[ https://issues.apache.org/jira/browse/HBASE-2462?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12923181#action_12923181 ]
stack commented on HBASE-2462: ------------------------------ Looking at production regionserver compacting with J-D, we're just doing too much compacting. Its running near back-to-back. Need to fix. Our default compaction trips and count of files seems too little. Also, it looks like we'll merrily go ahead and compact together a 10MB file with a 1G file; we need to have in place barriers to this ever happening. This issue is important to fix.. but yeah, 0.92. > Review compaction heuristic and move compaction code out so standalone and > independently testable > ------------------------------------------------------------------------------------------------- > > Key: HBASE-2462 > URL: https://issues.apache.org/jira/browse/HBASE-2462 > Project: HBase > Issue Type: Improvement > Reporter: stack > Priority: Critical > > Anything that improves our i/o profile makes hbase run smoother. Over in > HBASE-2457, good work has been done already describing the tension between > minimizing compactions versus minimizing count of store files. This issue is > about following on from what has been done in 2457 but also, breaking the > hard-to-read compaction code out of Store.java out to a standalone class that > can be the easier tested (and easily analyzed for its performance > characteristics). > If possible, in the refactor, we'd allow specification of alternate merge > sort implementations. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.