Open a8356555 opened 11 months ago
@a8356555 have a similar issue - how did you unblock yourself?
I am seeing a very similar issue when trying to write to S3 from pyflink using a query like the following.
INSERT INTO iceberg.devdb.table_name
SELECT * FROM transformed_metrics;
I can write to S3 fine with a manual/static insert of a single row using pyflink, but when I do this it silently fails. I'm unclear if I'm just using the platform wrong or if there's a bug. Since there are no glaring issues in the logs.
@parrik: The issue you have linked is for the cases where the checkpointing is not enabled. If I read the python code correctly, this is not the case here.
Still, the issue seems similar, likely the checkpointing is not triggered for the committer. Maybe the missing part is, that the notifyCheckpointComplete
is not called.
It would be good to see the DEBUG logs for the IcebergFilesCommitter
to see what is happening.
Apache Iceberg version
1.4.0
Query engine
Flink
Please describe the bug 🐞
I'm using following Dockerfile as my environment:
Here is my pyflink code (job.py)
I'm running this code using
However, the job never write data into my iceberg in s3 here is the dashboard
Why is Iceberg StreamWriter stuck? There's no error in the jobmanager's log