Description of problem: The header of the default /etc/openshift/resource_limits.conf file says: # NOTE: If you change these, please run the following commands: # service mcollective restart # oo-cgroup-enable --with-all-containers # oo-pam-enable --with-all-containers # oo-admin-ctl-tc restart # but this is not documented elsewhere. The current guides don't document these commands (and oo-cgroup-enable in particular) AFAICT, nor the process of updating that file. Version-Release number of selected component (if applicable): OSE 2.1 How reproducible: This request comes from a particular use case: if you try to increase memory_limit_in_bytes above your previousy set memory_memsw_limit_in_bytes, oo-cgroup-enable will fail; you have to increase the latter first. Raising a doc bug to see if this could be officially documented better, either in the admin guide, in the file itself, or both. Additional info: For now the mentioned particular use case is documented in kbase (attached). It seems it's not the first time this causes confusion, e.g. bug 1045540.
This info was added to the OSE Administration Guide as part of the updates that went in for BZ#1133936: https://access.redhat.com/documentation/en-US/OpenShift_Enterprise/2/html-single/Administration_Guide/index.html#Adding_New_Gear_Profiles