Linux Checkpoint/Restart BoFSDave Hansen (firstname.lastname@example.org)
Being able to move applications to cope with hardware availability or restarting an application from its exact state before a crash have been two very sought-after features in the computing community for many years. There have been countless implementations, some in userspace, others in kernels, even more in hypervisors.
However, many of these implementations have significant drawbacks. Out-of-tree kernel patches, proprietary products, and implementations in userspace that don't address a wide range of applications have been the some of the largest issues with these implementations. A comprehensive, Linux-distributor-included framework would help many users who have a need for checkpoint/restart, but want a better solution.
This BOF is intended for all those interested in such a common application checkpoint/restart framework for Linux, based on the existing containers work. We will discuss the current needs and look to the future for what work is next.