HA (hight availability) is a function provided by the esxi server cluster. Its main purpose is to run the virtual machine on the physical host, virtual machine operating system, and applications in the virtual machine.ProgramWhen a fault occurs, the VM can be restarted quickly. External services are not interrupted and data is not lost.
Fault Level: 1. esxi host faults. Ii. Virtual Machine operating system faults. 3. Application faults.
How to handle the first fault: An esxi physical host with HA enabled in a cluster, which is set every 5 seconds by default) it sends a heartbeat detection signal to other physical hosts in the cluster to prove that they are alive. If other hosts do not receive signals from the host after three cycles, that is, 15 seconds, other hosts believe that the host has encountered a physical fault, the VM running on this host will be restarted on another physical host to provide services again. Note that there is a very important premise: the files of each virtual machine are stored in other disk arrays in a shared manner instead of in the local storage mode (this is in my other article)ArticleVmfs data storage is described in the following section ). In this way, when host a fails, host B can copy the Virtual Machine file on host a in the shared disk array and re-run it on host B. Therefore, from the perspective of external clients, it is only one system restart.
The second and third types of faults are handled in a similar way: first, a wmware proxy tool is installed in the virtual machine operating system, the proxy tool is used to send heartbeat detection signals between the virtual machine and the software running on the virtual machine. If the heartbeat detection signal sent by the virtual machine is not received in three cycles, vcenter restarts the Virtual Machine (or the original physical host ). Similarly, if a software fault occurs, restart the virtual machine before restarting the software.
Original article:High Availability principle of VMware vsphere5.0 ha