diff --git a/ami-list-ubuntu/ami-list-ubuntu-amd64 b/ami-list-ubuntu/ami-list-ubuntu-amd64
new file mode 100644
index 00000000..ef760f62
--- /dev/null
+++ b/ami-list-ubuntu/ami-list-ubuntu-amd64
@@ -0,0 +1,88 @@
+ap-northeast-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-03d01062,hvm
+ap-northeast-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-90d111f1,hvm
+ap-northeast-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-49d31328,hvm
+ap-northeast-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-55d21234,hvm
+ap-northeast-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-47d41426,aki-176bf516
+ap-northeast-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-97d111f6,aki-176bf516
+ap-northeast-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-24d11145,aki-176bf516
+ap-northeast-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-55d01034,aki-176bf516
+ap-southeast-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-b84799db,hvm
+ap-southeast-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-f0459b93,hvm
+ap-southeast-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-5e429c3d,hvm
+ap-southeast-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-41409e22,hvm
+ap-southeast-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-af4997cc,aki-503e7402
+ap-southeast-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-af449acc,aki-503e7402
+ap-southeast-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-a2449ac1,aki-503e7402
+ap-southeast-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-dc459bbf,aki-503e7402
+ap-southeast-2,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-47f1c524,hvm
+ap-southeast-2,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-def2c6bd,hvm
+ap-southeast-2,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-25f3c746,hvm
+ap-southeast-2,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-41f1c522,hvm
+ap-southeast-2,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-0cf0c46f,aki-c362fff9
+ap-southeast-2,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-1df1c57e,aki-c362fff9
+ap-southeast-2,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-81f3c7e2,aki-c362fff9
+ap-southeast-2,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-fef2c69d,aki-c362fff9
+cn-north-1,trusty,14.04 LTS,amd64,hvm:ebs,20160714,ami-67af7a0a,hvm
+cn-north-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160714,ami-bfad78d2,hvm
+cn-north-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160714,ami-bead78d3,hvm
+cn-north-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160714,ami-02af7a6f,hvm
+cn-north-1,trusty,14.04 LTS,amd64,ebs,20160714,ami-58ae7b35,aki-9e8f1da7
+cn-north-1,trusty,14.04 LTS,amd64,ebs-io1,20160714,ami-f2ad789f,aki-9e8f1da7
+cn-north-1,trusty,14.04 LTS,amd64,ebs-ssd,20160714,ami-ccae7ba1,aki-9e8f1da7
+cn-north-1,trusty,14.04 LTS,amd64,instance-store,20160714,ami-79ae7b14,aki-9e8f1da7
+eu-central-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-d3cd3bbc,hvm
+eu-central-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-51cd3b3e,hvm
+eu-central-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-b1cf39de,hvm
+eu-central-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-6acb3d05,hvm
+eu-central-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-d1cd3bbe,aki-184c7a05
+eu-central-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-b6cf39d9,aki-184c7a05
+eu-central-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-09cf3966,aki-184c7a05
+eu-central-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-d3cf39bc,aki-184c7a05
+eu-west-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-a7412ad4,hvm
+eu-west-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-5a452e29,hvm
+eu-west-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-55452e26,hvm
+eu-west-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-5d42292e,hvm
+eu-west-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-e4452e97,aki-52a34525
+eu-west-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-3f472c4c,aki-52a34525
+eu-west-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-5b452e28,aki-52a34525
+eu-west-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-07492274,aki-52a34525
+sa-east-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-e09d0b8c,hvm
+sa-east-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-f89d0b94,hvm
+sa-east-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-97980efb,hvm
+sa-east-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-1a980e76,hvm
+sa-east-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-16980e7a,aki-5553f448
+sa-east-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-8b9a0ce7,aki-5553f448
+sa-east-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-98980ef4,aki-5553f448
+sa-east-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-6c9f0900,aki-5553f448
+us-east-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-d90d92ce,hvm
+us-east-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-1b0d920c,hvm
+us-east-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-8e0b9499,hvm
+us-east-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-08128d1f,hvm
+us-east-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-c70f90d0,aki-919dcaf8
+us-east-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-510c9346,aki-919dcaf8
+us-east-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-6808977f,aki-919dcaf8
+us-east-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-b266f9a5,aki-919dcaf8
+us-gov-west-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-3508b654,hvm
+us-gov-west-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-2e0bb54f,hvm
+us-gov-west-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-2a08b64b,hvm
+us-gov-west-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-2809b749,hvm
+us-gov-west-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-600cb201,aki-1de98d3e
+us-gov-west-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-780fb119,aki-1de98d3e
+us-gov-west-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-ab0bb5ca,aki-1de98d3e
+us-gov-west-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-2b0bb54a,aki-1de98d3e
+us-west-1,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-27743747,hvm
+us-west-1,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-577a3937,hvm
+us-west-1,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-547b3834,hvm
+us-west-1,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-fd74379d,hvm
+us-west-1,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-d47a39b4,aki-880531cd
+us-west-1,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-11753671,aki-880531cd
+us-west-1,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-31753651,aki-880531cd
+us-west-1,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-33773453,aki-880531cd
+us-west-2,trusty,14.04 LTS,amd64,hvm:ebs,20160809.1,ami-20be7540,hvm
+us-west-2,trusty,14.04 LTS,amd64,hvm:ebs-io1,20160809.1,ami-dcb57ebc,hvm
+us-west-2,trusty,14.04 LTS,amd64,hvm:ebs-ssd,20160809.1,ami-70b67d10,hvm
+us-west-2,trusty,14.04 LTS,amd64,hvm:instance-store,20160809.1,ami-a9b972c9,hvm
+us-west-2,trusty,14.04 LTS,amd64,ebs,20160809.1,ami-abb07bcb,aki-fc8f11cc
+us-west-2,trusty,14.04 LTS,amd64,ebs-io1,20160809.1,ami-88b07be8,aki-fc8f11cc
+us-west-2,trusty,14.04 LTS,amd64,ebs-ssd,20160809.1,ami-73b67d13,aki-fc8f11cc
+us-west-2,trusty,14.04 LTS,amd64,instance-store,20160809.1,ami-a3bf74c3,aki-fc8f11cc
diff --git a/copy-dir.sh b/copy-dir.sh
index e1d2bd07..54e176c1 100755
--- a/copy-dir.sh
+++ b/copy-dir.sh
@@ -37,7 +37,7 @@ DIR=`readlink -f "$1"`
 DIR=`echo "$DIR"|sed 's@/$@@'`
 DEST=`dirname "$DIR"`
 
