Feat Request: Use current-context in Kubernetes module when default context doesn't exist.

[webdog]

What should it do?

In the Kubernetes module, if a context isn't specified, the behavior is to revert to the default context. If this doesn't exist, the application panics, and returns an invalid memory address.

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x1e0 pc=0x4fd5d7c]
goroutine 38 [running]:
k8s.io/client-go/tools/clientcmd.(*DeferredLoadingClientConfig).createClientConfig(0xc000432050)
k8s.io/client-go@v10.0.0+incompatible/tools/clientcmd/merged_client_builder.go:77 +0x2bc
k8s.io/client-go/tools/clientcmd.(*DeferredLoadingClientConfig).ClientConfig(0xc000432050)
                                                                                                                                                                                                    k8s.io/client-go@v10.0.0+incompatible/tools/clientcmd/merged_client_builder.go:98 +0x25
github.com/wtfutil/wtf/modules/kubernetes.(*Widget).getKubeClient(0xc0000da640)
github.com/wtfutil/wtf/modules/kubernetes/client.go:37 +0x16a
                                                                                                                                                                                     github.com/wtfutil/wtf/modules/kubernetes.(*Widget).getInstance.func1()
                                                                                                                                                                                                                                                            github.com/wtfutil/wtf/modules/kubernetes/client.go:20 +0x65
sync.(*Once).doSlow(0x57597d6, 0x53db640)
sync/once.go:68 +0xd2
 sync.(*Once).Do(...)
sync/once.go:59
                                                                                                                                                                       github.com/wtfutil/wtf/modules/kubernetes.(*Widget).getInstance(0xc0000da640)
                                                                                                                                                                                                                                                        github.com/wtfutil/wtf/modules/kubernetes/client.go:18 +0x65
github.com/wtfutil/wtf/modules/kubernetes.(*Widget).Refresh(0xc0000da640)                                                                                                                               github.com/wtfutil/wtf/modules/kubernetes/widget.go:50 +0x4a                                                                                                                                        github.com/wtfutil/wtf/app.Schedule({0x5e2d7f8, 0xc0000da640})
github.com/wtfutil/wtf/app/scheduler.go:12 +0x31
created by github.com/wtfutil/wtf/app.(*WtfApp).scheduleWidgets
github.com/wtfutil/wtf/app/wtf_app.go:167 +0xb0

A current-context exists, but I do not use a default context. I programatically generate my kubeconfigs with the aws CLI, and then remove the kubeconfig after the cluster is removed. When aws CLI regenerates the config, it isn't expected to create a default context based on the parameters sent to the aws cli.

To reproduce:

• Have no kubeconfig file exist where the module config is expecting.
• Create/update a kubeconfig without a default context: aws eks update-kubeconfig --region us-west-2 --name my-eks-cluster --alias my-alias
• Confirm kubectl works:

❯ ~ kubectl get pods --all-namespaces
NAMESPACE     NAME                       READY   STATUS    RESTARTS   AGE
kube-system   aws-node-hwc8r             1/1     Running   0          48m
kube-system   aws-node-jb4gd             1/1     Running   0          49m
kube-system   aws-node-jhhvj             1/1     Running   0          48m
kube-system   coredns-657694c6f4-4w6gx   1/1     Running   0          56m
kube-system   coredns-657694c6f4-lndn5   1/1     Running   0          56m
kube-system   kube-proxy-2qdwc           1/1     Running   0          50m
kube-system   kube-proxy-82w46           1/1     Running   0          50m
kube-system   kube-proxy-vvnh4           1/1     Running   0          50m

• launch wtfutil, and witness similar error as above.

It seems reasonable (to me 😄 ) that a current-context can exist when a default doesn't, as this setup isn't necessarily breaking the kubeconfig. wtfutil seems to be state-dependent on the kubeconfig, adding support for the current would/could mitigate this somewhat.

Thank you!

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[webdog]

@senorprogrammer Any chance the wontfix label could be removed?

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.