vburenin commented on issue #2887: URL: https://github.com/apache/hudi/issues/2887#issuecomment-827670147
@rancylive can you take a look at spark execution plan, take snapshot how long each stage takes and print it out here? Also, would be interesting to know how many partitions you end up with. What is the CPU load on the executors during each stage? On my side every time I observe any such slowness without a meaningful CPU load it is typically IO bottleneck somewhere. You also may want to upgrade to Hudi 0.8.0 as it has a few several GCS related issues addressed. May not be the case for this specific issue, but still would be nice to exclude this factor. -- 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. For queries about this service, please contact Infrastructure at: [email protected]