-SLAVES=`cat /root/spark-ec2/slaves`
+SLAVES=`cat ~/spark-ec2/slaves`
 
 SSH_OPTS="-o StrictHostKeyChecking=no -o ConnectTimeout=5"
 
diff --git a/deploy.generic/root/spark-ec2/ec2-variables.sh b/deploy.generic/root/spark-ec2/ec2-variables.sh
old mode 100644
new mode 100755
diff --git a/deploy.ubuntu/home/ubuntu/spark-ec2/ec2-variables.sh b/deploy.ubuntu/home/ubuntu/spark-ec2/ec2-variables.sh
new file mode 100755
index 00000000..4f3e8da8
--- /dev/null
+++ b/deploy.ubuntu/home/ubuntu/spark-ec2/ec2-variables.sh
@@ -0,0 +1,34 @@
+#!/usr/bin/env bash
+
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements.  See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You under the Apache License, Version 2.0
+# (the "License"); you may not use this file except in compliance with
+# the License.  You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+# These variables are automatically filled in by the spark-ec2 script.
+export MASTERS="{{master_list}}"
+export SLAVES="{{slave_list}}"
+export HDFS_DATA_DIRS="{{hdfs_data_dirs}}"
+export MAPRED_LOCAL_DIRS="{{mapred_local_dirs}}"
+export SPARK_LOCAL_DIRS="{{spark_local_dirs}}"
+export MODULES="{{modules}}"
+export SPARK_VERSION="{{spark_version}}"
+export TACHYON_VERSION="{{tachyon_version}}"
+export HADOOP_MAJOR_VERSION="{{hadoop_major_version}}"
+export SWAP_MB="{{swap}}"
+export SPARK_WORKER_INSTANCES="{{spark_worker_instances}}"
+export SPARK_MASTER_OPTS="{{spark_master_opts}}"
+export AWS_ACCESS_KEY_ID="{{aws_access_key_id}}"
+export AWS_SECRET_ACCESS_KEY="{{aws_secret_access_key}}"
diff --git a/deploy_templates.py b/deploy_templates.py
index 895e55a4..9745971a 100755
--- a/deploy_templates.py
+++ b/deploy_templates.py
@@ -17,7 +17,7 @@
 master_ram_kb = int(
   os.popen(mem_command).read().strip())
 # This is the master's memory. Try to find slave's memory as well
-first_slave = os.popen("cat /root/spark-ec2/slaves | head -1").read().strip()
+first_slave = os.popen("cat ~/spark-ec2/slaves | head -1").read().strip()
 
