Add cluster usage balance CLI

[JySongWithZhangCe]

Is your feature request related to a problem? Please describe. I'm always frustrated when few Alluxio workers are full of data, while others are empty.

Describe the solution you'd like Add cluster usage balance CLI to manually balance the cache.

Describe alternatives you've considered

• Add a configuration about the maximum number of replicas in a single worker.
• Support clearing the cache of the specified worker.

Urgency Urgent

Additional context Nope.

[JySongWithZhangCe]

@jiacheliu3 Would you like to give some suggestions?

[jiacheliu3]

fyi @HelloHorizon

[JySongWithZhangCe]

For now, the applications has final and absolute control over how the cached data is distributed. In this case, excellent users can speed up data analysis well, while unprofessional users encounter various problems. When Alluxio provides services as a sharing platform, it is inevitable that unprofessional users will affect the use of other users.

At present, the administrator does not have a good way to intervene, and can only solve the problem by stopping the worker, deleting the cache, and starting the worker. Administrators need more tools to manage specific workers.

[JySongWithZhangCe]

The uneven distribution of data directly leads to the skew of task scheduling in analysis engines such as Spark, and hotspot workers are frequently under pressure from cache replacement.

[github-actions[bot]]

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