[SUPPORT] INSERT_OVERWRITE_TABLE on subsequent runs fails with a metadata file not found error (v0.14.0) #10445
Labels
on-call-triaged
priority:critical
production down; pipelines stalled; Need help asap.
writer-core
Issues relating to core transactions/write actions
Describe the problem you faced
I recently upgraded some of my AWS Glue processes to use Hudi v0.14.0, and the process overwrites an entire table on every run. The initial load runs just fine, but on every subsequent run, I'm getting a metadata error like below:
java.io.FileNotFoundException: No such file or directory 's3://bucket-name/5eda3c2b-38c0-4f3f-8163-8a9a2c88f8a4-0_0-22-75_20240104030715270.parquet'
The only changes to the settings from the older version used in Glue (v0.12.1) are:
Full Hoodie settings:
To Reproduce
Steps to reproduce the behavior:
INSERT_OVERWRITE_TABLE
, and the Hudi settings aboveExpected behavior
The table data should've been overwritten
Environment Description
Hudi version : 0.14.0
Spark version : 3.3.0 (AWS Glue 4.0)
Storage (HDFS/S3/GCS..) : S3
Running on Docker? (yes/no) : No, on AWS
Additional context
Add any other context about the problem here.
Stacktrace
The text was updated successfully, but these errors were encountered: