TODO: add a glossary of terms here and their corresponding values and how to get them. E.g.: service cluster CIDR, service cluster DNS IP, pod CIDR, etc. etc.
VPC subnet: the subnet of the GCE virtual private cloud network from which internal IP addresses for each VM instance will be taken.
- The VPC subnet has no relation to the cluster CIDR and must not overlap with the cluster CIDR.
- 10.240.0.0/24 in this tutorial.
Cluster CIDR: a subnet in private IP address space, from which IP addresses for all Kubernetes components will be taken. The cluster CIDR is carved into smaller pod CIDRs.
- The cluster CIDR has no relation to the VPC subnet and must not overlap with it.
- 10.200.0.0/16 in this tutorial.
Pod CIDR: a subnet of the cluster CIDR that is assigned to a specific worker node, from which pod IP addresses will be taken.
- This tutorial uses instance metadata as the mechanism for telling each
worker node its pod CIDR. To fetch the pod CIDR from within an instance,
run:
- Linux: TODO
- Windows: TODO
- For worker node N in this tutorial, the pod CIDR is 10.200.N.0/24
Admin: (TODO: is this a real component?)
API Server: TODO
Controller Manager: TODO
Kube Proxy: TODO
Kube Scheduler: TODO
Service Account: TODO