HDDS-13666. [DiskBalancer]Avoid reverse looping in destination selection when thresholds are low

[Gargi-jais11]

What changes were proposed in this pull request?

Sorting based on density is ok to choose source and possible destination, but the logic of choosing destination here can not be direct reverse loop, and need care for choosing destination should be make that node as candidate for reverse movement. And if no destination is possible to choose, should mark disk as balanced.
This reverse looping is especially seen in case of very low threshold such that the volume utilisation of source and destination disk is close making diskbalancer running forever without doing actual balancing.

As a solution to this while choosing container from source volume, we can check if movement of this container should satisfy the below condition.

if (newDestUtilization <= (average utilization + threshold)) {
// container move is productive
    return containerData;
}    

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-13666

How was this patch tested?

Added new UT TestDefaultContainerChoosingPolicy.

[Gargi-jais11]

@ChenSammi could you please review the changes? Thank you.

[ChenSammi]

The last patch looks good to me +1. Wait for the CI to pass.