[ https://issues.apache.org/jira/browse/HBASE-16417?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15480321#comment-15480321 ]
Anastasia Braginsky commented on HBASE-16417: --------------------------------------------- Hi again! As I started to explain in my previous comment, we believe this JIRA should be split into two. I have opened HBASE-16608 for this case. In HBASE-16608 we want to present the merge of the immutable segments of the pipeline into one new segment, while this is done without data-copy-compaction and without SQM. In this JIRA we plan to continue after HBASE-16608 and finalize the tuning of the in-memory flush policy for the best performance. All the performance issues are still meant to be resolved here. HBASE-16608 is for one more "algorithmic issue", not so sophisticated of course, but should resolve us the issues of two many segments (and later files) in the compaction pipeline. I would like to hear your opinion about all that! For more details look on HBASE-16608. Waiting for your fruitful thoughts! Thanks! > In-Memory MemStore Policy for Flattening and Compactions > -------------------------------------------------------- > > Key: HBASE-16417 > URL: https://issues.apache.org/jira/browse/HBASE-16417 > Project: HBase > Issue Type: Sub-task > Reporter: Anastasia Braginsky > Assignee: Anastasia Braginsky > Fix For: 2.0.0 > > -- This message was sent by Atlassian JIRA (v6.3.4#6332)