SEPURI-SAI-KRISHNA commented on PR #12182:
URL: https://github.com/apache/seatunnel/pull/12182#issuecomment-5578992974
Thanks for the detailed review @DanielLeens. One correction on the CI
reading, and then agreement on the conclusion.
The run has **two** failing jobs, not one. Alongside `engine-v2-it (11,
ubuntu-latest)` there is `transform-v2-it-part-1 (11, ubuntu-latest)`, which is
the more relevant of the two since it is the e2e module for the code this PR
changes. `kudu-connector-it (8, ubuntu-latest)` shows `cancelled` as
collateral. Overall: 80 passed, 2 failed, 1 cancelled, 11 skipped.
**`transform-v2-it-part-1`**:
`TestFilterRowKindIT.testFilterRowKindMultiTable:53 expected: <0> but was:
<1>`. This one is worth ruling out on its own terms rather than by module
boundary, because it does sit in the transforms e2e module. Its config,
`filter_row_kind_exclude_insert_multi_table.conf`, is `FakeSource` to
`FilterRowKind` to `Assert`, with no SQL transform in it at all.
`NumericFunction.mod` is reachable only through the Zeta SQL transform's
function dispatch, so nothing in this diff is on that job's path even in
principle.
It is also an already known flake. #12116, filed by @davidzollo on 09-05,
names this exact test and this exact config and lists three unrelated PRs
hitting it, including #11503 with the identical `expected: <0> but was: <1>`
exit-code symptom. The diagnosis there is `AssertSinkWriter`'s static JVM-wide
row counters being evaluated per subtask in `close()`.
The failing leg matches that issue too. `{TestContainer}[3]` is the third
template invocation in this run, which is
`tyrantlucifer/flink:1.18.0-scala_2.12_hadoop27`, and its 24.2 s elapsed time
lines up with that invocation's window in the log. Flink 1.18.0 is one of the
two legs #12116 already documents, and the Zeta and both Spark legs passed in
this same run, which is exactly the Flink-only pattern described there. So this
is another instance of #12116 rather than anything new.
**`engine-v2-it`**: confirmed as you described,
`SinkErrorToMysqlIT.testSinkMaxErrorRatioThreshold:605` with a
`ConditionTimeout` from an Awaitility `untilAsserted`, in `seatunnel-engine`'s
e2e base. 186 tests, 1 error.
Agreed on the rest. Both `unit-test` legs are green, including this PR's two
new tests. I am rerunning the two failed jobs rather than syncing with `dev`,
for the reason you gave: `dev` has no commits touching
`seatunnel-transforms-v2` since this branch point, so a sync would change
neither outcome.
--
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]
