Virtual Clusters
A virtual cluster is developed based on a large-scale cluster, which includes multiple
virtual sub-clusters (VCs). Each VC runs independently within the entire cluster and
shares a unified entry. With permission, VCs can access each other.
The Coordinator node (including Gcluster node and Gcware node) and VC's data
node support running on the same physical node.
Unified Management
By dividing the cluster into virtual sub-clusters, the cluster can reach a scale of
thousands of nodes while providing a unified management view, enhancing the
cluster's scalability.
Unified access entry to the cluster
Figure 4-2. Unified access entry.
GBase 8a MPP Cluster Technical White Paper
- 24 -
General Data Technology Co., Ltd.
Users select any Coordinator node in the Coordinator cluster (which includes the
Gcluster scheduling cluster and the Gcware management cluster, usually deployed
together and referred to as the Coordinator cluster) as the cluster entry point. The
Coordinator node determines the default VC based on user input.
Note: The Coordinator node stores metadata for all VC databases and tables.
Business Isolation
Virtual clusters isolate resources vertically in the cluster, enabling physical resource
isolation for different businesses in real-world business scenarios.
Transparent Access
Virtual clusters provide a unified access point, and accessing multiple physical
clusters is transparent to the application, which appears to be a unified cluster.