This document details the procedures for deploying (adding) and removing nodes in the Segura cluster, including prerequisites, step-by-step instructions, operational considerations, and recommendations for maintenance without downtime. These operations must be performed by authorized administrators to ensure continuity and stability.
Prerequisites
- Only administrators with appropriate profiles can perform cluster expansion or reduction.
- All nodes (new or existing) must be on the same Segura version and synchronized in time (NTP).
- Backup cluster configurations before any changes.
- Verify firewall rules and connectivity between nodes.
Adding a node to the cluster
Step-by-step
-
Provisioning:
- Install the approved OS and Segura package on the new physical machine or VM.
- Configure fixed IP, hostname, and access permissions.
-
Access the admin panel:
- In the Segura central console, go to “Cluster Management”.
-
Start the addition process:
- Select “Add Node”.
- Fill in node details (IP, port, role, authentication).
-
Validation:
- The system runs automatic tests for connectivity, version check, synchronization, and dependencies.
-
Synchronization:
- If approved, the node automatically receives configurations, policies, and files, entering replication until reaching “Healthy” status.
-
Activation:
- Once synchronized and validated, the node joins the active load balancing pool and starts accepting sessions.
-
Verification:
- Confirm the node’s “Healthy” status and normal operation via dashboard.
Note: Adding nodes can be done without downtime for connected users.
Removing a node from the cluster
Step-by-step
-
Access the admin panel:
- In “Cluster Management,” identify the node to remove.
-
Selection and confirmation:
- Select “Remove Node”.
- Confirm operation; the system may display active sessions for safe handling.
-
Isolation:
- The node is automatically removed from the active pool. Active sessions are migrated or terminated per policy.
-
Uninstallation/shutdown:
- Safely uninstall Segura software from the machine if necessary.
- Update records and inventory.
-
Verification and audit:
- All steps and outcomes are logged for compliance and auditing.
Operational considerations
- Check synchronization and consistency before removal to avoid replication delays or unexported logs.
- Prefer removing nodes without active sessions; otherwise, the system automatically redistributes sessions.
- All operations are audited, including responsible user, date/time, and justification (if required).
Maintenance without downtime
- Use rolling updates to add or remove nodes without service interruption.
- Schedule operations outside peak hours whenever possible.
- Keep internal documentation updated with cluster changes.