geomesa shell script for cloudera 5

geomesa.sh

#! /usr/bin/env bash
#
# Copyright (c) 2013-2016 Commonwealth Computer Research, Inc.
# All rights reserved. This program and the accompanying materials
# are made available under the terms of the Apache License, Version 2.0 which
# accompanies this distribution and is available at
# http://www.opensource.org/licenses/apache2.0.php.
#

setGeoHome ()
{
    SOURCE="${BASH_SOURCE[0]}"
    # resolve $SOURCE until the file is no longer a symlink
    while [ -h "${SOURCE}" ]; do
        bin="$( cd -P "$( dirname "${SOURCE}" )" && pwd )"
        SOURCE="$(readlink "${SOURCE}")"
        # if $SOURCE was a relative symlink, we need to resolve it relative to the path where
        # the symlink file was located
        [[ "${SOURCE}" != /* ]] && SOURCE="${bin}/${SOURCE}"
    done
    bin="$( cd -P "$( dirname "${SOURCE}" )" && cd ../ && pwd )"
    export GEOMESA_HOME="$bin"
    export PATH=${GEOMESA_HOME}/bin:$PATH
    echo >&2 "Warning: GEOMESA_HOME is not set, using $GEOMESA_HOME"
}

GEOMESA_OPTS="-Duser.timezone=UTC -DEPSG-HSQL.directory=/tmp/$(whoami)"
GEOMESA_DEBUG_OPTS="-Xmx8192m -XX:MaxPermSize=512m -XX:-UseGCOverheadLimit -Xdebug -Xnoagent -Xrunjdwp:transport=dt_socket,server=y,suspend=y,address=9898"
GEOMESA_CP=""
JAVA_LIBRARY_PATH=""



# Define GEOMESA_HOME and update the PATH if necessary.
if [ -z "$GEOMESA_HOME" ]; then
setGeoHome
else
    echo >&2 "Using GEOMESA_HOME = $GEOMESA_HOME"
    if [[ $1 = configure ]]; then
        read -p "Do you want to reset this? Y\n " -n 1 -r
        if [[  $REPLY =~ ^[Yy]$ ]]; then
            echo >&2 ""
            setGeoHome
        fi
        echo >&2 ""
    fi
fi

# GEOMESA paths, GEOMESA_LIB should live inside GEOMESA_HOME, but can be pointed elsewhere
if [ -z "$GEOMESA_LIB" ]; then
    GEOMESA_LIB=${GEOMESA_HOME}/lib
elif [[ $1 = configure ]]; then
    message="Warning: GEOMESA_LIB was already set, probably by a prior configuration."
    message="${message}\n The current value is ${GEOMESA_LIB}."
    echo >&2 ""
    echo -e >&2 "$message"
    echo >&2 ""
    read -p "Do you want to reset this to ${GEOMESA_HOME}/lib? Y\n " -n 1 -r
    if [[  $REPLY =~ ^[Yy]$ ]]; then
        GEOMESA_LIB=${GEOMESA_HOME}/lib
        echo >&2 ""
        echo "Now set to ${GEOMESA_LIB}"
    fi
    echo >&2 ""
fi

if [ -z "$GEOMESA_CONF_DIR" ]; then
    GEOMESA_CONF_DIR=${GEOMESA_HOME}/conf
elif [[ $1 = configure ]]; then
    message="Warning: GEOMESA_CONF_DIR was already set, probably by a prior configuration."
    message="${message}\n The current value is ${GEOMESA_CONF_DIR}."
    echo >&2 ""
    echo -e >&2 "$message"
    echo >&2 ""
    read -p "Do you want to reset this to ${GEOMESA_HOME}/conf? Y\n " -n 1 -r
    if [[  $REPLY =~ ^[Yy]$ ]]; then
        GEOMESA_CONF_DIR=${GEOMESA_HOME}/conf
        echo >&2 ""
        echo "Now set to ${GEOMESA_CONF_DIR}"
    fi
    echo >&2 ""
fi

GEOMESA_OPTS="${GEOMESA_OPTS} -Dlog4j.configuration=file://${GEOMESA_CONF_DIR}/log4j.properties"


# ACCUMULO paths, user can hard set these, or rely on this script to find set them via
if [ -n "$ACCUMULO_HOME" ]; then
    if [ -z "$ACCUMULO_LIB" -a -d "${ACCUMULO_HOME}/lib" ]; then
        ACCUMULO_LIB=${ACCUMULO_HOME}/lib
    fi
    if [ -z "$ACCUMULO_CONF_DIR" -a -d "${ACCUMULO_HOME}/conf" ]; then
        ACCUMULO_CONF_DIR=${ACCUMULO_HOME}/conf
    fi
fi

