kubernetes/minikube

profile list returns non-zero if a cluster is not running

Open

#6,743 建立於 2020年2月22日

在 GitHub 查看
 (4 留言) (0 反應) (0 負責人)Go (31,799 star) (5,222 fork)batch import
help wantedkind/bugkind/featurelifecycle/frozenpriority/backlog

描述

when I was looking at this intergeration test that failed because profile list says one of the clusters is not running yet...

https://storage.googleapis.com/minikube-builds/logs/6717/e653589/Docker_Linux.html#fail_TestFunctional%2fparallel%2fProfileCmd%2fprofile_list

and also one of the users who was using minikube in a test script,... made me think.

When are the resonable times to exit none-zero for profile list ?

currently we exit non-zero for the following:

  • If there is no profile .
  • if there are a few profiles are one of them has error to give status running.

貢獻者指南

技術棧
go
領域
cli
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
half day
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
needs investigation
前置要求
Go
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
35
研究方向
Investigate the current implementation of the `profile list` command in the minikube Go source code, likely in `cmd/minikube/cmd/profile list.go`. Understand the existing exit code logic and the cases where non zero is returned. Review the referenced integration test failure (https://storage.googleapis.com/minikube builds/logs/6717/e653589/Docker Linux.html#fail TestFunctional%2fparallel%2fProfileCmd%2fprofile list) to see the expected behavior. Participate in the issue discussion to clarify the desired exit code semantics for cases like a cluster not running.