Currently, operating systems have become increasingly complex and thus more sensitive to security problems and performance issues. In some circumstances, developers have to reboot their systems in order to patch security vulnerabilities, fix bugs, and add new features, which will bring unavoidable downtime for running applications. Live-patch is an elegant solution. But, it can only work in 30 percent of circumstances. In our project, we put another solution into practice. We use a checkpoint-and-restart mechanism to retore applications' status and then switch the kernel fastly.