[HADOOP-3576] hadoop dfs -mv throws NullPointerException

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 0.18.0
• Fix Version/s: 0.18.0
• Hadoop Flags: Reviewed
• Assignee: Tsz-wo Sze
• Reporter: Lohit Vijaya Renu
• Created: 17/Jun/08 01:08
• Updated: 08/Jul/09 16:43
• Resolved: 19/Jun/08 20:37

Description

hadoop dfs -mv command throws NullPointerException while moving a directory to its subdirecotry. In 0.17 version, such a move was not allowed.

Consider the example

[lohit@ hadoop-core-trunk]$ ./bin/hadoop dfs -mv /a/b /a/b/c
mv: java.io.IOException: java.lang.NullPointerException
[lohit@ hadoop-core-trunk]$ 

After this, the namespace of /a/b is gone

[lohit@ hadoop-core-trunk]$ ./bin/hadoop dfs -lsr /a
[lohit@ hadoop-core-trunk]$ 

Restarting the namenode recovers this namespace and everything seems to be normal.
On the other hand, before restarting the namenode, if we delete the directory /a, it succeeds.

[lohit@ hadoop-core-trunk]$ ./bin/hadoop dfs -rmr /a
Deleted hdfs://jamba.juice:8020/a
[lohit@ hadoop-core-trunk]$ 

But, restarting now, throws an exception on NameNode and NameNode wouldn't start.

2008-06-17 00:58:50,422 ERROR org.apache.hadoop.dfs.NameNode: java.lang.NullPointerException
  at org.apache.hadoop.dfs.FSNamesystem.changeLease(FSNamesystem.java:4339)
  at org.apache.hadoop.dfs.FSEditLog.loadFSEdits(FSEditLog.java:561)
  at org.apache.hadoop.dfs.FSImage.loadFSEdits(FSImage.java:846)
  at org.apache.hadoop.dfs.FSImage.loadFSImage(FSImage.java:675)
  at org.apache.hadoop.dfs.FSImage.recoverTransitionRead(FSImage.java:289)
  at org.apache.hadoop.dfs.FSDirectory.loadFSImage(FSDirectory.java:80)
  at org.apache.hadoop.dfs.FSNamesystem.initialize(FSNamesystem.java:273)
  at org.apache.hadoop.dfs.FSNamesystem.<init>(FSNamesystem.java:252)
  at org.apache.hadoop.dfs.NameNode.initialize(NameNode.java:148)
  at org.apache.hadoop.dfs.NameNode.<init>(NameNode.java:193)
  at org.apache.hadoop.dfs.NameNode.<init>(NameNode.java:179)
  at org.apache.hadoop.dfs.NameNode.createNameNode(NameNode.java:825)
  at org.apache.hadoop.dfs.NameNode.main(NameNode.java:834)

In hadoop 0.17, we never allowed such a move.

[lohit@ branch-0.17]$ ./bin/hadoop dfs -mv /a/b /a/b/c
mv: Failed to rename /a/b to /a/b/c
[lohit@ branch-0.17]$ 

This is the issue seen with HADOOP-3561. Opening this JIRA to fix the underlying problem while HADOOP-3561 could be committed.

Issue Links

• is related to: HADOOP-3561 With trash enabled, 'hadoop fs -rmr .' still fully deletes the working dir