[BUG] CalcitePPLTpchIT is flaky

[Swiddis]

What is the bug?

TestQ1 is occasionally failing in main, e.g. https://github.com/opensearch-project/sql/actions/runs/17594180049/job/49982291692.

CalcitePPLTpchIT > testQ1 FAILED
    java.lang.AssertionError: 
    Expected: iterable with items [[], [], [], []] in any order
         but: not matched: <["A","F",37474,37569624.64,35676192.096999995,37101416.222424,25.354533152909337,25419.231826792962,0.0508660351826793,1478]>
        at __randomizedtesting.SeedInfo.seed([7DBF59379F8C433E:DF95364E8E9BEC79]:0)
        at org.hamcrest.MatcherAssert.assertThat(MatcherAssert.java:18)
        at org.hamcrest.MatcherAssert.assertThat(MatcherAssert.java:6)
        at org.opensearch.sql.util.MatcherUtils.verify(MatcherUtils.java:193)
        at org.opensearch.sql.util.MatcherUtils.verifyDataRows(MatcherUtils.java:160)
        at org.opensearch.sql.calcite.tpch.CalcitePPLTpchIT.testQ1(CalcitePPLTpchIT.java:59)

At a quick glance, seems like data is being fetched before a refresh runs.

How can one reproduce the bug?
With great care.

What is the expected behavior?
Should pass consistently.

What is your host/environment?

• OS: Mac OS 13
• Java 24

Do you have any screenshots?
N/A

Do you have any additional context?
N/A

[Swiddis]

Attaching an example failure log for posterity: logs_45137514781.zip

[LantaoJin]

Not seen this issue any more.