Accumulo not getting initialised. - accumulo

Question

I am trying to initialise accumulo. I am configuring accumulo on hadoop2.0.0-cdh4.4.0.
I am making using tars on a MAC book.
I am getting an error when initialising accumulo : bin/accumulo init.
java.io.IOException: Mkdirs failed to create /accumulo/instance_id error.
The log says:
2014-05-24 01:24:33,935 [util.Initialize] FATAL: Failed to initialize filesystem
java.io.IOException: Mkdirs failed to create /accumulo/instance_id
at org.apache.hadoop.fs.ChecksumFileSystem.create(ChecksumFileSystem.java:447)
at org.apache.hadoop.fs.ChecksumFileSystem.create(ChecksumFileSystem.java:433)
at org.apache.hadoop.fs.FileSystem.create(FileSystem.java:886)
at org.apache.hadoop.fs.FileSystem.create(FileSystem.java:867)
at org.apache.hadoop.fs.FileSystem.create(FileSystem.java:829)
at org.apache.hadoop.fs.FileSystem.createNewFile(FileSystem.java:1129)
at org.apache.accumulo.server.util.Initialize.initFileSystem(Initialize.java:269)
at org.apache.accumulo.server.util.Initialize.initialize(Initialize.java:213)
at org.apache.accumulo.server.util.Initialize.doInit(Initialize.java:199)
at org.apache.accumulo.server.util.Initialize.main(Initialize.java:545)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:606)
at org.apache.accumulo.start.Main$1.run(Main.java:103)
at java.lang.Thread.run(Thread.java:744)
2014-05-24 01:24:33,937 [conf.Configuration] WARN : fs.default.name is deprecated. Instead, use fs.defaultFS
2014-05-24 01:24:33,937 [util.Initialize] FATAL: Default filesystem value ('fs.defaultFS' or 'fs.default.name') was found in the Hadoop configuration
2014-05-24 01:24:33,938 [util.Initialize] FATAL: Please ensure that the Hadoop core-site.xml is on the classpath using 'general.classpaths' in accumulo-site.xml
Please suggest me , I tried to fix this by creating the /accumulo, /user/accumulo on hdfs and gave 777 permissions also.

Answer 1

The root cause is that the Hadoop jars and configuration are not being placed on Accumulo's classpath. I'm not familiar with how Cloudera packages their Hadoop artifacts.
If you notice in your stack trace, it lists out the ChecksumFileSystem class instead of the DistributedFileSystem. This means that Accumulo doesn't know about the HDFS instance you're trying to write to and is falling back to using the local file system (that's what the ChecksumFileSystem is doing).
To fix this, check a couple of things in your Accumulo configuration files. First, make sure that you have correctly defined HADOOP_PREFIX and HADOOP_CONF_DIR in accumulo-env.sh. Second, make sure that the value you have configured for general.classpaths in accumulo-site.xml all exist, specifically the ones that reference HADOOP_PREFIX and HADOOP_CONF_DIR.