Management Node High Availability
The management node controls, monitors, schedules, allocates, and reclaims resources in ZSphere. If the management node becomes unavailable, management services stop, affecting ZSphere operations, monitoring and alarms, user access, and automated tasks. High availability for ZSphere management services therefore protects day-to-day platform operations.
Dual-Management-Node High Availability
ZSphere provides a dual-management-node high-availability solution with the following characteristics:
- If either management node loses power, the system continues operating and the UI remains available for task execution.
- After the failed node recovers, its databases, including MySQL and the monitoring database, synchronize automatically.
- The system can operate on a single management node for an extended period.
- Deployment and operation of dual management nodes remain comparable to those of a single management node.
An HA process runs on each management node and initializes, configures, monitors, and maintains the management-node environment.
- The HA process provides a command-line interface for configuring the system as a high-availability environment.
- It monitors critical services, including the management node process, UI process, and MySQL. If a service fails, Keepalived triggers VIP failover and the HA process attempts to recover the failed service.
- It provides a watchdog for Keepalived to ensure that Keepalived remains running.
- It provides commands for displaying cluster health information.
- A gateway serves as an arbitration endpoint to prevent split-brain between the two management nodes.
Management Node Monitoring
Management node monitoring displays the management node IP address, node status, VIP, and management service status for each node. For details, see Management Node Monitoring.
Management Node Monitoring Data Collection
For the monitoring-data collection mechanism, see Monitoring and Alarms.