# HADOOP paths, user can hard set these, or rely on this script to find set them via HADOOP_HOME
if [ -n "$HADOOP_HOME" ]; then
    if [ -z "$HADOOP_CONF_DIR" -a -d "${HADOOP_HOME}/etc/hadoop" ]; then
        HADOOP_CONF_DIR=${HADOOP_HOME}/etc/hadoop
    fi
fi

# Warn user about missing environment variables
if [[ (-z "$ACCUMULO_HOME") || (-z "$HADOOP_HOME") ]]; then
    message="Warning: you have not set ACCUMULO_HOME and/or HADOOP_HOME as environment variables."
    message="${message}\nGeoMesa tools will not run without the appropriate Accumulo and Hadoop jars in the tools classpath."
    message="${message}\nPlease ensure that those jars are present in the classpath by running 'geomesa classpath' ."
    message="${message}\nTo take corrective action, please place the necessary jar files in the lib directory of geomesa-tools."
    echo >&2 ""
    echo -e >&2 "$message"
    echo >&2 ""
elif [[ (-z "$ACCUMULO_CONF_DIR") || (-z "$HADOOP_CONF_DIR")]]; then
    message="Warning: you have not set ACCUMULO_CONF_DIR and/or HADOOP_CONF_DIR as environment variables."
    message="${message}\nGeoMesa tools looks for these folders relative to the paths of the ACCUMULO_HOME and HADOOP_HOME environment variables."
    message="${message}\nGeoMesa tools can run without these variables set, however the user will be required to provide the additional instanceId and zookeepers parameters."
    message="${message}\nTo take corrective action, please set these environment variable to the appropriate locations"
    echo >&2 ""
    echo -e >&2 "$message"
    echo >&2 ""
fi

function findJars() {
    home="$1"
    CP=""
    if [ -n "$home" -a -d "$home" ]; then
        for jar in $(find ${home} -name "*.jar"); do
            if [[ ("$2" != "true") || (("$jar" != *"test"*) && ("$jar" != *"slf4j"*)) ]]; then
                if [ "$jar" = "${jar%-sources.jar}" -a "$jar" = "${jar%-test.jar}" ]; then
                  CP="$CP:$jar"
                fi
            fi
        done
        if [ -d "$home/native" ]; then
            if [ -z "$JAVA_LIBRARY_PATH" ]; then
                JAVA_LIBRARY_PATH="$home/native"
            else
                JAVA_LIBRARY_PATH="$home/native:$JAVA_LIBRARY_PATH"
            fi
        fi
    fi
    if [ "${CP:0:1}" = ":" ]; then
        CP="${CP:1}"
    fi
    echo $CP
}

# Start constructing GEOMESA_CP (classpath)
# include geomesa first so that the correct log4j.properties is picked up
GEOMESA_CP="${GEOMESA_CONF_DIR}:$(findJars $GEOMESA_LIB)"
GEOMESA_CP="${GEOMESA_CP}:${ACCUMULO_CONF_DIR}:${HADOOP_CONF_DIR}"
ACCUMULO_CP="$(findJars $ACCUMULO_LIB)"

