On Mon, Jul 30, 2018 at 07:04:26AM -0700, Tejun Heo wrote: > Hello, > > On Mon, Jul 30, 2018 at 10:03:57AM +0200, Michal Hocko wrote: > > Please be careful when defining differen oom.group policies within the > > same hierarchy because OOM events at different hierarchy levels can > > have surprising effects. Example > > R > > | > > A (oom.group = 1) > > / \ > > B C (oom.group = 0) > > > > oom victim living in B resp. C. > > > > OOM event at R - (e.g. global OOM) or A will kill all tasks in A subtree. > > OOM event at B resp. C will only kill a single process from those > > memcgs. > > That behavior makes perfect sense to me and it maps to panic_on_oom==2 > which works. Roman, what do you think? I'm totally fine with this behavior, this is what I've suggested initially. I'll post the patchset soon. Thanks!
