HDDS-12040. ozone freon cr fails with NPE in ReplicationSupervisor

[len548]

What changes were proposed in this pull request?

"ozone freon cr" command runs into exceptions because requestCounter map in ReplicationSupervisor class is not initiated during the following reproducing steps. addTask method in the class is supposed to initiate it but it is not called, so I extracted that part into a new method called initCounters and call this new method before running the task.

How to reproduce the error:

// create a key
ozone freon ockg -n 10
// close a container
ozone admin container close 1
// run the command
ozone freon cr -n 1

Also this PR handles the followings to make it work:

• ClosedContainerReplicator needs update for HDDS-10804
• It needs DB volumes with directories (HDDS-10230).

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-12040

How was this patch tested?

Manually tested

[adoroszlai]

Thanks @len548 for the patch. Please update your branch by merging master into it and resolving the conflict.

[adoroszlai]

Thanks @len548 for updating the patch.

[len548]

@adoroszlai, thank you for the review and correction!

[adoroszlai]

Thanks @len548 for iterating on this.

Sorry for nitpicking about method order and whitespace. If you look at the changes in ReplicationSupervisor you may see why it was important: makes review much easier. Now it's clear that addTask was just split into 3 methods, instead of having to inspect blocks of code line-by-line.

[len548]

@adoroszlai thank you for teaching this practice throughly and sorry for causing unnecessary changes again and again. I often miss these, especially the unnecessary whitespaces because the changes are not visually obvious on the IntelliJ ;(