# Get the hadoop jars, ignoring jars with names containing slf4j and test
# Copied from accumulo classpath
hadoopDirs=(
  $ZOOKEEPER_HOME
  # Hadoop 2 requirements
  #$HADOOP_HOME/share/hadoop/common
  #$HADOOP_HOME/share/hadoop/hdfs/
  #$HADOOP_HOME/share/hadoop/mapreduce/
  #$HADOOP_HOME/share/hadoop/tools/lib
  #$HADOOP_HOME/share/hadoop/yarn/
  ## HDP 2.0 requirements
#  /usr/lib/hadoop/
#  /usr/lib/hadoop-hdfs/
#  /usr/lib/hadoop-mapreduce/
#  /usr/lib/hadoop-yarn/
  # HDP 2.2 requirements
  #/usr/hdp/current/hadoop-client/
  #/usr/hdp/current/hadoop-hdfs-client/
  #/usr/hdp/current/hadoop-mapreduce-client/
  #/usr/hdp/current/hadoop-yarn-client/
  ## IOP 4.1 requirements
  #/usr/iop/current/hadoop-client/
  #/usr/iop/current/hadoop-hdfs-client/
  #/usr/iop/current/hadoop-mapreduce-client/
  #/usr/iop/current/hadoop-yarn-client/
  # CDH5
#  $HADOOP_COMMON_HOME
  $HADOOP_HDFS_HOME
  $YARN_HOME
  $HADOOP_MAPRED_HOME
)
HADOOP_CP=/usr/lib/hadoop/hadoop-common-2.6.0-cdh5.5.2-tests.jar
HADOOP_CP=$HADOOP_CP:/usr/lib/hadoop/hadoop-common-2.6.0-cdh5.5.2.jar
HADOOP_CP=$HADOOP_CP:/usr/lib/hadoop/hadoop-aws-2.6.0-cdh5.5.2.jar
HADOOP_CP=$HADOOP_CP:/usr/lib/hadoop/hadoop-auth-2.6.0-cdh5.5.2.jar
HADOOP_CP=$HADOOP_CP:/usr/lib/hadoop/hadoop-annotations-2.6.0-cdh5.5.2.jar

for home in ${hadoopDirs[*]}; do
  tmp="$(findJars $home true)"
  if [ -n "$tmp" ]; then
    HADOOP_CP="$HADOOP_CP:$tmp"
  fi
  if [ "${HADOOP_CP:0:1}" = ":" ]; then
    HADOOP_CP="${HADOOP_CP:1}"
  fi
done

if [ -n "$JAVA_LIBRARY_PATH" ]; then
  GEOMESA_OPTS="$GEOMESA_OPTS,java.library.path=$JAVA_LIBRARY_PATH"
  export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:$JAVA_LIBRARY_PATH"
fi

# Configure geomesa logging directory
if [[ -z "${GEOMESA_LOG_DIR}" ]]; then
  export GEOMESA_LOG_DIR="${GEOMESA_HOME}/logs"
  export GEOMESA_OPTS="-Dgeomesa.log.dir=${GEOMESA_LOG_DIR} $GEOMESA_OPTS"
fi

# Geomesa tools setup completed, below we parse user arguments
#
# configure runs the geomesa configuration, sets both the GEOMESA_HOME and PATH vars
# classpath prints the class path, ie list of jars we are using
# else is running actual commands in the tools

if  [[ $1 = configure ]]; then
    echo >&2 "Using GEOMESA_HOME as set: $GEOMESA_HOME"
    read -p "Is this intentional? Y\n " -n 1 -r
    if [[  $REPLY =~ ^[Nn]$ ]]; then
        bin="$( cd -P "$( dirname "${SOURCE}" )" && cd ../ && pwd )"
        export GEOMESA_HOME="$bin"
        echo >&2 ""
        echo "Now set to ${GEOMESA_HOME}"
    fi

    if [ -z "$GEOMESA_LIB" ]; then
        GEOMESA_LIB=${GEOMESA_HOME}/lib
    else
        message="Warning: GEOMESA_LIB already set, probably by a prior configuration."
        message="${message}\n Current value is ${GEOMESA_LIB}."
        echo >&2 ""
        echo -e >&2 "$message"
        echo >&2 ""
        read -p "Is this intentional? Y\n " -n 1 -r
        if [[  $REPLY =~ ^[Nn]$ ]]; then
            GEOMESA_LIB=${GEOMESA_HOME}/lib
            echo >&2 ""
            echo "Now set to ${GEOMESA_LIB}"
        fi
        echo >&2 ""
    fi

    echo >&2 ""
    echo "To persist the configuration please update your bashrc file to include: "
    echo "export GEOMESA_HOME="$GEOMESA_HOME""
    echo "export PATH=\${GEOMESA_HOME}/bin:\$PATH"
elif [[ $1 = classpath ]]; then
    CLASSPATH="$GEOMESA_CP:$ACCUMULO_CP:$HADOOP_CP"
    for element in ${CLASSPATH//:/ } ; do
        echo ${element}
    done
else
    if [[ $1 = debug ]]; then
        GEOMESA_OPTS="$GEOMESA_OPTS $GEOMESA_DEBUG_OPTS"
        shift 1
    fi
    CLASSPATH="$GEOMESA_CP:$ACCUMULO_CP:$HADOOP_CP"
    java ${GEOMESA_OPTS} -cp ${CLASSPATH} org.locationtech.geomesa.tools.Runner "$@"
fi