 slave_mem_command = "ssh -t -o StrictHostKeyChecking=no %s %s" %\
         (first_slave, mem_command)
@@ -80,11 +80,17 @@
   "aws_secret_access_key": os.getenv("AWS_SECRET_ACCESS_KEY"),
 }
 
-template_dir="/root/spark-ec2/templates"
+deploy_root_dir = os.getenv("DEPLOY_ROOT_DIR")
+
+template_dir=deploy_root_dir+"/spark-ec2/templates"
+tmp_template_dir=os.getenv("TMP_TEMPLATE_DIR")
 
 for path, dirs, files in os.walk(template_dir):
   if path.find(".svn") == -1:
     dest_dir = os.path.join('/', path[len(template_dir):])
+    if dest_dir.startswith("/root"):
+      dest_dir = deploy_root_dir + dest_dir[5:]
+    dest_dir = "/tmp/templates/" + dest_dir
     if not os.path.exists(dest_dir):
       os.makedirs(dest_dir)
     for filename in files:
@@ -92,7 +98,6 @@
         dest_file = os.path.join(dest_dir, filename)
         with open(os.path.join(path, filename)) as src:
           with open(dest_file, "w") as dest:
-            print("Configuring " + dest_file)
             text = src.read()
             for key in template_vars:
               text = text.replace("{{" + key + "}}", template_vars[key] or '')
diff --git a/ephemeral-hdfs/init.sh b/ephemeral-hdfs/init.sh
index 0e18bca8..62d756d9 100755
--- a/ephemeral-hdfs/init.sh
+++ b/ephemeral-hdfs/init.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 
-pushd /root > /dev/null
+pushd ~ > /dev/null
 
 if [ -d "ephemeral-hdfs" ]; then
   echo "Ephemeral HDFS seems to be installed. Exiting."
@@ -14,8 +14,8 @@ case "$HADOOP_MAJOR_VERSION" in
     tar xvzf hadoop-1.0.4.tar.gz > /tmp/spark-ec2_hadoop.log
     rm hadoop-*.tar.gz
     mv hadoop-1.0.4/ ephemeral-hdfs/
-    sed -i 's/-jvm server/-server/g' /root/ephemeral-hdfs/bin/hadoop
-    cp /root/hadoop-native/* /root/ephemeral-hdfs/lib/native/
+    sed -i 's/-jvm server/-server/g' ~/ephemeral-hdfs/bin/hadoop
+    cp ~/hadoop-native/* ~/ephemeral-hdfs/lib/native/
     ;;
   2) 
     wget http://s3.amazonaws.com/spark-related-packages/hadoop-2.0.0-cdh4.2.0.tar.gz  
@@ -25,9 +25,9 @@ case "$HADOOP_MAJOR_VERSION" in
     mv hadoop-2.0.0-cdh4.2.0/ ephemeral-hdfs/
 
     # Have single conf dir
-    rm -rf /root/ephemeral-hdfs/etc/hadoop/
-    ln -s /root/ephemeral-hdfs/conf /root/ephemeral-hdfs/etc/hadoop
-    cp /root/hadoop-native/* /root/ephemeral-hdfs/lib/native/
+    rm -rf ~/ephemeral-hdfs/etc/hadoop/
+    ln -s ~/ephemeral-hdfs/conf ~/ephemeral-hdfs/etc/hadoop
+    cp ~/hadoop-native/* ~/ephemeral-hdfs/lib/native/
     ;;
   yarn)
     wget http://s3.amazonaws.com/spark-related-packages/hadoop-2.4.0.tar.gz
@@ -37,14 +37,14 @@ case "$HADOOP_MAJOR_VERSION" in
     mv hadoop-2.4.0/ ephemeral-hdfs/
 
