K
Q

Problems with pod-network in kubernetes

6/29/2018

I created 3 vbox VM for testing k8s. Each vm has 2 netwoks: Nat and internal. K8s was initialized with:

kubeadm init --apiserver-advertise-address 192.168.1.1 --service-cidr 192.168.1.0/24 --pod-network-cidr 192.168.1.0/24

192.168.1.0/24 is the internal network. Nodes were joined by

kubeadm join 192.168.1.1:6443 --token some_token --discovery-token-ca-cert-hash hash

When I'm trying to deploy flannel I get errors like:

I0629 09:25:06.640787       1 main.go:475] Determining IP address of default interface
I0629 09:25:06.645316       1 main.go:488] Using interface with name enp0s3 and address 10.0.2.15
I0629 09:25:06.645335       1 main.go:505] Defaulting external address to interface address (10.0.2.15)
I0629 09:25:06.656691       1 kube.go:131] Waiting 10m0s for node controller to sync
I0629 09:25:06.656794       1 kube.go:294] Starting kube subnet manager
I0629 09:25:07.657778       1 kube.go:138] Node controller sync successful
I0629 09:25:07.657795       1 main.go:235] Created subnet manager: Kubernetes Subnet Manager - kub2
I0629 09:25:07.657798       1 main.go:238] Installing signal handlers
I0629 09:25:07.657892       1 main.go:353] Found network config - Backend type: vxlan
I0629 09:25:07.657928       1 vxlan.go:120] VXLAN config: VNI=1 Port=0 GBP=false DirectRouting=false
E0629 09:25:07.658039       1 main.go:280] Error registering network: failed to acquire lease: node "kub2" pod cidr not assigned

what am I doing wrong?

-- Kirill Ponomarev
kubeadm
kubernetes

Similar Questions

Keep a Kubernetes Pod in service when in a state of not ready
Fluentd logging on kubernetes skips logs on log rotation
Can I have multiple values.yaml files for Helm
How to install rabbitmq plugin on kubernetes?
Why doesn't the kubernetes documentation show the valid mappings for requests or capacity?
What is the best way to get a sample k8 yaml to edit as per requirement?

1 Answer

6/29/2018

As per error string in your logs:

E0629 09:25:07.658039 1 main.go:280] Error registering network: failed to acquire lease: node "kub2" pod cidr not assigned

It seems you used subnet 192.168.1.0/24 which was fully distributed for master node creation; however, there were not enough IP pools for work nodes building, as by default kube-controller-manager gives /24 network for each node. As the solution, you can use wider network subnet addressing, and also be aware that subnets you intend to use can't be a part of each other.

For example, if you have your local network 192.168.1.0/24, --pod-network-cidr should be like 192.168.2.0/24 and --pod-network-cidr like 172.17.0.0/16

-- mk_sta
Source: StackOverflow