mirror of https://github.com/kubernetes/kops.git
Forcing all DNS communication go through TCP can lead to increased latency and slower DNS responses since a lot of upstream DNS servers are not optimized for TCP traffic. An example is AWS's VPC DNS server that sending all the DNS traffic through TCP cause it to increase the latencies for this requests. The kubernetes version has already changed this configmap to skip force_tcp flag for the . zone, so it makes sense to follow their example. Reference https://github.com/kubernetes/kubernetes/blob/master/cluster/addons/dns/nodelocaldns/nodelocaldns.yaml#L100 |
||
|---|---|---|
| .. | ||
| k8s-1.12.yaml.template | ||