HDDS-14702. Make lazy source container replica deletion interval configurable

[Gargi-jais11]

What changes were proposed in this pull request?

Currently source container replica deletion is postponed by 1 hour after the container is successfully moved from source volume to destination volume during disk balancing.
This PR aims to make this configurable, and reduce the default value from 1 hour to 5 minutes.

<property>
    <name>hdds.datanode.disk.balancer.replica.deletion.delay</name>
    <value>5m</value>
    <tag>OZONE, DATANODE, DISKBALANCER</tag>
    <description>The delay after a container is successfully moved from source
      volume to destination volume before the source container replica is deleted.
      By default this is set to 5 minutes. Unit: ns, ms, s, m, h, d.
    </description>
  </property>

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-14702

How was this patch tested?

Passed existing test cases for replica deletion delay.

[Gargi-jais11]

@ChenSammi Please review the patch.