satybald commented on pull request #15185: URL: https://github.com/apache/beam/pull/15185#issuecomment-893912744
The fast Avro pipeline finished. And it indeed way faster than python Avro version. It's has the same ~5GiB/s thought put as a regular batch Extract job. So, I believe we're here worker bound(however, it's an assumption that would be nice to back up with data) **Elapsed time** Batch Extract Job - 58 min BQ Storage with Fast Avro - 1 hours 27 min But in terms of elapsed time, it got 30min slower. I believe this case because, the job has 3 GRPC errors. Thus, the master had to fail work item and retry on the different place. Each such fail contributed to ~10 min to the total execution time.  -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
