Provide Configurable Timeout for Getting Kubeconfig During Phase Run

[lb4368]

Problem description
The clusterctl client that is used to retrieve the kubeconfig for site cluster(s) currently does not provide any way to specify a wait timeout so a long default (300 seconds) is used. This can cause the airshipctl cluster get-kubeconfig command or any phase that gets the kubeconfig for one or more clusters to run long when a kubeconfig is inaccessible. There is an upstream clusterctl issue (kubernetes-sigs/cluster-api#4334) meant to provide this capability within the clusterctl client, but it is not implemented yet and will have to be incorporated as part of a clusterctl upgrade once it is.

Proposed change
Update the process for getting the kubeconfig when needed for a specified cluster to allow for a shorter timeout period.

Until the upstream issue is resolved, a workaround would be implemented. This may involve getting the kubeconfig using an older method (e.g. https://github.com/airshipit/airshipctl/blob/63421c8fcc11426e7af10d58733dfcaef6ad8589/pkg/k8s/kubeconfig/secret.go) and adding a timeout setting.

[raliev12]

Hi @lb4368,

I can work on this one, at least to prepare a workaround you mentioned.

[lb4368]

Hi @raliev12. Thanks!

[airshipbot]

Related Change #793472

Subject: Allow to setup timeout and retries for get kubeconfig request
Link: https://review.opendev.org/c/airship/airshipctl/+/793472
Status: NEW
Owner: Ruslan Aliev (raliev@mirantis.com)

Approvals

Code-Review
! None
Verified
-1 ATT Airship2.0 CI
+1 Zuul
Workflow
! None

Last Updated: 2021-05-28 01:59:56 CDT