Deleting the CNI directory
First run command fails at setup storage step with an error. Pod is stuck in pending state on the host for a long time.
To fix this issue:
Delete the CNI directory on the host failing to launch pods:
ssh
root@ecs-ha1-p-7.vpc.cloudera.com rm -rf /var/lib/cni
Restart the canal pod running on that host:
kubectl get pods -n kube-system -o wide | grep ecs-ha1-p-7.vpc.cloudera.com kube-proxy-ecs-ha1-p-7.vpc.cloudera.com 1/1 Running 0 11h 10.65.52.51 ecs-ha1-p-7.vpc.cloudera.com <none> <none> rke2-canal-llkc9 2/2 Running 0 11h 10.65.52.51 ecs-ha1-p-7.vpc.cloudera.com <none> <none> rke2-ingress-nginx-controller-dqtz8 1/1 Running 0 11h 10.65.52.51 ecs-ha1-p-7.vpc.cloudera.com <none> <none> kubectl delete pod rke2-canal-llkc9 -n kube-system