Gatsby-Lee commented on issue #4839: URL: https://github.com/apache/hudi/issues/4839#issuecomment-1055016784
@glory9211 Like @nsivabalan said, if your data latency is not critical, then CoW works as well. I use CoW for some datasets that have incoming data around 600-800/sec. I designed to use MoR initially, but due to the Async Compaction issue in AWS Glue, I moved to CoW for now. If the data latency is not big concern, then try with the increased windowSize. You can set the windowSize when you setup forEachBatch. What I learned is that as Big Data, Hudi write has high latency + high throughput. So, rather than writing many small batch, try with a bigger size batch with less write. I believe that you can get some different ideas @glory9211 Good question. Yes. I'd like to ask the same that if Async Clustering is triggered or run in AWS Glue when it is enabled. @nsivabalan Does the Async Clustering has to run independently? -- 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]
