1

We are using CNI plugins bridge/host-local on nodes with limited pool of IPs which can be exhaused before the cpu/mem are exhaused.

In such case, how can I ensure the pods are not scheduled to the pods where IPs are exhaused but the resources are still available.

Is there any way to taint the nodes once the IPs are exhausted.

  • when the IP pool is exhausted does this add any annotations to the node? How do you know when it is exhausted? – aurelius Jul 02 '19 at 14:34

0 Answers0