     # Have single conf dir
-    rm -rf /root/ephemeral-hdfs/etc/hadoop/
-    ln -s /root/ephemeral-hdfs/conf /root/ephemeral-hdfs/etc/hadoop
+    rm -rf ~/ephemeral-hdfs/etc/hadoop/
+    ln -s ~/ephemeral-hdfs/conf ~/ephemeral-hdfs/etc/hadoop
     ;;
 
   *)
      echo "ERROR: Unknown Hadoop version"
      return 1
 esac
-/root/spark-ec2/copy-dir /root/ephemeral-hdfs
+~/spark-ec2/copy-dir ~/ephemeral-hdfs
 
 popd > /dev/null
diff --git a/ephemeral-hdfs/setup-slave.sh b/ephemeral-hdfs/setup-slave.sh
index a85c4df7..f91a86a0 100755
--- a/ephemeral-hdfs/setup-slave.sh
+++ b/ephemeral-hdfs/setup-slave.sh
@@ -1,5 +1,15 @@
 #!/bin/bash
 
+#learn the linux distribution
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
+
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  [[ ! -e /var/hadoop ]] && sudo mkdir /var/hadoop
+  sudo chmod 777 /var/hadoop
+fi
+
 # Setup ephemeral-hdfs
 mkdir -p /mnt/ephemeral-hdfs/logs
 mkdir -p /mnt/hadoop-logs
@@ -24,3 +34,4 @@ create_hadoop_dirs /mnt
 create_hadoop_dirs /mnt2
 create_hadoop_dirs /mnt3
 create_hadoop_dirs /mnt4
+
diff --git a/ephemeral-hdfs/setup.sh b/ephemeral-hdfs/setup.sh
index 1c171056..5cdcb2d8 100755
--- a/ephemeral-hdfs/setup.sh
+++ b/ephemeral-hdfs/setup.sh
@@ -1,21 +1,27 @@
 #!/bin/bash
 
-EPHEMERAL_HDFS=/root/ephemeral-hdfs
+#learn the linux distribution
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
+
+EPHEMERAL_HDFS=~/ephemeral-hdfs
+USER=`whoami`
 
 # Set hdfs url to make it easier
 HDFS_URL="hdfs://$PUBLIC_DNS:9000"
 echo "export HDFS_URL=$HDFS_URL" >> ~/.bash_profile
 
-pushd /root/spark-ec2/ephemeral-hdfs > /dev/null
+pushd ~/spark-ec2/ephemeral-hdfs > /dev/null
 source ./setup-slave.sh
 
 for node in $SLAVES $OTHER_MASTERS; do
   echo $node
-  ssh -t -t $SSH_OPTS root@$node "/root/spark-ec2/ephemeral-hdfs/setup-slave.sh" & sleep 0.3
+  ssh -t -t $SSH_OPTS $USER@$node "~/spark-ec2/ephemeral-hdfs/setup-slave.sh" & sleep 0.3
 done
 wait
 
-/root/spark-ec2/copy-dir $EPHEMERAL_HDFS/conf
+~/spark-ec2/copy-dir $EPHEMERAL_HDFS/conf
 
 NAMENODE_DIR=/mnt/ephemeral-hdfs/dfs/name
 
@@ -31,12 +37,24 @@ echo "Starting ephemeral HDFS..."
 # This is different depending on version.
 case "$HADOOP_MAJOR_VERSION" in
   1)
+    if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+      echo "ERROR: Unsupported hadoop version on Ubuntu"
+      return -1
+    fi
     $EPHEMERAL_HDFS/bin/start-dfs.sh
     ;;
   2)
+    if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+      [[ ! -e /var/hadoop ]] && sudo mkdir /var/hadoop
+      sudo chmod 777 /var/hadoop
+    fi
     $EPHEMERAL_HDFS/sbin/start-dfs.sh
     ;;
   yarn) 
+    if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+      [[ ! -e /var/hadoop ]] && sudo mkdir /var/hadoop 
+      sudo chmod 777 /var/hadoop
+    fi
     $EPHEMERAL_HDFS/sbin/start-dfs.sh
     echo "Starting YARN"
     $EPHEMERAL_HDFS/sbin/start-yarn.sh
diff --git a/ganglia/init.sh b/ganglia/init.sh
old mode 100644
new mode 100755
index d120a236..7b5fdf72
--- a/ganglia/init.sh
+++ b/ganglia/init.sh
@@ -1,26 +1,50 @@
-#!/bin/bash
 
