HDDS-16066. replicas after re-replication should span at least 2 racks

[hevinhsu]

What changes were proposed in this pull request?

This pull request addresses the flaky rack-aware placement test reported in HDDS-16066 and removes obsolete FindBugs exclusions for the nested tests in TestRackAwarePlacement.

The rack-aware placement test previously waited only for the container replica count to reach three after shutting down a datanode. Since the original replica set already contained three replicas, this condition could be satisfied before the dead datanode replica was removed and before re-replication completed.

The test now waits until:

• At least three replicas are present.
• The stopped datanode is no longer included in the replica set.

This prevents the test from checking rack placement while the container is temporarily under-replicated.

The obsolete SIC_INNER_SHOULD_BE_STATIC FindBugs exclusions for the nested test classes are also removed. The affected helper methods are made non-static to match their usage by the nested test classes.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-16066

How was this patch tested?

https://github.com/hevinhsu/ozone/actions/runs/30880442815

[adoroszlai]

Thanks @hevinhsu for the patch, LGTM. Can you please run 10x10 flaky-test-check for TestRackAwarePlacement*?

[hevinhsu]

Thanks @adoroszlai for the reviews!

The flaky-test-check run is here:
https://github.com/hevinhsu/ozone/actions/runs/30899377348

PTAL when you have time. Thanks!

[adoroszlai]

https://github.com/hevinhsu/ozone/actions/runs/30899377348

Thanks for triggering it, but the parameters seem wrong. It should be:

• Test Class = TestRackAwarePlacement* (instead of NA)
• Number of Iterations per split = 10 (instead of org.apache.hadoop.hdds.scm.TestRackAwarePlacement)

[hevinhsu]

Oops, my bad 😅 I mixed up the fields. I’ll retrigger it with the correct parameters. Thanks!

[adoroszlai]

Thanks @hevinhsu for triggering the test. There are 5 failures without the fix (on master) and 4 with the fix. So there may be another source of the problem left?

[yandrey321]

lgtm