dlmarion commented on pull request #2330: URL: https://github.com/apache/accumulo/pull/2330#issuecomment-952977486
I wonder if there is a consequence to performing compactions on tablets with a greater number of files first versus using a random distribution (or some other scheme). I'm thinking about the case where you have unbalanced ingest into a table, will you run into a case where compactions never happen for tablets that have a light amount of ingest because the tablets that have a large amount of ingest activity are constantly compacting? Same problem for multiple tables being assigned to the same compaction service. To be clear, my concern is starvation and the users inability to change it. I wonder, if instead of using some hardcoded priority, we let the user decide a mode to use instead. For example: FIFO - compact tablets that meet the compaction ratio threshold in time order LIFO - compact tablets that meet the compaction ratio threshold in reverse time order TOTAL_FILES - compact tablets with the largest number of files first (the current approach) MINIMUM_RESULTING_FILES - the approach contained in this PR etc... -- 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]
