[Bug] Parallelism cannot be controlled in `remove_orphan_files`

### Search before asking

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

### Paimon version

1.3.1

### Compute Engine

flink 1.20.3

### Minimal reproduce step

```sql
CALL sys.remove_orphan_files(`table` => 'ods_mms_db.*', older_than => '2026-01-29 10:00:00', parallelism => 10, dryRun => true);
```
The database contains more than 600 tables.

### What doesn't meet your expectations?

The parallelism of the job should be 10.
The actual parallelism is much higher than 10, which will exhaust the cluster resources.

### Anything else?

_No response_

### Are you willing to submit a PR?

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

Start with the sys.remove_orphan_files procedure and reproduce the call against a database with more than 600 tables, using parallelism => 10 and dryRun => true. Trace how the procedure schedules work and verify that the resulting job parallelism is limited to 10 without exhausting cluster resources.

Written by the indexing model from the issue text.