Resource management across different tools

This proposal has been accepted as a session.


One Line Summary

Effectively managing machine resources using different cgroup tools.


Different cgroup tools setup an application’s cgroup slightly differently, sometimes depending on the underlying distro they are running on. Each approach present a different model for managing the whole machine and type of guarantees that applications can expect. We’ll cover some resource models for each type and show some example use-cases.

We’d like to discuss ways to minimize/hide the differences and present a manageable machine model to higher layers like cluster manager or monitoring services.


QOS, cgroups, lxc, docker, lmctfy