[ https://issues.apache.org/jira/browse/HBASE-12988?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14543201#comment-14543201 ]
Lars Hofhansl commented on HBASE-12988: --------------------------------------- That'd save CPU on the replication source. I have the feeling that would not do much to speed things up. What's slow (and I say this without having verified in detail) is the sequential reading of the WAL and non-streaming writes to the replication sink with absolutely no parallelism. Maybe I'll put instrumentation in to measure exactly where it's slow, but I am fairly certain it is shipping the edit to the sink and applying them there. > [Replication]Parallel apply edits on row-level > ---------------------------------------------- > > Key: HBASE-12988 > URL: https://issues.apache.org/jira/browse/HBASE-12988 > Project: HBase > Issue Type: Improvement > Components: Replication > Reporter: hongyu bi > Assignee: hongyu bi > > we can apply edits to slave cluster in parallel on table-level to speed up > replication . > update : per conversation blow , it's better to apply edits on row-level in > parallel -- This message was sent by Atlassian JIRA (v6.3.4#6332)