Ryan Faircloth

I'm experiencing this same issue, and I've set the job to use `Spark 2.4, Python 2 (Glue Version 1.0)`. If there are any suggestions/ tips on debugging this issue that...

Hi @dacort. Thanks for getting back to me. I do see that message: ``` INFO:athena_glue_service_logs.catalog_manager:Creating database table s3_access_optimized null_fields [] INFO:athena_glue_service_logs.converter:No data returned, skipping conversion. INFO:athena_glue_service_logs.job:Initial run with source NullPartitioner,...

@dacort The access logs are using another prefix `s3:///service=s3/bucket=/06-05-2021-alongstring` But I did test with specifying a more specific path (e.g. `s3:///service=s3/bucket=` as the `--s3_source_location`. Yes, I can query the `s3_access_raw`...

I just tried this with version bazel version 1.1.0 and I have the same results.

@uhthomas I did sort of get it to work. I mean it worked, but I had to re-disable the sandbox mode that bazel uses. I started looking at bazel's [sandboxfs](https://blog.bazel.build/2017/08/25/introducing-sandboxfs.html),...

@uhthomas I think this is a mixture of both hugo and bazel. Bazel is choosing to fallback on creating a sandbox using symlinks to run everything from what it considers...