voonhous opened a new issue, #19954:
URL: https://github.com/apache/hudi/issues/19954

   The Trino E2E workflow (`hudi_trino_e2e.yml`) has not run its tests on 
master since #19642, but every run still reports success.
   
   #19642 moved hudi-trino from released Trino builds to trinodb/trino master 
at a pinned SHA (`trino.sha`). The E2E image kept using the released 
`trinodb/trino:<trino.e2e.version>` server, so it added an SPI drift gate that 
skips every later step when SPI or filesystem paths changed between the release 
and the pin.
   
   Master tracks Trino master, so the gate always fires. The current pin 
(484-SNAPSHOT, `40b70400a8a6`) is 29 boundary commits past 483, and the 
pin-advance bot moves it further each time.
   
   | Run | Result |
   |---|---|
   | 
[2026-08-14](https://github.com/apache/hudi/actions/runs/31783295622/job/94713532009),
 before #19642 | tests ran, 26 min |
   | 
[2026-09-08](https://github.com/apache/hudi/actions/runs/34191374126/job/101949963378),
 #19863 merge | gate skipped all steps, 11 s, green |
   
   Breakages now show up only on release branches, which pin back to a released 
Trino. The Docker Hub MinIO removal, for example, surfaced first on 
`release-1.2.1-staging`.
   
   Proposed fix: build the E2E server image from the pinned commit so server 
and connector always match, and remove the drift gate.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to