the cluster problem

Hello,

You can check the following documentation about Cluster management and the benefits of configuring the 3VCN as cluster: Cluster Management or the Staking management.

Stack:

Application Scenario
Due to limited access capabilities of a single VCN, large-scale sites often deploy a large number of VCNs to meet service demands. The stacking management feature can efficiently manage VCNs and cameras in a centralized manner.

Customer Benefits
•Centralized management, reducing operation and maintenance costs.
•Unified service operation portal, making service operations more convenient.
Functions
The stacking management feature provides the following functions:

A stack consists of 2 to 16 VCNs. One of the VCNs functions as the management server to centrally manage other slave servers. The management server provides a unified login entry, configuration management interface, and service operation interface, facilitating centralized management of servers and cameras. Figure 1 shows the stack functions.

Cluster:

Application Scenario
To continue using cameras connected to a faulty MPU, a user can set up an MPU cluster so that cameras can be automatically migrated from the faulty MPU to normal MPUs. This ensures service continuity.

Customer Benefits
Multiple MPUs in the local domain can be deployed in cluster mode to enhance system reliability. When an MPU is faulty, its services and cameras can be automatically migrated to a normal MPU to ensure service continuity.

Functions

Fault migration
The cluster possesses the capabilities of automatic fault detection, service migration, and fault recovery. The platform collects the operating status of each node in real time. Upon detecting a faulty node, the platform can quickly schedule another normal node in the cluster to take over services from the faulty node, ensuring service continuity.

NOTE:
It is recommended that the total number or media bandwidth of cameras connected to a cluster not reach the maximum performance of the VCN.

For example, consider a VCN540 that supports connection to a maximum of 64 cameras. If a cluster needs to connect to 320 cameras, it is recommended to use at least six VCN540s. In this way, if a VCN540 in the cluster is faulty, the cameras connected to the faulty VCN540 can be migrated to other VCN540s whose performance has not reached the maximum. A VCN540 supports a maximum of 320 Mbit/s media access. If the bit rate of a camera is 8 Mbit/s, a VCN540 can connect to a maximum of 40 cameras. If a cluster needs to connect to 200 cameras whose bit rate is 8 Mbit/s, it is recommended to use at least six VCN540s.

Load balancing
The cluster can dynamically balance service loads on each node based on the node status to prevent overload on a single node.

When a new node is added to the cluster or the faulty node recovers, the system can automatically balance service loads on each node to achieve load balancing.

NOTE:
Load balancing in a cluster is implemented based on the recording storage space first and then the number of cameras without recording plans.

For example, consider a cluster that consists of MPUs A, B, C, and D, each of which is connected to 50 cameras. The same recording plan has been set for A, B, and C under the proper operating and specification requirements. No recording plan is set for D. Therefore, the recording storage space of D is the lowest and some cameras connected to A, B, and C will be migrated to D. In addition, some cameras that do not have recording plans on D will be migrated to A, B, and C to balance the load.

The recording storage space is determined by the bit rate and number of cameras. For example, the recording storage space required by 64 cameras with a bit rate of 4 Mbit/s equals the recording storage space required by 32 cameras with a bit rate of 8 Mbit/s. In this example, the service loads of the two MPUs connected to 64 cameras and 32 cameras respectively are balanced. Therefore, the number of cameras connected to a cluster member may be greater than the number connected to another cluster member.

Elastic scaling
When the service system expands, users can continuously increase the node quantity to a maximum of 8 nodes per cluster. New nodes can automatically share cluster services. If the service volume is small and the service load on each node is light, users can decrease the node quantity. The services on the nodes that are removed from the cluster can be automatically switched to other cluster nodes, allowing services to continue running properly.

Cloud-based management of clusters
Allows users to:

•View the storage space of a cluster, cluster member status, camera status in a cluster, and platform recording plan status of a cluster.
•Export offline camera information into an Excel file.

Understanding the Stack
This topic describes the stack function and typical stack networks.
Introduction to the Concept of Domain
Before understanding the stack function, understand the concept of domain by referring to Configuration Precautions.
Functions
The stacking management feature provides the following functions:
A stack consists of 2 to 16 VCNs. One of the VCNs functions as the management server to centrally manage other slave servers. The management server provides a unified login entry, configuration management interface, and service operation interface, facilitating centralized management of servers and cameras. Figure 1 shows the stack functions.
Figure 1 Stack functions

Generally, a stack is a domain. The management server is called the master server. Other servers in the stack are called slave servers.
Typical Network
Figure 2 shows a typical network. The total number of master and slave servers cannot exceed 16. If the VCN540, VCN520, and VCN510 coexist in a stack network, it is recommended that a device with the most powerful performance, for example, VCN540, function as the master server.
Figure 2 Typical network



In the case of a single cluster:
VCN3000 series: Allows 3–16 VCN3000s to be deployed as a cluster.
VCN500 series: Allows 3–8 VCN500s to be deployed as a cluster.
The VCN500 and VCN3000 cannot be deployed in the same cluster

very nice :)

Hello, bro! Hoping you are fine. A stack consists of 2 to 16 VCNs. One of the VCNs functions as the management server to centrally manage other slave servers. The management server provides a unified login entry, configuration management interface, and service operation interface, facilitating centralized management of servers and cameras. The cluster possesses the capabilities of automatic fault detection, service migration, and fault recovery. The platform collects the operating status of each node in real time. Upon detecting a faulty node, the platform can quickly schedule another normal node in the cluster to take over services from the faulty node, ensuring service continuity.

bro.

Hoping you are fine.

1.Stacking is the centralized management of multiple servers. We can log in to the domain master server to view the status of all servers.

2.The cluster is load-balanced and unified for all servers on a stack basis. It must be composed of three or more servers.

Stacking is the foundation of a cluster, which is an extension of a stack.

for cluster:
To continue using cameras connected to a faulty MPU, a user can set up an MPU cluster so that cameras can be automatically migrated from the faulty MPU to normal MPUs. This ensures service continuity.
for stacking:
Due to the limited access capabilities of a single VCN, large-scale sites often deploy a large number of VCNs to meet service demands. The stacking management feature can efficiently manage VCNs and cameras in a centralized manner.