chectl fail on version comparison

[cb3zopp]

Fixed for the next channel.
Could you try to check if problem has gone ?

1. chectl update next
2. chectl server:deploy ...

Originally posted by @tolusha in #19165 (comment)

[cb3zopp]

Did you fix this also für chectl for Windows?
If I install with Windows 10 Hyper-V without a proxy it works.
But on an Azure VM behind a HTTPS proxy I still get the error also with the 'next' channel.

[tolusha]

@cb3zopp
Could you pls attach logs please?
You can use chectl server:logs command

[cb3zopp]

PS C:\cloud> set_net
PS C:\cloud> minikube start --addons=ingress --vm=true --memory=8192

• minikube v1.18.1 on Microsoft Windows 10 Enterprise 10.0.19042 Build 19042
• Using the hyperv driver based on existing profile
• Starting control plane node minikube in cluster minikube
• Restarting existing hyperv VM for "minikube" ...
• Found network options:
  • HTTP_PROXY=185.46.212.88:80
  • HTTPS_PROXY=185.46.212.88:80
  • NO_PROXY=192.168.99.0/24,192.168.39.0/24,192.168.49.0/24,10.96.0.0/12,10.113.41.0/24,172.0.0.0/8
  • http_proxy=185.46.212.88:80
  • https_proxy=185.46.212.88:80
  • no_proxy=192.168.99.0/24,192.168.39.0/24,192.168.49.0/24,10.96.0.0/12,10.113.41.0/24,172.0.0.0/8
• Preparing Kubernetes v1.20.2 on Docker 20.10.3 ...
  • env HTTP_PROXY=185.46.212.88:80
  • env HTTPS_PROXY=185.46.212.88:80
  • env NO_PROXY=192.168.99.0/24,192.168.39.0/24,192.168.49.0/24,10.96.0.0/12,10.113.41.0/24,172.0.0.0/8
• Verifying Kubernetes components...
  • Using image us.gcr.io/k8s-artifacts-prod/ingress-nginx/controller:v0.40.2
  • Using image gcr.io/k8s-minikube/storage-provisioner:v4
  • Using image jettech/kube-webhook-certgen:v1.2.2
  • Using image jettech/kube-webhook-certgen:v1.3.0
• Verifying ingress addon...
• Enabled addons: ingress, storage-provisioner, default-storageclass
• Done! kubectl is now configured to use "minikube" cluster and "default" namespace by default
  PS C:\cloud> kubectl get po -A
  NAMESPACE NAME READY STATUS RESTARTS AGE
  kube-system coredns-74ff55c5b-svqmh 1/1 Running 2 3d5h
  kube-system etcd-minikube 1/1 Running 0 52s
  kube-system ingress-nginx-admission-create-gkjfn 0/1 Completed 0 3d5h
  kube-system ingress-nginx-admission-patch-chn6q 0/1 Completed 0 3d5h
  kube-system ingress-nginx-controller-65cf89dc4f-zj2ln 1/1 Running 3 3d5h
  kube-system kube-apiserver-minikube 1/1 Running 0 52s
  kube-system kube-controller-manager-minikube 0/1 Running 2 3d5h
  kube-system kube-proxy-cgh27 1/1 Running 2 3d5h
  kube-system kube-scheduler-minikube 1/1 Running 2 3d5h
  kube-system storage-provisioner 1/1 Running 5 3d5h
  PS C:\cloud> chectl --version
  chectl/0.0.20210408-next.cf44212 win32-x64 node-v12.22.0
  PS C:\cloud> chectl server:deploy --platform minikube
  TypeError: Invalid Version: undefined
  PS C:\cloud> chectl server:logs
  Eclipse Che logs will be available in 'C:\Users\cb3zopp\AppData\Local\Temp\chectl-logs\1618229297428'
  › Current Kubernetes context: 'minikube'
  × Verify Kubernetes API
  → Failed to connect to Kubernetes API, error: E_K8S_API_NO_RESPONSE - Endpoint: https://172.21.186.161:8443/healthz
  …
  Verify if namespace 'eclipse-che' exists
  Read Operator logs
  Read Eclipse Che server logs
  Read Postgres logs
  Read Keycloak logs
  Read Plugin registry logs
  Read Devfile registry logs
  Read namespace events
  » Error: Error: Failed to connect to Kubernetes API, error: E_K8S_API_NO_RESPONSE - Endpoint:
  » https://172.21.186.161:8443/healthz - Error message: getaddrinfo ENOTFOUND 80. If you're sure that your Kubernetes
  » cluster is healthy - you can skip this check with '--skip-kubernetes-health-check' flag.
  » Command server:logs failed. Error log: C:\Users\cb3zopp\AppData\Local\chectl\error.log Eclipse Che logs:
  » C:\Users\cb3zopp\AppData\Local\Temp\chectl-logs\1618229297428

[cb3zopp]

I guess chectl server:logs can only work after a successful deployment.
TypeError: Invalid Version: undefined
is the primary error.

[tolusha]

@cb3zopp
I got the problem.
Thank you for reporting.

[tolusha]

Meanwhile could you remove next-update-info.json file and try again ?
On windows it is located in <user>\AppData\local\chectl directory.

[tolusha]

pls.
try chectl update next and then chectl server:deploy ... to check if issue is fixed.

[cb3zopp]

I think it works. I'm one step further but get a new error.
My problem is that the no proxy configurations works for the kubernetes unix tools but not for the windows system itself.
For windows I can set the external proxy to get internet. In this case the direct 172.x.x.x adresses don't work. I get only access to the kubernetes cluster when the 172.x.x.x adresses are rebound to localhost.
Or I set a pac-file which should excludes the internal adresses from the sending to the proxy. In this case I don't get internet access. Must find error in the windows pacfile/configuration to get both at the same time.

× Verify Kubernetes API
→ Failed to connect to Kubernetes API, error: E_K8S_API_NO_RESPONSE - Endpoint: https://172.24.147.37:8443/healthz
-…
👀 Looking for an already existing Eclipse Che instance
» Error: Error: Failed to connect to Kubernetes API, error: E_K8S_API_NO_RESPONSE - Endpoint:
» https://172.24.147.37:8443/healthz - Error message: getaddrinfo ENOTFOUND 80. If you're sure that your Kubernetes
» cluster is healthy - you can skip this check with '--skip-kubernetes-health-check' flag.
» Command server:deploy failed. Error log: C:\Users\cb3zopp\AppData\Local\chectl\error.log Eclipse Che logs:
» C:\Users\cb3zopp\AppData\Local\Temp\chectl-logs\1618302145186

[tolusha]

Use --skip-kubernetes-health-check flag to skip Verify Kubernetes API step.

[cb3zopp]

I did. This helps to avoid the health check but some steps later there is the next E_K8S_API_NO_RESPONSE error. There must be a way to have a proxy set but exclude the local container adresses from the proxy routing.

[tolusha]

Do you mind create a dedicated issue?
I would like not to mix several problems into a single one.