K
Q

unable to upgrade connection: pod does not exist

1/7/2020

I have a error when i want to access to my pod:

error: unable to upgrade connection: pod does not exist

it's a cluster with 3 nodes, below some details. Thanks in advance

root@kubm:~/deploy/nginx# kubectl get nodes -o wide

NAME       STATUS   ROLES    AGE   VERSION   INTERNAL-IP   EXTERNAL-IP   OS-IMAGE             KERNEL-VERSION      CONTAINER-RUNTIME

kubm       Ready    master   37h   v1.17.0   10.0.2.15     <none>        Ubuntu 16.04.6 LTS   4.4.0-150-generic   docker://19.3.5

kubnode    Ready    <none>   37h   v1.17.0   10.0.2.15     <none>        Ubuntu 16.04.6 LTS   4.4.0-150-generic   docker://19.3.5

kubnode2   Ready    <none>   37h   v1.17.0   10.0.2.15     <none>        Ubuntu 16.04.6 LTS   4.4.0-150-generic   docker://19.3.5

root@kubm:~/deploy/nginx# kubectl get pods -o wide
NAME                    READY   STATUS    RESTARTS   AGE   IP           NODE       NOMINATED NODE   READINESS GATES

nginx-59c9f8dff-v7dvg   1/1     Running   0          16h   10.244.2.3   kubnode2   <none>           <none>

root@kubm:~/deploy/nginx# kubectl exec -it nginx-59c9f8dff-v7dvg -- /bin/bash

**error: unable to upgrade connection: pod does not exist**
-- maroki
kubernetes

Similar Questions

How to access kubernetes services externally on bare-metal cluster
Jenkins deployment to K8 cluster
K8S Auditing - how to get all the policy options?
PostgreSQL HA in Kubernetes with already replicated persistent volume
Kube-proxy with IPVS mode doesn&#39;t keep a connection
How MetalLB works in Kubernetes
Airflow DAG serialization: TypeError: Object of type &#39;V1Pod&#39; is not JSON serializable
Fixing DataDog agent congestion issues in Amazon EKS cluster

3 Answers

5/7/2020

The 10.0.2.15 IP address is the default for virtualbox NAT

If you deploy a VM using a vagrantfile, your eth0 adapter will use the 10.0.2.15 IP address and the eth1 adapter will be assigned an other IP address.

K8s uses the eth0 adapter to route packets between pods.

-- Dmitry Zadorov
Source: StackOverflow
5/12/2020

I had the same issue running a cluster with Vagrant and Virtualbox the first time.

Adding KUBELET_EXTRA_ARGS=--node-ip=x.x.x.x where x.x.x.x is your VM's IP in /etc/default/kubelet (this can be part of the provisioning script for example) and then restarting kubelet (systemctl restart kubelet) fixes the issues.

This is the recommended way to add extra runtime arguments to kubelet as you can see in /etc/systemd/system/kubelet.service.d/10-kubeadm.conf. Alternatively you can also edit the kubelet config file under /etc/kubernetes/kubelet.conf

-- IppX
Source: StackOverflow
1/7/2020

I use below command to get into a pod.

kubectl exec -i -t <pod-name> -- /bin/bash

Note -i and -t flag have a space on the command..

If you have multi-container pod you should pass container name with -c flag or it will by default connect to first container in POD.

ubuntu@cluster-master:~$ kubectl exec -i -t nginx -- /bin/bash
root@nginx:/# whoami
root
root@nginx:/# date
Tue Jan  7 14:12:29 UTC 2020
root@nginx:/#

Refer help section of command kubectl exec --help

-- DT.
Source: StackOverflow