-# NOTE: Remove all rrds which might be around from an earlier run
-rm -rf /var/lib/ganglia/rrds/*
-rm -rf /mnt/ganglia/rrds/*
 
-# Make sure rrd storage directory has right permissions
-mkdir -p /mnt/ganglia/rrds
-chown -R nobody:nobody /mnt/ganglia/rrds
+USER=`whoami`
 
-# Install ganglia
-# TODO: Remove this once the AMI has ganglia by default
+#learn the linux distribution
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
 
-GANGLIA_PACKAGES="ganglia ganglia-web ganglia-gmond ganglia-gmetad"
+if [[ $DISTRIB_ID = "Centos" ]]; then
+
+  # NOTE: Remove all rrds which might be around from an earlier run
+  sudo rm -rf /var/lib/ganglia/rrds/*
+  sudo rm -rf /mnt/ganglia/rrds/*
+
+  # Make sure rrd storage directory has right permissions
+  mkdir -p /mnt/ganglia/rrds
+  chown -R nobody:nobody /mnt/ganglia/rrds
+
+  # Install ganglia
+  # TODO: Remove this once the AMI has ganglia by default
+
+
+  GANGLIA_PACKAGES="ganglia ganglia-web ganglia-gmond ganglia-gmetad"
+  if ! rpm --quiet -q $GANGLIA_PACKAGES; then
+    yum install -q -y $GANGLIA_PACKAGES;
+  fi
+  for node in $SLAVES $OTHER_MASTERS; do
+    ssh -t -t $SSH_OPTS root@$node "if ! rpm --quiet -q $GANGLIA_PACKAGES; then yum install -q -y $GANGLIA_PACKAGES; fi" & sleep 0.3
+  done
+  wait
+
+  # Post-package installation : Symlink /var/lib/ganglia/rrds to /mnt/ganglia/rrds
+  if [[ -d /var/lib/ganglia/rrds ]]; then sudo rmdir /var/lib/ganglia/rrds; fi
+  sudo ln -s /mnt/ganglia/rrds /var/lib/ganglia/rrds
+
+elif [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  echo "WARNING: Skipping ganglia on ubuntu..."
+  #GANGLIA_PACKAGES="ganglia-webfrontend ganglia-monitor gmetad"
+  #sudo apt-get install -y $GANGLIA_PACKAGES
+  #for node in $SLAVES $OTHER_MASTERS; do
+  #  ssh -t -t $SSH_OPTS $USER@$node "sudo apt-get install -y $GANGLIA_PACKAGES; sudo dpkg --configure -a" & sleep 0.3
+  #done
+  #wait
+
+  ## Post-package installation : Symlink /var/lib/ganglia/rrds to /mnt/ganglia/rrds
+  #if [[ -d /var/lib/ganglia/rrds ]]; then sudo rmdir /var/lib/ganglia/rrds; fi
+  #sudo ln -s /mnt/ganglia/rrds /var/lib/ganglia/rrds
 
-if ! rpm --quiet -q $GANGLIA_PACKAGES; then
-  yum install -q -y $GANGLIA_PACKAGES;
 fi
-for node in $SLAVES $OTHER_MASTERS; do
-  ssh -t -t $SSH_OPTS root@$node "if ! rpm --quiet -q $GANGLIA_PACKAGES; then yum install -q -y $GANGLIA_PACKAGES; fi" & sleep 0.3
-done
-wait
-
-# Post-package installation : Symlink /var/lib/ganglia/rrds to /mnt/ganglia/rrds
-rmdir /var/lib/ganglia/rrds
-ln -s /mnt/ganglia/rrds /var/lib/ganglia/rrds
diff --git a/ganglia/setup.sh b/ganglia/setup.sh
old mode 100644
new mode 100755
index 8719d528..738d6dee
--- a/ganglia/setup.sh
+++ b/ganglia/setup.sh
@@ -1,20 +1,33 @@
 #!/bin/bash
 
-/root/spark-ec2/copy-dir /etc/ganglia/
+USER=`whoami`
 
-# Start gmond everywhere
-/etc/init.d/gmond restart
+#learn the linux distribution
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
 
-for node in $SLAVES $OTHER_MASTERS; do
-  ssh -t -t $SSH_OPTS root@$node "/etc/init.d/gmond restart"
-done
+if [[ $DISTRIB_ID = "Centos" ]]; then
 
