[ https://issues.apache.org/jira/browse/HBASE-23634?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17161899#comment-17161899 ]
Guanghao Zhang edited comment on HBASE-23634 at 7/21/20, 9:51 AM: ------------------------------------------------------------------ {quote}What if the split is stopped in between. (WAL split RS died).. Also there may be other systems seeing directly the files under CF (Like offline compaction tool). How we can give a consistent view of the CF files to these other systems. {quote} For the WAL replay, it replay edits and may flush many times if the memstore is big. And if RS died before finishing replay, the other systems cannot get a consistent view of the CF files, too? So I thought HBase doesn't have this guarantee before. was (Author: zghaobac): {quote}What if the split is stopped in between. (WAL split RS died).. Also there may be other systems seeing directly the files under CF (Like offline compaction tool). How we can give a consistent view of the CF files to these other systems. {quote} For the WAL replay, it replay edits and may flush many times if the memstore is big. And if RS died before replay finished, the other systems cannot get a consistent view of the CF files, too? > Enable "Split WAL to HFile" by default > -------------------------------------- > > Key: HBASE-23634 > URL: https://issues.apache.org/jira/browse/HBASE-23634 > Project: HBase > Issue Type: Task > Affects Versions: 3.0.0-alpha-1, 2.3.0 > Reporter: Guanghao Zhang > Priority: Blocker > Fix For: 3.0.0-alpha-1, 2.4.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)