HDDS-6676. KeyValueContainerData#getProtoBufMessage() should set block count

[kaijchen]

What changes were proposed in this pull request?

When implementing HDDS-6518, I found KeyValueContainerData#getProtoBufMessage() does not set block count.

Although it is an optional field in ContainerData, I think we should set it to make ReadContainer command more useful.

HDDS-6518 may not get merged since another approach HDDS-6665 is taken. So I filed a separate Jira for this improvement.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-6676

How was this patch tested?

See TestReplicationService#testReadContainer in #3351.
But I don't think there is a need to add new test in this PR.

[kerneltime]

Can we add a unit test where all attributes unless explicitly skipped are part of the serialized protobuf?

[kaijchen]

Can we add a unit test where all attributes unless explicitly skipped are part of the serialized protobuf?

I think we can check x.equals(x.fromProto(x.toProto)), do you have any better idea?

[neils-dev]

A unit test can be added to TestKeyValueContainer (hadoop-hdds/container-service/src/test/java/org/apache/hadoop/ozone/container/keyvalue/TestKeyValueContainer.java). This test, TestKeyValueContainer#testKeyValueDataProtoBufMsg, can exercise the protobuf and all attributes set:
ie.

populate(numberOfKeysToWrite);
ContainerDataProto proto = containerData.getProtoBufMessage();
assertEquals(proto.getState(), keyValueContainerData.getState());
...
assertEquals(proto.getBlockCount(), numberOfKeysToWrite);

[kaijchen]

Hi @neils-dev, thanks for the advice. I have added the test.

[adoroszlai]

Thanks @kaijchen for the patch, @kerneltime and @neils-dev for the review.

[kaijchen]

Thanks @kerneltime @neils-dev and @adoroszlai for the review.