Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Plain View
HDFS >> mail # dev >> test failures on trunk


+
Eli Collins 2011-12-15, 05:51
+
Uma Maheswara Rao G 2011-12-15, 14:16
+
Suresh Srinivas 2011-12-15, 15:57
Copy link to this message
-
RE: test failures on trunk
Thanks a lot, Suresh.

Here is the Jira for this issue https://issues.apache.org/jira/browse/HDFS-2687

Regards,
Uma
________________________________________
From: Suresh Srinivas [[EMAIL PROTECTED]]
Sent: Thursday, December 15, 2011 9:27 PM
To: [EMAIL PROTECTED]
Subject: Re: test failures on trunk

I will fix this bug today.
On Thursday, December 15, 2011, Uma Maheswara Rao G <[EMAIL PROTECTED]>
wrote:
> Yes Eli,  Agree with you.
>
>  I think we must take a look immediately now.
>
> I just debugged the failure related to this class cast exception.
>
> Following is the initial analysis:
>
> Previous behaviour of getListing in ClientNamenodeProtocolTranslatorR23 is
>  It will pass needLocation flag as true and server side will be packed
empty LocatedBlocks with HdfsLocatedFileStatus object if getBlocations are
null and needLocation flag is true. So, cleint will reconstruct the
HdfsLocatedFileStatus directly  if HdfsfileStatus is instanceOf
HdfsLocatedFileStatus.
>
> But with new implementation of protocolBuffers,  It is just ignoring the
LocatedBlocks are server side if the getBlocklocations are null. At client
side, to reconstruct the HDFSFileStatus object, it is depending on
locations field. Since there is no locations field here, it is considering
HDFSFileStatus as just HDFSFileStatus.
>
> Here test assumes that HDFSFileStatus  obj will be type of
HdfsLocatedFileStatus  and trying cast. This is failing here.
>
> I think Suresh and Sanjay can give more hint on this.
>
> Regards,
> Uma
> ________________________________________
> From: Eli Collins [[EMAIL PROTECTED]]
> Sent: Thursday, December 15, 2011 11:21 AM
> To: [EMAIL PROTECTED]
> Subject: test failures on trunk
>
> Hey gang,
>
> Looks like a number of the trunk tests
> (https://builds.apache.org/job/Hadoop-Hdfs-trunk) started failing on
> the 10th, due to the following. Ring a bell?  Maybe due to all the
> recent protocol changes?
>
> Error Message
>
> org.apache.hadoop.hdfs.protocol.HdfsFileStatus cannot be cast to
> org.apache.hadoop.hdfs.protocol.HdfsLocatedFileStatus
> Stacktrace
>
> java.lang.ClassCastException:
> org.apache.hadoop.hdfs.protocol.HdfsFileStatus cannot be cast to
> org.apache.hadoop.hdfs.protocol.HdfsLocatedFileStatus
>        at
org.apache.hadoop.hdfs.DistributedFileSystem$1.hasNext(DistributedFileSystem.java:452)
>        at org.apache.hadoop.fs.FileSystem$5.hasNext(FileSystem.java:1551)
>        at org.apache.hadoop.fs.FileSystem$5.next(FileSystem.java:1581)
>        at org.apache.hadoop.fs.FileSystem$5.next(FileSystem.java:1541)
>        at
org.apache.hadoop.fs.TestListFiles.testDirectory(TestListFiles.java:146)
>
> Thanks,
> Eli
>
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB