Closed luoyuqi-lab closed 1 year ago
BTW, I make a mistake that directly pushed the working branch to master, so I can not PR at first...
Then I roll-back and make a PR, now.
Please rebase your PR
Moving forward, observe the single responsibility principle that each PR is only for one change, either delete aptuser or add dagobah.
Scrum or agile software development frame, coincidentally, trains developers to be lazy in the sense that they only do minimal work that satisfy requirements. If the requirements are to remove aptuser, they don't add dagobah. Agile trades task size for task quantity, meaning each task is smaller but more tasks are done.
Got the idea of a single PR handling a single issue, next time I'll use it.
Please help me take a look at whether my rebase is right. @gqqnbig
Well.. I realized we have #129. The removed eureka would serve #129, right?
Well.. I realized we have #129. The removed eureka would serve #129, right?
Eureka is fixed and we must add it to the cluster? Ok, it seems we should not delete Eureka in cluster-add-user/cluster-add-user.sh
Considering the GPU issue on Hoth, I figure the easiest way to deal with this PR is to include only the aptuser changes, and 86 all other stuff.
Because now we no need aptuser group, so we do not create it for users.
And we use add dagobah and hoth nodes, instead of down eureka.