Starting DataNode 0 with dfs.data.dir:
/Users/billg/ws/svn/chukwa/target/test/classes/target/test-data/f104e7b0-0999-4234-9df8-290189ec3674/dfs/data/data1,/Users/billg/ws/svn/chukwa/target/test/classes/target/test-data/f104e7b0-0999-4234-9df8-290189ec3674/dfs/data/data2
java.lang.NullPointerException
at
org.apache.hadoop.hdfs.MiniDFSCluster.startDataNodes(MiniDFSCluster.java:422)
at
org.apache.hadoop.hdfs.MiniDFSCluster.<init>(MiniDFSCluster.java:280)
at
org.apache.hadoop.hbase.HBaseTestingUtility.startMiniDFSCluster(HBaseTestingUtility.java:223)
at
org.apache.hadoop.hbase.HBaseTestingUtility.startMiniCluster(HBaseTestingUtility.java:345)
at
org.apache.hadoop.hbase.HBaseTestingUtility.startMiniCluster(HBaseTestingUtility.java:308)
at
org.apache.hadoop.chukwa.datacollection.writer.TestHBaseWriter.<init>(TestHBaseWriter.java:66)
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native
Method)
at
sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:39)
at
sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:27)
at java.lang.reflect.Constructor.newInstance(Constructor.java:513)
at junit.framework.TestSuite.createTest(TestSuite.java:61)
at junit.framework.TestSuite.addTestMethod(TestSuite.java:294)
at
junit.framework.TestSuite.addTestsFromTestCase(TestSuite.java:150)
at junit.framework.TestSuite.<init>(TestSuite.java:129)
On Fri, Dec 9, 2011 at 11:34 PM, Eric Yang <[email protected]> wrote:
> TestHBaseWriter works on my side without setting HBASE_CONF_DIR. In
> the exception, it has printed additional stacktrace to stdout. Could
> you paste the output in
>
> target/test-reports/org.apache.hadoop.chukwa.datacollection.writer.TestHBaseWriter-output.txt?
>
> regards,
> Eric
>
> On Fri, Dec 9, 2011 at 11:11 PM, Bill Graham <[email protected]> wrote:
> > When I ran ant package the HBase test failed, so I though it was because
> I
> > hadn't set HBASE_CONF_DIR. If that's not required, then we have a test
> > problem:
> >
> > $ less
> >
> target/test-reports/org.apache.hadoop.chukwa.datacollection.writer.TestHBaseWriter.txt
> > Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 1.625 sec
> > <<< FAILURE!
> > warning(junit.framework.TestSuite$1) Time elapsed: 0.005 sec <<<
> FAILURE!
> > junit.framework.AssertionFailedError: Exception in constructor:
> testWriters
> > (junit.framework.AssertionFailedError
> > at junit.framework.Assert.fail(Assert.java:48)
> > at
> >
> org.apache.hadoop.chukwa.datacollection.writer.TestHBaseWriter.<init>(TestHBaseWriter.java:75)
> > at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native
> > Method)
> > at
> >
> sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:39)
> > at
> >
> sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:27)
> > at
> java.lang.reflect.Constructor.newInstance(Constructor.java:513)
> > ...
> >
> >
> > On Fri, Dec 9, 2011 at 9:16 PM, Eric Yang <[email protected]> wrote:
> >>
> >> mvn clean package works fine, maven will automatically download Hadoop
> >> and HBase jar in the build process from maven central. The
> >> -DHBASE_CONF_DIR and -DHADOOP_CONF_DIR is for developer who wants to
> >> preconfigure the environment variable in chukwa-env.sh. Without
> >> -DHBASE_CONF_DIR and -DHADOOP_CONF_DIR, chukwa-env.sh is default to
> >> ${HBASE_CONF_DIR} and ${HADOOP_CONF_DIR}. Unit tests are using
> >> minicluster with Hadoop/hbase unit test configuration. There is no
> >> configuration required to run Hadoop or HBase dependent unit tests.
> >>
> >> regards,
> >> Eric
> >>
> >> On Fri, Dec 9, 2011 at 5:09 PM, Bill Graham <[email protected]>
> wrote:
> >> > I agree that users should be able to checkout the source and run 'mvn
> >> > package' without separately download and installing both Hadoop and
> >> > HBase.
> >> > Can't these unit tests be refactored to not require this by instead
> >> > referencing Hadoop and HBase conf dirs that are committed?
> >> >
> >> >
> >> > On Tue, Dec 6, 2011 at 8:03 PM, Eric Yang <[email protected]> wrote:
> >> >
> >> >> HBase is optional. You can build it with "mvn clean package", then
> >> >> switch the PipelineWriter to SeqFileWriter. However, the default
> >> >> behavior is to make everything work as seamless as possible for using
> >> >> Chukwa as a whole system.
> >> >>
> >> >> regards,
> >> >> Eric
> >> >>
> >> >> On Tue, Dec 6, 2011 at 12:27 PM, Ariel Rabkin <[email protected]>
> >> >> wrote:
> >> >> > I am unhappy with a build that requires HBase. Is there a way we
> can
> >> >> > make this optional, both at compile and run time?
> >> >> >
> >> >> > --Ari
> >> >> >
> >> >> > On Mon, Dec 5, 2011 at 7:31 PM, Eric Yang <[email protected]>
> wrote:
> >> >> >> Hi all,
> >> >> >>
> >> >> >> I have made some progress in polishing Chukwa 0.5.0. I think it
> is
> >> >> >> almost ready for public consumption. Documents have been updated
> >> >> >> to reflect the current code base. I generated a preview of the
> >> >> >> r0.5.0
> >> >> >> document at:
> >> >> >>
> >> >> >> http://people.apache.org/~eyang/chukwa-0.5.0-docs/
> >> >> >>
> >> >> >> Please review and try out the build by running:
> >> >> >>
> >> >> >> mvn clean package site \
> >> >> >> -DHADOOP_CONF_DIR=/path/to/${HADOOP_CONF_DIR} \
> >> >> >> -DHBASE_CONF_DIR=/path/to/${HBASE_CONF_DIR}
> >> >> >>
> >> >> >> If all goes well, we can vote for branching for 0.5.0, then follow
> >> >> >> by
> >> >> >> release votes.
> >> >> >> Your input to this matter is important. Thanks for your
> attention.
> >> >> >>
> >> >> >> regards,
> >> >> >> Eric
> >> >> >
> >> >> >
> >> >> >
> >> >> > --
> >> >> > Ari Rabkin [email protected]
> >> >> > UC Berkeley Computer Science Department
> >> >>
> >
> >
>
