How do rebootless updates (
vzreboot) work on Parallels Cloud Server.
Rebootless updates eliminate services outages for end users during system updates.
In a few words during the update the following is performed:
- Virtual Machines and Containers are suspended and memory dumps are saved to a memory region which is preserved during
- kexec (kernel execution), which is a mechanism of the Linux kernel that allows "live" booting of a new kernel "over" the currently running kernel;
- Virtual Machines and Containers are getting resumed instead of rebooting.
This greatly reduces the downtime of the server and the virtual environments stored on it.
To enable rebootless updates, the corresponding option should be included into the Parallels Cloud Server license
In earlier versions of Parallels Server Bare Metal, applying updates to your system often required rebooting the server. During reboot, all virtual machines and Containers running on the server were shut down and started again. If the number of running virtual machines and Containers was high enough, stopping and starting each of them could result in significant downtime.
Starting with version 6.0, Parallels Cloud Server supports rebootless updates. During such an update, all running virtual machines and Containers are suspended in the server's memory and then resumed rather than shut down and started again. This greatly reduces their downtime and virtually eliminates service outage or interruption for end users.
To perform a rebootless update:
Download and install the new Parallels Cloud Server kernel:
# yum install vzkernel
Run the vzreboot command:
This command checks the GRUB configuration file to determine the default kernel, loads this kernel into memory, and then boots the new kernel. You can also specify the kernel version manually, for example:
# vzreboot 2.6.32-042stab054.2
In this case, the command will load and then boot the specified kernel instead of the one set as default in the GRUB configuration file.