The GitHub Actions job "Backport Approval Check" on 
texera.git/sshiv012/surface-region-launch-failures has failed.
Run started by GitHub user Yicong-Huang (triggered by Yicong-Huang).

Head commit for run:
f932270f62a86583b7a5d56a11e06bcc38f3867c / suryaacharan 
<[email protected]>
fix(amber): surface region launch failures instead of hanging

When a region has dependee input ports (e.g. a HashJoin probe), its
non-dependee operators are launched from a later advance round via
syncStatusAndTransitionRegionExecutionPhase(). advanceRegionExecutions
discarded those futures, so if a worker failed there (e.g. a Python UDF
whose code raises NameError at load time) the error was lost: no
FatalError reached the client, no operator in the region started, and
the execution stayed RUNNING forever.

- Keep the sync futures and fail the advance future if any fails, so
  AdvanceRegionExecutionsHandler reports a FatalError.
- Log ControlError contents in AsyncRPCClient; it is not a Throwable, so
  the logger silently dropped it.
- Tests: manager unit tests for a failing/succeeding second phase, and
  e2e csv->join->Python UDF tests, including a UDF missing
  `from pytexera import *`, which now fails in ~1s instead of timing out.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

Report URL: https://github.com/apache/texera/actions/runs/36073663792

With regards,
GitHub Actions via GitBox

Reply via email to