphical sessions. The OCR option allows delegating the processing of excess sessions to the selected node when the number of connections exceeds the number of local processes responsible for processing and indexing content.

Linking Nodes with the Initial Node

  1. Make sure you are in the Fudo Enterprise Administration Panel where the cluster was initialized.

  1. Go to the main view of the Cluster tab.

  2. Copy the cluster SSH public key.

../../_images/5-6-copy-ssh-key.png

  1. Log in to the Administration Panel of the joining node.

Note

You can use the shortcut by clicking the address of the selected node in the list of added nodes in the cluster view.

  1. From the left menu, select Settings > Cluster.

  2. Go to the JOIN CLUSTER tab.

../../_images/5-6-join-cluster.png

  1. Paste the copied key into the Cluster SSH public key field.

  2. Click Save.

Warning

After accepting the data wipe warning, all data on this node will be deleted.

  1. Read the data wipe message and click I understand the consequences. Proceed.

  2. Wait until the cluster is initialized and the objects are copied between cluster nodes.

Note

The added node will be restarted. If disk encryption is used (USB key or password), connect the key or provide the password in the hypervisor terminal.

Cluster Initialization Completion Verification

The cluster initialization time depends on the amount of data on the initial node. For a new, unused instance, the process may take several minutes. In environments containing large amounts of data (e.g., logs, session recordings), initialization may take from several hours to several days.


After replication is complete, log in to the newly added node and verify synchronization by:

  • checking the cluster node status on the Dashboard,

  • analyzing the event log,

  • verifying cluster settings.

Additionally, it is recommended to confirm replication of configuration data (e.g., users, servers) by reviewing the appropriate tabs in the administrative interface.


Completion of the initial replication is confirmed by log entries on the initial node (the one where the cluster was created):

  • Initial objects replication to cluster node <name> completed.

  • Initial logs replication to cluster node <name> completed.


Related topics: