DcAwarePolicyTest.Integration_Cassandra_AllowRemoteDcsForLocalCl is flakky

Sometimes it fails:

[ RUN      ] DcAwarePolicyTest.Integration_Cassandra_AllowRemoteDcsForLocalCl
1758628238072 [DEBUG] (scylla_cpp_driver::logging:178) message: Log level is set to TRACE
ccm> ccm list
ccm> ccm list
Testing with `allow_remote_dcs_for_local_cl`=`false`
Running nonlocal statement with local DC available
Running local statement with local DC available
ccm> ccm node1 stop --not-gently
ccm> ccm node2 stop --not-gently
Running nonlocal statement with local DC unavailable
/home/runner/work/cpp-rs-driver/cpp-rs-driver/tests/src/integration/objects/future.hpp:152: Failure
      Expected: CASS_OK
To be equal to: wait_code
      Which is: CASS_ERROR_SERVER_READ_TIMEOUT [Read timeout]
Read timeout: Database returned an error: Not enough nodes responded to the read request in time to satisfy required consistency level (consistency: Two, received: 1, required: 2, data_present: true), Error message: Operation timed out - received only 1 responses.
unknown file: Failure
C++ exception with description "Attempted to use null object" thrown in the test body.
ccm> ccm node1 start --wait-other-notice --wait-for-binary-proto
ccm> ccm node2 start --wait-other-notice --wait-for-binary-proto
[  FAILED  ] DcAwarePolicyTest.Integration_Cassandra_AllowRemoteDcsForLocalCl (37067 ms)

Start with DcAwarePolicyTest.Integration_Cassandra_AllowRemoteDcsForLocalCl and reproduce it using the ccm node stop/start sequence in the report. Inspect tests/src/integration/objects/future.hpp at line 152 and the surrounding integration test code to determine why the read timeout and null-object exception occur; done means the test reliably passes when run repeatedly.