-# gmeta needs rrds to be owned by nobody
-chown -R nobody /var/lib/ganglia/rrds
-# cluster-wide aggregates only show up with this. TODO: Fix this cleanly ?
-ln -s /usr/share/ganglia/conf/default.json /var/lib/ganglia/conf/
+  /root/spark-ec2/copy-dir /etc/ganglia/
 
-/etc/init.d/gmetad restart
+  # Start gmond everywhere
+  /etc/init.d/gmond restart
 
-# Start http server to serve ganglia
-/etc/init.d/httpd restart
+  for node in $SLAVES $OTHER_MASTERS; do
+    ssh -t -t $SSH_OPTS root@$node "/etc/init.d/gmond restart"
+  done
+
+  # gmeta needs rrds to be owned by nobody
+  chown -R nobody /var/lib/ganglia/rrds
+  # cluster-wide aggregates only show up with this. TODO: Fix this cleanly ?
+  ln -s /usr/share/ganglia/conf/default.json /var/lib/ganglia/conf/
+
+  /etc/init.d/gmetad restart
+
+  # Start http server to serve ganglia
+  /etc/init.d/httpd restart
+
+elif [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  echo "WARNING: Skipping ganglia on ubuntu..."
+fi
diff --git a/install-java-on-ubuntu.sh b/install-java-on-ubuntu.sh
new file mode 100755
index 00000000..f48df69f
--- /dev/null
+++ b/install-java-on-ubuntu.sh
@@ -0,0 +1,10 @@
+#!/bin/bash
+
+if [[ "x$JAVA_HOME" == "x" ]]; then
+  #also install java on the first master since stock ubuntu ami does not
+  #come with java pre-installed; also install git
+  sudo apt-get update -q
+  sudo apt-get install -y -q openjdk-7-jdk
+  sudo sh -c 'echo "export JAVA_HOME=/usr/lib/jvm/java-7-openjdk-amd64" >> /etc/environment'
+  source /etc/environment
+fi
diff --git a/mapreduce/init.sh b/mapreduce/init.sh
index 2e952799..90dda056 100755
--- a/mapreduce/init.sh
+++ b/mapreduce/init.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 
-pushd /root > /dev/null
+pushd ~ > /dev/null
 case "$HADOOP_MAJOR_VERSION" in
   1)
     echo "Nothing to initialize for MapReduce in Hadoop 1"
@@ -19,5 +19,5 @@ case "$HADOOP_MAJOR_VERSION" in
      echo "ERROR: Unknown Hadoop version"
      return -1
 esac
-/root/spark-ec2/copy-dir /root/mapreduce
+~/spark-ec2/copy-dir ~/mapreduce
 popd > /dev/null
diff --git a/mapreduce/setup.sh b/mapreduce/setup.sh
index 4f71e0c7..4625c294 100755
--- a/mapreduce/setup.sh
+++ b/mapreduce/setup.sh
@@ -1,11 +1,12 @@
 #!/bin/bash
-MAPREDUCE=/root/mapreduce
+MAPREDUCE=~/mapreduce
+USER=`whoami`
 
 mkdir -p /mnt/mapreduce/logs
 for node in $SLAVES $OTHER_MASTERS; do
-  ssh -t $SSH_OPTS root@$node "mkdir -p /mnt/mapreduce/logs && chown hadoop:hadoop /mnt/mapreduce/logs && chown hadoop:hadoop /mnt/mapreduce" & sleep 0.3
+  ssh -t $SSH_OPTS $USER@$node "mkdir -p /mnt/mapreduce/logs && chown hadoop:hadoop /mnt/mapreduce/logs && chown hadoop:hadoop /mnt/mapreduce" & sleep 0.3
 done
 wait
 
 chown hadoop:hadoop /mnt/mapreduce -R
-/root/spark-ec2/copy-dir $MAPREDUCE/conf
+~/spark-ec2/copy-dir $MAPREDUCE/conf
diff --git a/persistent-hdfs/init.sh b/persistent-hdfs/init.sh
index 735cebcc..14d901ff 100755
--- a/persistent-hdfs/init.sh
+++ b/persistent-hdfs/init.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 
-pushd /root > /dev/null
+pushd ~ > /dev/null
 
