SEPURI-SAI-KRISHNA commented on issue #12116:
URL: https://github.com/apache/seatunnel/issues/12116#issuecomment-5595369250
Two more occurrences for the record, both on #12182, and one detail that may
narrow the cause.
Fork run `34135130164` on head `08b4cd2a7`, same test and same symptom as
the three cases in the description,
`TestFilterRowKindIT.testFilterRowKindMultiTable:53 expected: <0> but was: <1>`:
- attempt 1, job `101784974271`, `transform-v2-it-part-1 (11,
ubuntu-latest)`: failed on `{TestContainer}[3]`, which is
`tyrantlucifer/flink:1.18.0-scala_2.12_hadoop27`.
- attempt 3, job `102038075593`, same job name, same head: failed on
`{TestContainer}[2]`, which is `tyrantlucifer/flink:1.15.3-scala_2.12_hadoop27`.
The detail that seems worth having on record: this is the same unchanged
commit failing twice across reruns, and it moved between Flink minor versions
rather than staying on one. In both runs the Zeta leg and both Spark legs of
the same test passed, and the `transform-v2-it-part-1 (8, ubuntu-latest)` leg
passed. That is consistent with the shared-static-counter diagnosis in the
description and against anything data-dependent or version-specific, since the
config, the data and the head are identical across the two failures and only
the leg that lost the race changed.
For completeness on relevance: #12182 changes `NumericFunction.mod` and two
unit test classes under `seatunnel-transforms-v2`. It touches nothing in
`connector-assert`, nothing in the `FilterRowKind` transform, and nothing in
`seatunnel-e2e`.
Anecdotally the rate looks high rather than rare. Two failures in two
executions of that job on the same head, so a rerun has not been a reliable way
around it.
--
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]
