[ https://issues.apache.org/jira/browse/HBASE-20475?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16454202#comment-16454202 ]
Zheng Hu commented on HBASE-20475: ---------------------------------- Found an unrelated bug in RecoveredReplicationSourceShipper#getRecoveredQueueStartPos() {code} private long getRecoveredQueueStartPos() { long startPosition = 0; String peerClusterZnode = source.getQueueId(); try { startPosition = this.replicationQueues.getWALPosition(source.getServerWALsBelongTo(), peerClusterZnode, this.queue.peek().getName()); if (LOG.isTraceEnabled()) { LOG.trace("Recovered queue started with log " + this.queue.peek() + " at position " + startPosition); } } catch (ReplicationException e) { terminate("Couldn't get the position of this recovered queue " + peerClusterZnode, e); } return startPosition; } {code} When we start run to RecoveredReplicationSourceShipper, all WALs of dead server has been pushed into the new RS's queue, So it will always return -1 for following method, because the path does not exist any more . {code} startPosition = this.replicationQueues.getWALPosition(source.getServerWALsBelongTo(), peerClusterZnode, this.queue.peek().getName()); {code} > Fix the flaky TestReplicationDroppedTables unit test. > ----------------------------------------------------- > > Key: HBASE-20475 > URL: https://issues.apache.org/jira/browse/HBASE-20475 > Project: HBase > Issue Type: Bug > Affects Versions: 2.1.0 > Reporter: Zheng Hu > Assignee: Zheng Hu > Priority: Major > Fix For: 3.0.0, 2.1.0 > > Attachments: HBASE-20475-addendum.patch, HBASE-20475.patch > > > See > https://builds.apache.org/job/HBASE-Find-Flaky-Tests/lastSuccessfulBuild/artifact/dashboard.html -- This message was sent by Atlassian JIRA (v7.6.3#76005)