The physical characteristics of HDDs and potential hardware issues with HBAs may cause a state where the physical disk can still perform I/O operations, but with significantly reduced performance. This condition is referred to as a sub-healthy state. ABS actively identifies sub-healthy disks by monitoring the historical performance of physical disks and, while ensuring data safety, isolates them to minimize I/O access to these disks. At the same time, ABS safely reconstructs the associated data on sub-healthy disks onto other physical disks or nodes to ensure continuous data safety.
In terms of networking, network port or optical module abnormalities, or misconfigured uplink network policies, may cause packet loss, increased latency, and other instability. In such cases, if a port bonding policy is configured on the network, ABS automatically switches the network port to bypass sub-healthy NICs. When the node's network is globally abnormal, ABS proactively isolates the node, ensuring that data links, critical services, and data are migrated from the abnormal node to other available nodes, maintaining service stability.
