Skip to content

Bump org.bouncycastle:bcprov-jdk18on from 1.79 to 1.85 (5.2) - #2695

Merged
apurtell merged 1 commit into
apache:5.2from
apurtell:backport-2681-5.2
Nov 22, 2026
Merged

apurtell merged 1 commit into
apache:5.2from
apurtell:backport-2681-5.2

Conversation

@apurtell

Copy link
Copy Markdown
Contributor

Backport of #2681 to 5.2. Cherry-picked with -x from the master commit; it applies cleanly.

Posted by Claude Code on behalf of Andrew Purtell.

🤖 Generated with Claude Code

Bumps [org.bouncycastle:bcprov-jdk18on](https://github.com/bcgit/bc-java) from 1.79 to 1.85.
- [Changelog](https://github.com/bcgit/bc-java/blob/main/docs/releasenotes.md)
- [Commits](https://github.com/bcgit/bc-java/commits)

---
updated-dependencies:
- dependency-name: org.bouncycastle:bcprov-jdk18on
  dependency-version: '1.85'
  dependency-type: direct:development
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
(cherry picked from commit a4b9357)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@apurtell

Copy link
Copy Markdown
Contributor Author

Posted by Claude Code on behalf of Andrew Purtell.

The two failing checks are unrelated to this change (pom-only bcprov bump).

Test coverage: This is the known 5.2 categorization gap. Some ITs have no @Category, and SortOrderWithRegionMoves*IT runs in two jobs. Pending category fixes cover it.

ParallelStatsDisabledTest D-I: GlobalMutableTxIndexIT.testLastDDLTimestampOnAsyncIndexes[OMID] fails with an NPE in IndexToolIT.verifyMapper. The NPE hides the real error. IndexTool.run failed earlier, and getJob() was null. The test output shows the cause:

RejectedExecutionException ... ScheduledThreadPoolExecutor[Terminated, ... completed tasks = 478]
  at org.apache.omid.tso.client.TSOClient.getNewStartTimestamp
  at OmidTransactionContext.begin <- PostIndexDDLCompiler <- IndexTool.configureJobForAsyncIndex

Root cause (test-order dependent, latent on all branches):

  1. GlobalImmutableTxIndexIT ran just before in the same fork and on the same mini-cluster. Its test CQS started a TSO and injected that transaction manager into the process-wide OmidTransactionProvider singleton.
  2. That class ran IndexTool. ConnectionUtil.getInputConnection goes through PhoenixDriver, not the test driver. This creates a production ConnectionQueryServicesImpl, and initTransactionClient caches the singleton's current manager in it.
  3. At the end of that class, ConnectionQueryServicesTestImpl.close() closes its Omid service. This terminates that TSOClient.
  4. GlobalMutableTxIndexIT injects a new manager. Its IndexTool run reuses the PhoenixDriver CQS, which is still cached and still holds the closed manager. The run then fails.

The failure needs both of these conditions:

  • An earlier class in the same fork runs IndexTool on an OMID table.
  • The mini-cluster is not torn down between that class and this one.

The relevant code is identical on 5.2 and master. Other branches passed only because the fork assignment was different.

Conclusion: these failures are not caused by this PR. A follow-up JIRA can make OmidTransactionClient read the provider's current manager each time it is used, instead of capturing it at creation.

@apurtell
apurtell merged commit 3dff012 into apache:5.2 Nov 22, 2026
11 of 13 checks passed
@apurtell
apurtell deleted the backport-2681-5.2 branch November 22, 2026 20:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant