On Mon, Mar 12, 2018 at 05:57:53PM -0700, David Rientjes wrote: > The cgroup aware oom killer is needlessly enforced for the entire system > by a mount option. It's unnecessary to force the system into a single > oom policy: either cgroup aware, or the traditional process aware. Can you, please, provide a real-life example, when using per-process and cgroup-aware OOM killer depending on OOM scope is beneficial? It might be quite confusing, depending on configuration. >From inside a container you can have different types of OOMs, depending on parent's cgroup configuration, which is not even accessible for reading from inside. Also, it's probably good to have an interface to show which policies are available. Thanks!
