zhuangchong edited a comment on pull request #5387: URL: https://github.com/apache/dolphinscheduler/pull/5387#issuecomment-826712422
When you deploy with one click of docker, the alert plugin report does not exist, which leads to the failure of the whole deployment What you do: add a loop in the alert server startup to check whether the plugin table exists. If it does not exist, always print the warning log (I see that you just added a loop number judgment) I don't know if I understand this issue If I understand correctly, I reserve my opinion. Formally, from a global perspective, I suggest that dependencies be detected before starting the program during deployment. What I don't understand is that the one click deployment is successful, but the alert server can't alarm normally because the table doesn't exist, and the service process is still there. This is to improve the user experience from a global perspective? If there is no difference, there will be no discussion. I reserve my opinion --- 你在使用docker 一键部署的时候遇到alert plugin表报错不存在，导致整个部署失败 你的做法：在alert-server启动里面加了一个循环检查plugin表是否存在功能，如果不存在一直打印警告日志（我看你刚刚加看一个循环次数的判断） 不知道这是不是理解了这个issue 如果我理解的正确，我保留我的意见。 正式从全局角度出发，我建议是在部署过程中启动程序前去检测依赖项。我不理解的是让用户部署一键部署成功了，但是alert-server因为表不存在，不能正常告警，服务进程还在，这个就是从全局角度出发，提升了用户体验？ 如果没有分歧，就不会发生讨论，我保留自己的意见。 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
