Published Release toolkit CRIU 3.16 (Checkpoint and Restore in Userspace), designed to save and restore processes in the user space. The toolkit allows you to save the state of one or group of processes, and then resume work from the saved position, including after restarting the system or on another server without breaking already installed network connections. Project code extends under the GPLV2 license.
From the applications of the CRIU technology, the OS restarting is noted without disrupting the continuity of the execution of a long time Processes, Live-migration of insulated containers, accelerate the start of slow processes (you can start working from the status saved after initialization), conducting the kernel updates without restarting services, the periodic preservation of the state of debt computing tasks to resume work in case of collaborators, balancing the load on nodes in clusters , Duplication of processes to another machine (fork on a remote system), creating snapshots of user applications during operation to analyze them on another system or in case you need to cancel further actions in the program. CRIU is used in container control systems such as OpenVZ, LXC / LXD and Docker. The changes required for the work of CRIU are included in the main composition of the Linux kernel.
In a new release:
- Added Team CRIU-NS , allowing you to restore the stored process With the new PID identifier and in a separate space of the namespace (Mount Namespace). Running with another PID may be required, for example, if the old PID is already used in the system.
- implements the ability to save and restore the casting of the state of the apparmor nested profiles.
- Locked and unlock network resources based on NFTables.
- Added support for recovery of pre-created VETH devices.
- improved support for the restoration of containers in existing POD-s.
- For RPC clients, the ability to determine the PID reuse is added, implemented using the PIDFD mechanism.
- License for all proto-files in the images / Changed on Mit.