[
https://issues.apache.org/jira/browse/ATLAS-5338?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18095061#comment-18095061
]
ASF subversion and git services commented on ATLAS-5338:
--------------------------------------------------------
Commit 319a1d0d082042db90e333b61a76352808927936 in atlas's branch
refs/heads/master from Ramachandran Krishnan
[ https://gitbox.apache.org/repos/asf?p=atlas.git;h=319a1d0d0 ]
ATLAS-5338: Add Trino extractor integration and unit test suite (#691)
* ATLAS-5338: Add Trino extractor integration and unit test suite
Implement live integration tests that run the standalone extractor tarball
against Atlas + Trino, plus unit tests for cron validation and concurrent
job guards. Adds the trino-extractor-it Maven profile and README.
https://issues.apache.org/jira/browse/ATLAS-5338
* ATLAS-5338: Document components required for Trino extractor ITs
* ATLAS-5338: Skip Trino extractor ITs in default Maven verify.
Live-stack integration tests require Atlas and Trino; opt in with
-Ptrino-extractor-it so CI mvn verify does not invoke Failsafe here.
* ATLAS-5338: Use Java 8 compatible List helpers in TrinoExtractorITSupport
Replace List.of() with Arrays.asList() so test code compiles on JDK 8
(CI docker-build matrix).
* Harden atlas-hbase Docker startup for CI compose smoke test
RegionServer often lags Master on slow runners; the old healthcheck
(start_period 40s, 30s interval) and tail-on-HMaster-only entrypoint
caused flaky "atlas-hbase is unhealthy" failures during compose up --wait.
- Add shared atlas-hbase-healthcheck.sh (wget/curl) for Master + RS
- Wait up to 180s for both endpoints before marking container ready
- Keep container alive while Master or RegionServer JVMs are running
- Install wget/curl in atlas-base; tune healthcheck (120s start_period)
- CI: --wait-timeout 300 and dump logs when compose up fails
* Stabilize ImportTaskListenerImplTest async consumer tests.
Use per-test import request mocks and a synchronous executor so Mockito stubs
are not accessed concurrently, fixing the JDK 8 ClassCastException flake in
testStartImportConsumer_Successful.
Co-authored-by: Cursor <[email protected]>
---------
Co-authored-by: ramk <[email protected]>
Co-authored-by: Cursor <[email protected]>
> Implement Trino extractor integration test suite (TrinoExtractorIT)
> -------------------------------------------------------------------
>
> Key: ATLAS-5338
> URL: https://issues.apache.org/jira/browse/ATLAS-5338
> Project: Atlas
> Issue Type: Task
> Components: atlas-intg
> Reporter: Ramachandran Krishnan
> Assignee: Ramachandran Krishnan
> Priority: Major
> Fix For: 3.0.0
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> h3. Summary
> Implement the Trino extractor integration and unit test suite that was
> previously only sketched as comments in {{{}TrinoExtractorIT{}}}. Tests run
> the standalone extractor tarball ({{{}run-trino-extractor.sh{}}}) against a
> live Atlas + Trino stack, matching production layout rather than the Maven
> test classpath.
> h3. Background
> {{TrinoExtractorIT}} contained a placeholder list of test cases with no
> implementation. The Trino extractor needs automated verification for metadata
> import, stale-entity cleanup, Hive hook bridging, and CLI/cron validation
> before release.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
