[HBASE-28076] NPE on initialization error in RecoveredReplicationSourceShipper

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.6.0, 2.4.17, 2.5.5
• Fix Version/s: 2.6.0, 2.4.18, 2.5.6
• Hadoop Flags: Reviewed
• Assignee: Istvan Toth
• Reporter: Istvan Toth
• Created: 12/Sep/23 13:14
• Updated: 18/Sep/23 04:55
• Resolved: 14/Sep/23 23:57

Description

When we run into problems starting RecoveredReplicationSourceShipper, we try to stop the reader thread which we haven't initialized yet, resulting in an NPE.

ERROR org.apache.hadoop.hbase.replication.regionserver.ReplicationSource: Unexpected exception in redacted currentPath=hdfs://redacted
java.lang.NullPointerException
        at org.apache.hadoop.hbase.replication.regionserver.RecoveredReplicationSourceShipper.terminate(RecoveredReplicationSourceShipper.java:100)
        at org.apache.hadoop.hbase.replication.regionserver.RecoveredReplicationSourceShipper.getRecoveredQueueStartPos(RecoveredReplicationSourceShipper.java:87)
        at org.apache.hadoop.hbase.replication.regionserver.RecoveredReplicationSourceShipper.getStartPosition(RecoveredReplicationSourceShipper.java:62)
        at org.apache.hadoop.hbase.replication.regionserver.ReplicationSource.lambda$tryStartNewShipper$3(ReplicationSource.java:349)
        at java.util.concurrent.ConcurrentHashMap.compute(ConcurrentHashMap.java:1853)
        at org.apache.hadoop.hbase.replication.regionserver.ReplicationSource.tryStartNewShipper(ReplicationSource.java:341)
        at org.apache.hadoop.hbase.replication.regionserver.ReplicationSource.initialize(ReplicationSource.java:601)
        at java.lang.Thread.run(Thread.java:750)

A simple null check should fix this.

Issue Links

• links to: GitHub Pull Request #5397