Mailing-List: contact user-help@accumulo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@accumulo.apache.org
MIME-Version: 1.0
In-Reply-To: 
 <CA+zmODjzJf+QYaQbcuUCyWdeM42MuS4T_fxiTt7QxgMFMprKCw@mail.gmail.com>
References: 
 <CA+zmODjzJf+QYaQbcuUCyWdeM42MuS4T_fxiTt7QxgMFMprKCw@mail.gmail.com>
Date: Thu, 9 May 2013 13:34:36 -0400
Message-ID: 
 <CAL5zq9YGU7dDuc5BNv6hn5MxNn58SScrdVGr9wgJcEhHqkSLBQ@mail.gmail.com>
Subject: Re: Accumulo and Bigtop Hadoop
From: Christopher <ctubbsii@apache.org>
To: user@accumulo.apache.org
Content-Type: text/plain; charset=ISO-8859-1

This is precisely why these classpaths are configurable. My personal
goal is that, in the future (1.6.0 and later), the RPMs/DEBs that we
provide will better support BigTop packages out of the box without the
need to edit the configuration.

(Also see John's response about hadoop 2 support.)

--
Christopher L Tubbs II
http://gravatar.com/ctubbsii


On Thu, May 9, 2013 at 1:22 PM, Roshan Punnoose <roshanp@gmail.com> wrote:
> I have a feeling this has already been asked, but I could not find the
> answer on google.
>
> Bigtop is installing hadoop 2.0.2-alpha into /usr/lib/hadoop,
> /usr/lib/hadoop-hdfs, /usr/lib/hadoop-yarn, etc. However, Accumulo is setup
> to have the "general.classpaths" of accumulo-site to point
> /usr/lib/hadoop/(share)/(hdfs|yarn)/*.jar.
>
> As a quick workaround I had to edit the accumulo-env.sh and
> accumulo-site.sh:
>
> accumulo-env.sh
> ###for bigtop
> export HADOOP_COMMON_DIR="./"
> export HADOOP_COMMON_LIB_JARS_DIR="lib"
> export HADOOP_COMMON_LIB_NATIVE_DIR="lib/native"
> export HDFS_DIR="./"
> export HDFS_LIB_JARS_DIR="lib"
> export YARN_DIR="./"
> export YARN_LIB_JARS_DIR="lib"
> export MAPRED_DIR="./"
> export MAPRED_LIB_JARS_DIR="lib"
>
> export HADOOP_LIBEXEC_DIR=${HADOOP_LIBEXEC_DIR:-"/usr/lib/hadoop/libexec"}
> export HADOOP_CONF_DIR=${HADOOP_CONF_DIR:-"/etc/hadoop/conf"}
> export HADOOP_COMMON_HOME=${HADOOP_COMMON_HOME:-"/usr/lib/hadoop"}
> export HADOOP_HDFS_HOME=${HADOOP_HDFS_HOME:-"/usr/lib/hadoop-hdfs"}
> export HADOOP_MAPRED_HOME=${HADOOP_MAPRED_HOME:-"/usr/lib/hadoop-mapreduce"}
> export HADOOP_YARN_HOME=${HADOOP_YARN_HOME:-"/usr/lib/hadoop-yarn"}
> ###for bigtop
>
> accumulo-site.sh (added to the general.classpaths)
>       $HADOOP_MAPRED_HOME/.*.jar,
>       $HADOOP_MAPRED_HOME/$MAPRED_LIB_JARS_DIR/.*.jar,
>       $HADOOP_YARN_HOME/.*.jar,
>       $HADOOP_YARN_HOME/$YARN_LIB_JARS_DIR/.*.jar,
>       $HADOOP_HDFS_HOME/[^.].*.jar,
>       $HADOOP_HDFS_HOME/$HDFS_LIB_JARS_DIR/.*.jar,
>
> This above is working. Has anyone ran into this and found a better solution?
>
> Thanks!
>
> Roshan