Open debianmaster opened 6 years ago
possible reasons
cleanup of ips for dead pods is not properly done. (may be bug in older ocp versions)
debugging
check for timestamps here ls -l /var/lib/cni/networks/openshift-sdn/
Drain the node:
oc adm drain $node systemctl stop atomic-openshift-node rm -rf /var/lib/cni/networks/openshift-sdn/* systemctl start atomic-openshift-node oc adm uncordon $node
Also check https://github.com/cloudnativelabs/kube-router/issues/383. There is a (automatic) workaround using https://github.com/jsenon/api-cni-cleanup.
thanks @tobiasdenzler
Is there a built-in solution for this issue in Openshift 4 ? I have this issue in Openshift 4.7.
@donnydavis no clue.