[ https://issues.apache.org/jira/browse/HBASE-14004?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15050165#comment-15050165 ]
Phil Yang commented on HBASE-14004: ----------------------------------- I have revised the doc https://docs.google.com/document/d/1tTwXrip18qxxsSiPu_y4fGB-26SxqwMvZnFREswyQow/edit?usp=sharing which removes the logic about hsync. And previously the doc quotes Duo's comment {quote} close old file using any length that larger than the previous succeeded hflushed length {quote} I changed this to close the file using the previous succeeded hflushed length which is more clear for implementation. And in section "Some issues still need discussion" I quote [~carp84]'s comment: {quote} We also need to consider the cross-RS case when region assign happens before wal sync acked. {quote} Yu, is it still a problem after we not using hsync? > [Replication] Inconsistency between Memstore and WAL may result in data in > remote cluster that is not in the origin > ------------------------------------------------------------------------------------------------------------------- > > Key: HBASE-14004 > URL: https://issues.apache.org/jira/browse/HBASE-14004 > Project: HBase > Issue Type: Bug > Components: regionserver > Reporter: He Liangliang > Priority: Critical > Labels: replication, wal > > Looks like the current write path can cause inconsistency between > memstore/hfile and WAL which cause the slave cluster has more data than the > master cluster. > The simplified write path looks like: > 1. insert record into Memstore > 2. write record to WAL > 3. sync WAL > 4. rollback Memstore if 3 fails > It's possible that the HDFS sync RPC call fails, but the data is already > (may partially) transported to the DNs which finally get persisted. As a > result, the handler will rollback the Memstore and the later flushed HFile > will also skip this record. -- This message was sent by Atlassian JIRA (v6.3.4#6332)