Closed mrhillsman closed 5 years ago
Running the following resulted in the pods running:
root@master:~# kubectl -n kube-system taint nodes node00 node.cloudprovider.kubernetes.io/uninitialized:NoSchedule-
root@master:~# kubectl -n kube-system taint nodes node01 node.cloudprovider.kubernetes.io/uninitialized:NoSchedule-
I waited a significant amount of time before running the above commands; +2hrs
@mrhillsman by default dns is deployed to non-tainted nodes (workers). Did you have any worker nodes?
@jakolehm I do have worker nodes; node00 and node01. Here is my cluster.yml
hosts:
- address: "10.145.81.191"
private_interface: ens3
user: ubuntu
ssh_key_path: ~/.ssh/id_rsa
role: master
- address: "10.145.81.192"
private_interface: ens3
user: ubuntu
ssh_key_path: ~/.ssh/id_rsa
role: worker
- address: "10.145.81.193"
private_interface: ens3
user: ubuntu
ssh_key_path: ~/.ssh/id_rsa
role: worker
# network:
# provider: weave
# weave:
# trusted_subnets:
# - "172.31.0.0/16"
network:
provider: calico
service_cidr: 172.31.0.0/16
pod_network_cidr: 172.32.0.0/16
cloud:
provider: external
config: ./cloud-config
addons:
kubernetes-dashboard:
enabled: true
Cannot reproduce this. @mrhillsman did you figure out why nodes have node.cloudprovider.kubernetes.io/uninitialized=true:NoSchedule
taint? (it sounds like cni/network problem)
@jakolehm unfortunately not and had to move on to another deployment tool though I was really liking pharos.
Sent with GitHawk
I am using the latest version available via releases v1.3.0-rc.3 and the CoreDNS pods are not coming online due to taints:
http://paste.mrhillsman.com/paste/2w4IIuGPmTw4p97HQeqe1w
I am sure this was an issue with the latest 1.2 branch as well but have not confirmed since I installed v1.3.0-rc.3 just before deploying again.