From Standalone to Self-Maintaining: Tracing the Evolution of Database High Availability Architectures

Published on 2025-05-09

Since the advent of database technology, users' demand for business continuity has grown increasingly stronger, driving continuous innovation in high availability architectures. This article takes you through the evolution of database high availability architectures from standalone to self-maintaining.

The Standalone Era: The 'Bare' Database

Scenario

The business is just starting up, and the database runs on a single server — simple, direct, and low-cost.  

Problems

Prone to downtime: Hardware failure, system crash, or human error can all cause service interruptions.  

Risk of data loss: Without backup, a disk failure can lead to permanent data loss.  

Complex maintenance: Every upgrade or maintenance requires stopping the service, leading to constant complaints from business teams.  

 

HAC Cluster: The First Step to High Availability

Solution

Deploy an HAC (High Availability Cluster).  

Improvements

Active-standby switchover: When the primary database fails, the standby automatically takes over, ensuring business continuity.  

Data synchronization: Real-time sync between primary and standby prevents single-point data loss.  

New Problems

Complex connections: Applications must configure addresses for all cluster nodes.

Slow recovery: After the standby becomes primary, the failed node requires manual repair, putting heavy pressure on O&M.

 

HAC+VIP: Simplified Access, But O&M Still Complex

Optimization

Introduce a virtual IP (VIP) so that applications access the database through a fixed IP.  

Benefits

Simple connections: Applications connect directly to "VIP:port" without configuration changes.  

Seamless switchover: When the primary fails, the VIP automatically drifts to the standby, with no perceptible impact on the business.  

Remaining Issues

Single-node risk: If the primary fails and the standby takes over, only a single node remains; any subsequent problem will cause a complete business interruption.  

Manual repair: The failed node requires manual repair, which is slow and inefficient, especially in remote areas.  

 

HAC+VIP+hac_alline: The Birth of Perfect High Availability!

Ultimate Solution

GBase database adds a cluster daemon (hac_alline) on top of HAC+VIP, achieving self-maintaining high availability. 

 Core Capabilities

1. Automatic Failure Detection 

The daemon monitors database and network health in real time. 

2. Automatic VIP Failover  

When the primary fails, the VIP switches within seconds to a healthy node.  

3. Node Self-Healing  

The failed node is automatically restarted, and after recovery it can automatically rejoin the cluster without manual intervention.  

4. Split-Brain Prevention  

Pre-checks before restarting the failed node avoid dual primaries, ensuring data consistency.  

Results

  • 7x24 business continuity  

  • Dramatically reduced O&M costs (no need to rush to the data center!)  

  • Zero data loss (real-time replication + automatic repair)  

 

Summary: The Evolution of High Availability  


Our database's 'comeback' story is not just a process of technological evolution, but a perfect practice of reliability, continuity, and high availability concepts.

Today, GBase database's "HAC+VIP+hac_alline" high availability architecture has been deployed at scale in critical infrastructure sectors such as power and water utilities, providing rock-solid data support for industry customers.