Cloudera 5.12 documentation suggests that static resource pools configure cgroups to control memory limits between applications.
I’m trying to understand how these pools are configured on the server.
I’ve verified that our servers use cgroups v1 (Redhat 7.4) – but I can’t see how there are any limits configured or how these services are in fact in separate groups.
Our static service pools are configured as follows:
Below is output of cgroup configuration information, it appears yarn, impala are all part of user.slice and not in separate cgroups.