[Bug] flaky case pax_inject_fault

### Apache Cloudberry version

_No response_

### What happened

another flaky case: https://github.com/apache/cloudberry/actions/runs/15015871210/job/42200944780?pr=1098

```diff

Found regression.diffs at: ./contrib/pax_storage/regression.diffs
diff -I HINT: -I CONTEXT: -I GP_IGNORE: -U3 /__w/cloudberry/cloudberry/contrib/pax_storage/expected/pax_inject_fault.out /__w/cloudberry/cloudberry/contrib/pax_storage/results/pax_inject_fault.out
--- /__w/cloudberry/cloudberry/contrib/pax_storage/expected/pax_inject_fault.out 2025-05-14 03:20:02.329295420 -0700
+++ /__w/cloudberry/cloudberry/contrib/pax_storage/results/pax_inject_fault.out 2025-05-14 03:20:02.330295422 -0700
@@ -24,7 +24,10 @@

-- failed because of fault injection
insert into t_insert select generate_series(1,10);
+DETAIL: The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
ERROR: fault triggered, fault name:'orc_writer_write_tuple' fault type:'panic'
+HINT: In a moment you should be able to reconnect to the database and repeat your command.
+WARNING: terminating connection because of crash of another server process
GP_IGNORE:-- start_ignore
GP_IGNORE:-- clear the fault inject, so the next insert will success.
GP_IGNORE:-- put the reset operation in ignore range
```

Lack of enough resources?

### What you think should happen instead

_No response_

### How to reproduce

CI

### Operating System

CI

### Anything else

_No response_

### Are you willing to submit PR?

- [ ] Yes, I am willing to submit a PR!

### Code of Conduct

- [x] I agree to follow this project's [Code of Conduct](https://github.com/apache/cloudberry/blob/main/CODE_OF_CONDUCT.md).