[
https://issues.apache.org/jira/browse/HADOOP-3388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12596671#action_12596671
]
dhruba borthakur commented on HADOOP-3388:
------------------------------------------
svn diff
Index: src/test/org/apache/hadoop/dfs/TestDatanodeBlockScanner.java
===================================================================
--- src/test/org/apache/hadoop/dfs/TestDatanodeBlockScanner.java
(revision 656131)
+++ src/test/org/apache/hadoop/dfs/TestDatanodeBlockScanner.java
(working copy)
@@ -174,7 +174,7 @@
fs = cluster.getFileSystem();
Path file1 = new Path("/tmp/testBlockVerification/file1");
DFSTestUtil.createFile(fs, file1, 1024, (short)3, 0);
- String block = DFSTestUtil.getFirstBlock(fs, file1).toString();
+ String block = DFSTestUtil.getFirstBlock(fs, file1).getBlockName();
dfsClient = new DFSClient(new InetSocketAddress("localhost",
cluster.getNameNodePort()), conf);
The above patch fixes only part of the problem. The other part is probably
related to the CorruptReplicaMap that has recently been committed.
> TestDatanodeBlockScanner failed while trying to corrupt replicas
> ----------------------------------------------------------------
>
> Key: HADOOP-3388
> URL: https://issues.apache.org/jira/browse/HADOOP-3388
> Project: Hadoop Core
> Issue Type: Bug
> Components: test
> Reporter: dhruba borthakur
> Assignee: dhruba borthakur
>
> Blocks now have generation stamp associated with them. This unit test does a
> Block.toString() to find out the name of the block. Instead it should use
> Block.getBlockName().
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
