Research And Implementation Of The Novel Heartbeat Inspecting Technique

High availability cluster technology can make the system to service the outside request seamlessly, it has great use in critical transaction processes. Based on the description about the knowledge of high availability cluster, the system structure of the high availability cluster and the technologies about the system fault diagnosis, checkpoint setting and roll-back recovery are discussed. Then, a dual computer hot-standby high availability (HA) system is designed and implied on the Windows NT platform.High availability cluster technology does focus on the data backup, node connection and storage technologies. HA has two nodes and uses the dual computer hot-standby working mode; the nodes are joined by two heartbeat lines; the nodes use SCSI to connect with a RAID, which has the same driver symbols in them. The detection and recovery of the system failure is the key technology for the high availability system. By using the different agents, HA implements the functions as follows: monitoring the status of the data base services, Web services, Email services and file sharing services, detecting the hardware failures. Through the Heartbeat-Gear heartbeat inspecting technique, the nodes can watch on the status of each other, this makes HA to implement the function of monitoring the nodes' failure.Availability is an important factor to be used in evaluating the performance of a high availability system. Using the Markov mode to analyze the heartbeat cycle's action on the dual-computer hot-standby system's availability, a new heartbeat inspecting technique named as Heartbeat-Gear is designed and realized. By dynamically adjusting the heartbeat cycle in a gear-stage method, Heartbeat-Gear can improve the HA system's Heartbeat Environment Adaptability (HEA) efficiently in a changing service-request environment, and can finally increase the whole system's high availability.