On Mon, Jul 29, 2019 at 11:48 AM Michal Hocko <mhocko@xxxxxxxxxx> wrote: > > On Mon 29-07-19 10:28:43, Yang Shi wrote: > [...] > > I don't worry too much about scale since the scale issue is not unique > > to background reclaim, direct reclaim may run into the same problem. > > Just to clarify. By scaling problem I mean 1:1 kswapd thread to memcg. > You can have thousands of memcgs and I do not think we really do want > to create one kswapd for each. Once we have a kswapd thread pool then we > get into a tricky land where a determinism/fairness would be non trivial > to achieve. Direct reclaim, on the other hand is bound by the workload > itself. Yes, I agree thread pool would introduce more latency than dedicated kswapd thread. But, it looks not that bad in our test. When memory allocation is fast, even though dedicated kswapd thread can't catch up. So, such background reclaim is best effort, not guaranteed. I don't quite get what you mean about fairness. Do you mean they may spend excessive cpu time then cause other processes starvation? I think this could be mitigated by properly organizing and setting groups. But, I agree this is tricky. Typically, the processes are placed into different cgroups according to their importance and priority. For example, in our cluster, system processes would go to system group, then latency sensitive jobs and batch jobs (they are usually second class citizens) go to different groups. The memcg kswapd would be enabled for latency sensitive groups only. The memcg kswapd threads would have the same priority with global kswapd. > -- > Michal Hocko > SUSE Labs
