High system‑reserved cpu usage when autoSizingReserved: true is enabled

Environment

• Red Hat OpenShift Container Platform (RHOCP)
  • 4
• Red Hat Enterprise Linux CoreOS (RHCOS)
  • 4

Issue

• When the autoSizingReserved: true feature is configured in the Kubelet configuration, the system-reserved CPU allocation is observed to be lower than the default reservation of 500m.
• After configuring autoSizingReserved: true, the node is still showing high system-reserved CPU usage.

Resolution

Red Hat is aware of this issue and it has been reported to Engineering, and it has been fixed in the following releases.

Workaround

If an immediate upgrade is not possible, refer to which amount of CPU and memory are recommended to reserve for the system in OpenShift 4 nodes to calculate the recommended values, and then manually adjust CPU and memory reservations to prevent the dashboard from showing usage above 80% by following the documentation for manually allocating resources for nodes.

Note: when configuring the system resource reservations manually, it is needed to configure both, the CPU and also the memory.

Root Cause

This is a known bug in the automatic resource reservation calculation, causing the reservation of CPU for system reserved resources for specific numbers of CPUs is calculated lower than the default reservation, which causes the UI metrics to show very high system reserved CPU usage.

Diagnostic Steps

When autoSizingReserved: true is enabled in the cluster, verify the resource allocation values in the /etc/node-sizing.env file on each node using the following command.

$ oc debug node/[node_name]
[...]
sh-4.4# chroot /host bash
# cat /etc/node-sizing.env

This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.