| 摘要 |
Systems and techniques are described for managing virtual machines crashes. A described technique includes determining that a virtual machine is unresponsive, in response to determining that the virtual machine is unresponsive, identifying one or more files in which a guest operating system (OS) of the virtual machine writes data when the guest OS abnormally terminates, the one or more files being stored in virtual storage, determining, based on contents of the one or more files, whether the guest OS has abnormally terminated, in response to determining that the guest OS has abnormally terminated, monitoring requests initiated by the virtual machine to determine whether the guest OS has finished writing to the one or more files, and in response to determining that the guest OS has finished writing to the one or more files, copying the one or more files to a location in hardware storage, and restarting the virtual machine. |
