[ https://issues.apache.org/jira/browse/HBASE-19923?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16358722#comment-16358722 ]
Hudson commented on HBASE-19923: -------------------------------- FAILURE: Integrated in Jenkins build HBase-Trunk_matrix #4555 (See [https://builds.apache.org/job/HBase-Trunk_matrix/4555/]) HBASE-19923 Reset peer state and config when refresh replication source (zghao: rev 2b63af376edbca53eb0475e929478bd38b580aae) * (edit) hbase-server/src/main/java/org/apache/hadoop/hbase/replication/regionserver/PeerProcedureHandlerImpl.java * (edit) hbase-replication/src/main/java/org/apache/hadoop/hbase/replication/ReplicationPeerImpl.java > Reset peer state and config when refresh replication source failed > ------------------------------------------------------------------ > > Key: HBASE-19923 > URL: https://issues.apache.org/jira/browse/HBASE-19923 > Project: HBase > Issue Type: Bug > Components: Replication > Affects Versions: 3.0.0 > Reporter: Guanghao Zhang > Assignee: Guanghao Zhang > Priority: Major > Attachments: HBASE-19923.master.001.patch, > HBASE-19923.master.001.patch, HBASE-19923.master.002.patch, > HBASE-19923.master.003.patch > > > Now we use procedure for replication. When peer state changed, the RS will > read peer state from storage to cache. If RS found the peer state changed, > then it will refresh replication source. If refresh failed, the Master will > retry the procedure. Then RS will read peer state again, but now the peer > state in cache is right. So it don't refresh replication source...... So we > need reset the peer state to old peer state when refresh failed. -- This message was sent by Atlassian JIRA (v7.6.3#76005)