Jason Fehr has posted comments on this change. ( http://gerrit.cloudera.org:8080/23385 )
Change subject: IMPALA-14401: Deflake/Improve OpenTelemetry Tracing Tests ...................................................................... Patch Set 5: (3 comments) http://gerrit.cloudera.org:8080/#/c/23385/4/tests/custom_cluster/test_otel_trace.py File tests/custom_cluster/test_otel_trace.py: http://gerrit.cloudera.org:8080/#/c/23385/4/tests/custom_cluster/test_otel_trace.py@547 PS4, Line 547: def test_retry_select_failed(self): > This test seems flaky in my machine. Can you please loop it and see if you This is a known issue, but I only ever saw it during the QueryExecution span. Makes sense that it could hit the Close span too though. Events are no longer asserted in the Close span. IMPALA-14334 will take care of fixing the asserts for the span events for both QueryExecution and Close spans. http://gerrit.cloudera.org:8080/#/c/23385/4/tests/custom_cluster/test_otel_trace.py@691 PS4, Line 691: def test_ddl_createtable_fail(self, vector, unique_name): > This flaky to me when loop it by itself. Can you try loop it in your machin I'm puzzled by this one. I was unable to replicate the issue. The 'parse_session_id' function is called on each individual test (when the test calls 'assert_trace'). I've never seen this failure happen in all the tests runs I've done locally and in Jenkins builds. I don't think the '--exploration=exhaustive' flag should have any impact on this test. Claude Sonnet suggested the issue is because the call to `self.query_id_from_ui` on lines 698-699 allows the query to not be found. In that case, profile will be None. That guess seems very likely to be the case. Thus, I added a new function 'query_id_from_ui_retry' that will try repeatedly (for 30 seconds) to get the query id and profile from the UI debug webpage. http://gerrit.cloudera.org:8080/#/c/23385/4/tests/util/otel_trace.py File tests/util/otel_trace.py: http://gerrit.cloudera.org:8080/#/c/23385/4/tests/util/otel_trace.py@588 PS4, Line 588: return span_log.group(2), span_log.group(1) : > nit: It can be helpful to print few last lines of log_file_path before rais Added logging to output the log file being read. That will be very helpful when troubleshooting build failures. -- To view, visit http://gerrit.cloudera.org:8080/23385 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I0c3e0075df688c7ae601c6f2e5743f56d6db100e Gerrit-Change-Number: 23385 Gerrit-PatchSet: 5 Gerrit-Owner: Jason Fehr <[email protected]> Gerrit-Reviewer: Abhishek Rawat <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Jason Fehr <[email protected]> Gerrit-Reviewer: Michael Smith <[email protected]> Gerrit-Reviewer: Riza Suminto <[email protected]> Gerrit-Comment-Date: Tue, 09 Sep 2025 22:53:19 +0000 Gerrit-HasComments: Yes
