[SUPPORT] is not a Parquet file (length is too low: 0)

**Environment:**

Hudi MOR+ Spark

**Hudi version :**

0.12.1

**Spark version :**

spark3

**Hadoop version:**

2.8.5

**Problem:**

java.lang.RuntimeException: hdfs://test/warehouse/test.db/t_test/20220811/b4a54eb9-0e0c-406a-a2bd-c2f07b021277-0_8-10-991_20230421093519585.parquet is not a Parquet file (length is too low: 0)

**SQL:**

`insert into table_a_rt select * from table_b_rt;`

Start by reproducing the reported failure in the Hudi MOR + Spark environment using `insert into table_a_rt select * from table_b_rt;`. Inspect how the referenced zero-length `.parquet` file is produced or read, then confirm the issue is resolved when the insert no longer raises the reported Parquet error; no source file or test is identified in the issue.

Written by the indexing model from the issue text.