- Green versionCheck
- Green versionCheck
- Green versionCheck
- Green versionCheck
The two servers (hosts) of the ServHA dual-machine system are connected to the disk array (shared storage) system. The user's operating system, application software and ServHA software are installed on the two hosts respectively, and shared data such as databases are stored on the storage system. A heartbeat connection is established between two hosts through a private (or public) network.
After the configured system host starts working, the ServHA service starts to monitor the system. Through the heartbeat connection, the ServHA service on each host can monitor the status of the other host. When the working host fails, the heartbeat information will change, and this change can be captured by the ServHA service through the heartbeat connection.
After capturing this change, ServHA will control the system to perform host switching, that is, the backup machine starts the same application as the working host to take over the work of the working host (including providing TCP/IP network services, storage system access, etc.) and An alarm is issued to prompt management personnel to repair the faulty host.
After the repair is completed, it can be switched back automatically or manually according to the settings of ServHA, or it can not be switched. At this time, the repaired host will be used as a backup machine, and the dual-machine system will continue to work.
The key to ServHA's fault tolerance function is that the host is transparent to the client. When a system error occurs and a switch is performed, the switch of the host does not appear to the client to have changed, and all host-based applications will still run normally. ServHA uses IP drift technology to achieve this function.
The client communicates with the working host through the virtual address. Regardless of whether the system switches or not, the virtual address always points to the working host.
When performing network services, ServHA provides a logical virtual address, and any client only needs to use this virtual address when requesting services. During normal operation, virtual addresses and network services are provided by the main server. When the main server fails, ServHA will transfer the virtual address to the network card of another server to continue to provide network services.
After the switchover is completed, it appears to the client that the system has not failed and network services can still be used. In addition to IP addresses, ServHA can also provide virtual computer aliases for client access. For database services, when a main server fails, another server will automatically take over and start the database and applications at the same time, so that the user database can operate normally.
it works
it works
it works