Test is unstable because waitUntilPartitions timeout

### Search before asking

- [x] I searched in the [issues](https://github.com/alibaba/fluss/issues) and found nothing similar.

### Fluss version

0.6.0 (latest release)

### Please describe the bug 🐞

Even in one day, there are many times that the test fail with "expect 2 table partition has not been created"

For example:
In https://github.com/alibaba/fluss/actions/runs/14839399255/job/41658380202,
```java
Error: Failures:
Error: Flink120TableSinkITCase>FlinkTableSinkITCase.testWritePartitionedTable:543 expect 2 table partition has not been created
```

In https://github.com/alibaba/fluss/actions/runs/14836873591/job/41650048948 ,
```java
Error: Failures:
Error: Flink120TableSourceBatchITCase>FlinkTableSourceBatchITCase.testLimitLogTableScan:274->FlinkTableSourceBatchITCase.preparePartitionedLogTable:429->FlinkTestBase.waitUntilPartitions:209->FlinkTestBase.waitUntilPartitions:221 expect 2 table partition has not been created

```

### Solution

_No response_

### Are you willing to submit a PR?

- [x] I'm willing to submit a PR!

Start with FlinkTestBase.waitUntilPartitions, especially the lines reported in the failures, and inspect the callers FlinkTableSinkITCase.testWritePartitionedTable and FlinkTableSourceBatchITCase.preparePartitionedLogTable. Reproduce the failing integration tests and determine why the timeout is insufficient or unstable; done means the affected tests reliably observe both table partitions without intermittent timeout failures.

Written by the indexing model from the issue text.