# OpenShift v4.6.43 - CRI-O PID Leak Causing High Memory Utilization # - RefURL # [1]: https://access.redhat.com/solutions/6304881 # [2]: https://bugzilla.redhat.com/show_bug.cgi?id=1997062 apiVersion: apps/v1 kind: DaemonSet metadata: name: execpid-cleaner namespace: kube-system spec: selector: matchLabels: name: execpid-cleaner template: metadata: labels: name: execpid-cleaner spec: tolerations: - operator: Exists affinity: nodeAffinity: requiredDuringSchedulingIgnoredDuringExecution: nodeSelectorTerms: - matchExpressions: - key: node.openshift.io/os_id operator: In values: - "rhcos" terminationGracePeriodSeconds: 5 containers: - name: execpid-cleaner image: [your-private-registry-ip-or-domain:port]/openshift4/release:4.6.43-tools command: - sh - -c - | set -euo pipefail while true; do printf "%s Running cleanup:\n" "$(date)" set -x find /run/crio/exec-pid-dir -type f -mmin +5 -exec rm -f {} \; sleep 5m set +x done securityContext: privileged: true resources: limits: memory: 64Mi requests: memory: 64Mi volumeMounts: - name: exec-pid-dir mountPath: /run/crio/exec-pid-dir volumes: - name: exec-pid-dir hostPath: path: /run/crio/exec-pid-dir
previous article
OpenShift v4.x – Self-hosted Load Balancer for OpenShift
next article