Another thing you can try is
export HADOOP_ROOT_LOGGER=DEBUG,console and run the hadoop dfs -cat command with hftp and you should get more logs on the client.
Also since you are running cdh it might be better to ask on the cdh mailing lists.
On Dec 18, 2012, at 3:28 PM, Robert Rapplean <[EMAIL PROTECTED]> wrote:
> The cluster says this:
> Hadoop 2.0.0-cdh4.0.0
> Subversion file:///data/1/jenkins/workspace/generic-package-rhel64-6-0/topdir/BUILD/hadoop-2.0.0-cdh4.0.0/src/hadoop-common-project/hadoop-common -r 5d678f6bb1f2bc49e2287dd69ac41d7232fc9cdc
> Compiled by jenkins on Mon Jun 4 16:52:21 PDT 2012
> From source with checksum 64f877fc49f5adc0d7d55c13089e866e
> Which put a tiny strain on my knowledge to retrieve. Can you make a suggestion regarding which logs you want to look at?
> Robert Rapplean
> Senior Software Engineer
> 303-872-2256 direct | 303.438.9597 main | www.trueffect.com
> -----Original Message-----
> From: Harsh J [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, December 18, 2012 4:17 PM
> To: <[EMAIL PROTECTED]>
> Subject: Re: hftp can list directories but won't send files
> What version/distribution of Hadoop is your source cluster?
> Also, I'd take a look at a your NN's and a few of your DN's logs right after encountering this issue, to see the reason+stacktrace printed for the Server Error 500 (a code for a server-end fault). That'd give us more ideas on whys.
> On Wed, Dec 19, 2012 at 4:13 AM, Robert Rapplean <[EMAIL PROTECTED]> wrote:
>> Hey, everone. Just got finished reading about all of the unsubscribe messages in Sept-Oct, and was hoping someone has a clue about what my system is doing wrong. I suspect that this is a configuration issue, but I don't even know where to start looking for it. I'm a developer, and my sysadmin is tied up until the end of the year.
>> I'm trying to move files from one cluster to another using distcp, using the hftp protocol as specified in their instructions.
>> I can read directories over hftp, but when I attempt to get a file I get a 500 (internal server error). To eliminate the possibility of network and firewall issues, I'm using hadoop fs -ls and hadoop fs -cat commands on the source server in order to attempt to figure out this issue.
>> This provides a directory of the files, which is correct.
>> hadoop fs -ls ourlogs/day_id=19991231/hour_id=1999123123
>> -rw-r--r-- 3 username supergroup 812 2012-12-16 17:21 logfiles/day_id=19991231/hour_id=1999123123/000008_0
>> This gives me a "file not found" error, which is also correct because the file isn't there:
>> hadoop fs -cat
>> 231/hour_id=1999123123/000008_0x': No such file or directory
>> This line gives me a 500 internal server error. The file is confirmed to be on the server.
>> hadoop fs -cat
>> cat: HTTP_OK expected, received 500
>> Here is a stack trace of what distcp logs when I attempt this:
>> java.io.IOException: HTTP_OK expected, received 500
>> at org.apache.hadoop.hdfs.HftpFileSystem$RangeHeaderUrlOpener.connect(HftpFileSystem.java:365)
>> at org.apache.hadoop.hdfs.ByteRangeInputStream.openInputStream(ByteRangeInputStream.java:119)
>> at org.apache.hadoop.hdfs.ByteRangeInputStream.getInputStream(ByteRangeInputStream.java:103)
>> at org.apache.hadoop.hdfs.ByteRangeInputStream.read(ByteRangeInputStream.java:187)
>> at java.io.DataInputStream.read(DataInputStream.java:83)
>> at org.apache.hadoop.tools.DistCp$CopyFilesMapper.copy(DistCp.java:424)
>> at org.apache.hadoop.tools.DistCp$CopyFilesMapper.map(DistCp.java:547)
>> at org.apache.hadoop.tools.DistCp$CopyFilesMapper.map(DistCp.java:314)