[BP-1.20][FLINK-40070][tests] Fix flaky DynamicParameterITCase reading rolled JobManager logs

[MartijnVisser]

What is the purpose of the change

Backport of #28636 (f9e0756fbfd) to release-1.20.

The distribution rolls its log file during JobManager startup, so the startup banner with the
program arguments frequently ends up in a rolled .log.N file that
FlinkDistribution.searchAllLogs skips. DynamicParameterITCase then waits for that banner in
an unbounded loop, and the class has no @Timeout, so the whole Java e2e leg runs until the CI
watchdog kills it at its time budget. On release-2.3 that took out E2E group 2 on the 2026-09-04,
09-11, 09-15 and 09-16 nightlies. The readiness loop, FlinkDistribution and the distribution
log4j configuration are identical on this branch.

Brief change log

• FlinkDistribution.searchAllLogs gains an includeRolledLogs overload, so the banner is
  found when it has been rolled. The existing two-argument method delegates with false, so
  the other callers are unchanged.
• DynamicParameterITCase waits with CommonTestUtils.waitUtil(..., 1 minute, 500 ms, msg)
  instead of while (!allProgramArgumentsLogged(dist)) Thread.sleep(500), so a banner that
  never arrives fails in a minute with a message rather than consuming the leg's budget.

Verifying this change

FlinkDistribution.java is byte-identical to the master commit. DynamicParameterITCase.java was
resolved by hand: release-1.20 passes the host and port positionally and asserts on
EntrypointClusterConfiguration#getHostname and #getRestPort, so the argument building and the
assertions are left alone. Master's loadConfiguration helper and its Configuration based
assertions come from a later change that is not on this branch, and are not introduced here. The
three helpers the fix adds or rewrites are byte-identical to master's.

DynamicParameterITCase green twice against a built 1.20 distribution, 6 tests in 29.5 s and
29.8 s. spotless and checkstyle green under JDK 11, since this branch pins google-java-format 1.7
rather than 1.24.0.

Does this pull request potentially affect one of the following parts:

• Dependencies (does it add or upgrade a dependency): no
• The public API, i.e., is any changed class annotated with @Public(Evolving): no
• The serializers: no
• The runtime per-record code paths (performance sensitive): no
• Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: no
• The S3 file system connector: no

Documentation

• Does this pull request introduce a new feature? no

---

Was generative AI tooling used to co-author this PR?

• Yes (Claude Opus 5, via Claude Code)

Generated-by: Claude Opus 5 (1M context)

[flinkbot]

CI report:

• 676e53f Azure: FAILURE

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot run azure re-run the last Azure build

[spuru9]

Clean BP. LGTM.