[HDFS-3100] failed to append data

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.23.1, 2.0.0-alpha
• Fix Version/s: 2.0.0-alpha
• Component/s: datanode
• Hadoop Flags: Reviewed
• Assignee: Brandon Li
• Reporter: Zhanwei Wang
• Created: 15/Mar/12 19:07
• Updated: 28/Sep/15 20:58
• Resolved: 22/Mar/12 00:14

Description

STEP：
1, deploy a single node hdfs 0.23.1 cluster and configure hdfs as:
A) enable webhdfs
B) enable append
C) disable permissions
2, start hdfs
3, run the test script as attached

RESULT:
expected: a file named testFile should be created and populated with 32K * 5000 zeros, HDFS should be OK.
I got: script cannot be finished, file has been created but not be populated as expected, actually append operation failed.

Datanode log shows that, blockscaner report a bad replica and nanenode decide to delete it. Since it is a single node cluster, append fail. It makes no sense that the script failed every time.

Datanode and Namenode logs are attached.

Issue Links

• is broken by: HDFS-3082 Clean up FSDatasetInterface
• relates to: HDFS-2525 Race between BlockPoolSliceScanner and append