[
https://issues.apache.org/jira/browse/HADOOP-5267?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12673891#action_12673891
]
Steve Loughran commented on HADOOP-5267:
----------------------------------------
full stack trace :
{code}
[sf-system-test-junit] Map: Task Id : attempt_200902161423_0001_m_000000_0,
Status : TIPFAILED
[sf-system-test-junit] java.io.IOException: Cannot open filename
/tests/mrtestsequence/in/in.txt
[sf-system-test-junit] at
org.apache.hadoop.hdfs.DFSClient$DFSInputStream.openInfo(DFSClient.java:1352)
[sf-system-test-junit] at
org.apache.hadoop.hdfs.DFSClient$DFSInputStream.<init>(DFSClient.java:1343)
[sf-system-test-junit] at
org.apache.hadoop.hdfs.DFSClient.open(DFSClient.java:312)
[sf-system-test-junit] at
org.apache.hadoop.hdfs.DistributedFileSystem.open(DistributedFileSystem.java:177)
[sf-system-test-junit] at
org.apache.hadoop.fs.FileSystem.open(FileSystem.java:347)
[sf-system-test-junit] at
org.apache.hadoop.mapreduce.lib.input.LineRecordReader.initialize(LineRecordReader.java:67)
[sf-system-test-junit] at
org.apache.hadoop.mapred.MapTask$NewTrackingRecordReader.initialize(MapTask.java:412)
[sf-system-test-junit] at
org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:510)
[sf-system-test-junit] at
org.apache.hadoop.mapred.MapTask.run(MapTask.java:303)
[sf-system-test-junit] at org.apache.hadoop.mapred.Child.main(Child.java:155)
{code}
I'd rather an error message like
{code}
java.io.FileNotFoundException: File not Found "tests/mrtestsequence/in/in.txt"
{code}
or even
{code}
java.io.FileNotFoundException: File not Found
"hfds://server4:9000/tests/mrtestsequence/in/in.txt"
{code}
subtyping for code to handle better; including the full URL makes it easier to
detect problems where the URL itself is at fault
> DFSClient could throw a FileNotFound exception when a file could not be opened
> ------------------------------------------------------------------------------
>
> Key: HADOOP-5267
> URL: https://issues.apache.org/jira/browse/HADOOP-5267
> Project: Hadoop Core
> Issue Type: Improvement
> Components: dfs
> Affects Versions: 0.21.0
> Reporter: Steve Loughran
> Priority: Minor
>
> DfsClient.openInit() throws an IOE when a file can't be found, that is, it
> has no blocks
> [sf-startdaemon-debug] 09/02/16 12:38:47 [IPC Server handler 0 on 8012] INFO
> mapred.TaskInProgress : Error from attempt_200902161238_0001_m_000000_2:
> java.io.IOException: Cannot open filename /tests/mrtestsequence/in/in.txt
> [sf-startdaemon-debug] at
> org.apache.hadoop.hdfs.DFSClient$DFSInputStream.openInfo(DFSClient.java:1352)
> [sf-startdaemon-debug] at
> org.apache.hadoop.hdfs.DFSClient$DFSInputStream.<init>(DFSClient.java:1343)
> [sf-startdaemon-debug] at
> org.apache.hadoop.hdfs.DFSClient.open(DFSClient.java:312)
> [sf-startdaemon-debug] at
> org.apache.hadoop.hdfs.DistributedFileSystem.open(DistributedFileSystem.java:177)
> [sf-startdaemon-debug] at
> org.apache.hadoop.fs.FileSystem.open(FileSystem.java:347)
> I propose turning this into a FileNotFoundException, which is more specific
> about the underlying problem. Including the full dfs URL would be useful too.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.