Load balancing mode
All members in the server pool are primary members. Traffic is evenly distributed to backend servers according to the scheduling policy, achieving load balancing.
Primary-secondary mode
The server pool contains one primary member and one secondary member. As long as the primary member is healthy, traffic is not forwarded to the secondary member. When the primary member becomes unavailable, the secondary member takes over and handles traffic. When the primary member recovers, traffic is automatically migrated back to the primary member.
Whether the secondary member takes effect depends on whether a health monitor is configured for the server pool. If no health monitor is configured, the primary member is always considered healthy, and no primary-secondary switchover is triggered.
Multiple primaries–one secondary mode
The server pool contains multiple primary members and one secondary member. The secondary member is activated only when all primary members fail. In this mode, the secondary member mainly serves as a fallback, for example, to indicate that the server pool has failed, and does not normally handle actual traffic. If a health monitor is configured for the server pool, the secondary member is always considered healthy by default and does not participate in health checks.
