On 23 Feb 2016, at 08:22, Arunkumar Pillai
<arunkumar1...@gmail.com<mailto:arunkumar1...@gmail.com>> wrote:
at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:121)
at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)
Caused by: java.lang.reflect.InvocationTargetException
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
at
sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)
at
sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
at java.lang.reflect.Constructor.newInstance(Constructor.java:422)
at
org.apache.hadoop.util.ReflectionUtils.newInstance(ReflectionUtils.java:129)
... 33 more
Caused by: java.lang.UnsatisfiedLinkError:
org.apache.hadoop.security.JniBasedUnixGroupsMapping.anchorNative()V
at
org.apache.hadoop.security.JniBasedUnixGroupsMapping.anchorNative(Native Method)
at
org.apache.hadoop.security.JniBasedUnixGroupsMapping.<clinit>(JniBasedUnixGroupsMapping.java:49)
at
org.apache.hadoop.security.JniBasedUnixGroupsMappingWithFallback.<init>(JniBasedUnixGroupsMappingWithFallback.java:39)
looks supiciously like there's a mismatch between the libhadoop.so on your path
and the Hadoop version. Is there a Hadoop 2.6 installed on the same system?
you could try to skip that bit of JNI code by switching to the shell:
spark.hadoop.security.group.mapping
org.apache.hadoop.security.ShellBasedUnixGroupsMapping
...but that will just postpone the problem
Best to find all copies of libhadoop.so on your filesystem, and make sure the
one that gets loaded is the Hadoop 2.7 one
