joshelser commented on PR #4338: URL: https://github.com/apache/hbase/pull/4338#issuecomment-1098257270
> That change does not allow concurrent compaction against a given store, respecting that Compactor is not thread safe for now. It works well. The performance of the test scenario is unchanged from baseline without any SFT changes. As an option to unblock us we could use it for now and come back to the implementation issues with SFT and compaction in a follow up issue. I agree with your feelings, Andrew. Correctness first and then optimization. > we do allow concurrent compactions happen at the same time in the past... Acknowledging this too: yes, we should be able to compact two distinct subsets of the files in a store concurrently. And, I could see value in doing so (e.g. compacting three larger files in a store and also wanting to compact a few smaller hfiles created from memstore flushes -- we don't want to wait for that big compaction to finish to run the smaller one). -- 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]
