Cluster Limitations and Redundancy
While clustering allows VMS Gateway to scale across multiple machines, it is important to understand the current redundancy and failover limitations of VMS Gateway 9.2.
Primary Node Responsibilities
The primary node hosts:
The Kubernetes (K3s) control plane.
VMS Gateway configuration and management services.
The VMS Gateway web interface.
Any VODs deployed to the primary node.
Worker nodes host only VOD workloads and contribute processing capacity to the cluster.
Primary Node Failure
If the primary node becomes unavailable:
Existing video streams running on worker nodes continue to operate.
Existing video streams running on the primary node continue only while the primary remains operational.
No configuration changes can be made.
Additional VODs cannot be created, modified, or removed.
Worker nodes cannot be promoted to become the new primary node.
Currently, complete loss of the primary node typically requires restoration from backup or a full redeployment of the cluster.
Worker Node Failure
VODs are permanently assigned to the node on which they are created.
If a worker node fails:
VODs running on that node stop operating.
VODs are not automatically redeployed to other cluster nodes.
Client systems connected to those VODs will lose connectivity until the node is restored or the VODs are recreated elsewhere.
To recover from the loss of a worker node, the recommended approach is to provision a replacement server using the same hostname and reinstall VMS Gateway.
Workload Placement
VMS Gateway 9.2 does not currently provide:
Automatic load balancing of VODs between nodes.
Automatic workload migration.
Automatic failover of VODs to another node.
High-availability primary node functionality.
When additional capacity is added to the cluster, existing VODs remain on their current nodes. Moving a VOD between nodes requires deleting it from the existing node and recreating it on the target node.
Backup and Recovery
VMS Gateway 9.2 does not currently provide a dedicated backup and restore mechanism for on-premises deployments.
As an interim best practice, we recommend:
Regularly backing up the primary VM.
Retaining deployment records and installation details.
Maintaining documented network and system configuration information to assist with disaster recovery scenarios.
Planning for Resilience
When designing a VMS Gateway deployment, consider that clustering in version 9.2 provides capacity scaling rather than high availability. Additional nodes increase the number of VODs that can be hosted, but they do not provide automatic redundancy or failover for workloads running on other nodes.