 if [ -d "persistent-hdfs" ]; then
   echo "Persistent HDFS seems to be installed. Exiting."
@@ -14,7 +14,7 @@ case "$HADOOP_MAJOR_VERSION" in
     tar xvzf hadoop-1.0.4.tar.gz > /tmp/spark-ec2_hadoop.log
     rm hadoop-*.tar.gz
     mv hadoop-1.0.4/ persistent-hdfs/
-    cp /root/hadoop-native/* /root/persistent-hdfs/lib/native/
+    cp ~/hadoop-native/* ~/persistent-hdfs/lib/native/
     ;;
   2)
     wget http://s3.amazonaws.com/spark-related-packages/hadoop-2.0.0-cdh4.2.0.tar.gz
@@ -24,9 +24,9 @@ case "$HADOOP_MAJOR_VERSION" in
     mv hadoop-2.0.0-cdh4.2.0/ persistent-hdfs/
 
     # Have single conf dir
-    rm -rf /root/persistent-hdfs/etc/hadoop/
-    ln -s /root/persistent-hdfs/conf /root/persistent-hdfs/etc/hadoop
-    cp /root/hadoop-native/* /root/persistent-hdfs/lib/native/
+    rm -rf ~/persistent-hdfs/etc/hadoop/
+    ln -s ~/persistent-hdfs/conf ~/persistent-hdfs/etc/hadoop
+    cp ~/hadoop-native/* ~/persistent-hdfs/lib/native/
     ;;
   yarn)
     wget http://s3.amazonaws.com/spark-related-packages/hadoop-2.4.0.tar.gz
@@ -36,14 +36,14 @@ case "$HADOOP_MAJOR_VERSION" in
     mv hadoop-2.4.0/ persistent-hdfs/
 
     # Have single conf dir
-    rm -rf /root/persistent-hdfs/etc/hadoop/
-    ln -s /root/persistent-hdfs/conf /root/persistent-hdfs/etc/hadoop
+    rm -rf ~/persistent-hdfs/etc/hadoop/
+    ln -s ~/persistent-hdfs/conf ~/persistent-hdfs/etc/hadoop
     ;;
 
   *)
      echo "ERROR: Unknown Hadoop version"
      return 1
 esac
-/root/spark-ec2/copy-dir /root/persistent-hdfs
+~/spark-ec2/copy-dir ~/persistent-hdfs
 
 popd > /dev/null
diff --git a/persistent-hdfs/setup.sh b/persistent-hdfs/setup.sh
index d1713e12..fa7b1bb9 100755
--- a/persistent-hdfs/setup.sh
+++ b/persistent-hdfs/setup.sh
@@ -1,22 +1,32 @@
 #!/bin/bash
 
-PERSISTENT_HDFS=/root/persistent-hdfs
+PERSISTENT_HDFS=~/persistent-hdfs
+USER=`whoami`
 
-pushd /root/spark-ec2/persistent-hdfs > /dev/null
-source ./setup-slave.sh
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
 
-for node in $SLAVES $OTHER_MASTERS; do
-  ssh -t $SSH_OPTS root@$node "/root/spark-ec2/persistent-hdfs/setup-slave.sh" & sleep 0.3
-done
-wait
+if [[ $DISTRIB_ID = "Centos" ]]; then
+  pushd ~/spark-ec2/persistent-hdfs > /dev/null
+  source ./setup-slave.sh
 
-/root/spark-ec2/copy-dir $PERSISTENT_HDFS/conf
+  for node in $SLAVES $OTHER_MASTERS; do
+    ssh -t $SSH_OPTS $USER@$node "~/spark-ec2/persistent-hdfs/setup-slave.sh" & sleep 0.3
+  done
+  wait
 
-if [[ ! -e /vol/persistent-hdfs/dfs/name ]] ; then
-  echo "Formatting persistent HDFS namenode..."
-  $PERSISTENT_HDFS/bin/hadoop namenode -format
-fi
+  ~/spark-ec2/copy-dir $PERSISTENT_HDFS/conf
+
+  if [[ ! -e /vol/persistent-hdfs/dfs/name ]] ; then
+    echo "Formatting persistent HDFS namenode..."
+    $PERSISTENT_HDFS/bin/hadoop namenode -format
+  fi
 
-echo "Persistent HDFS installed, won't start by default..."
+  echo "Persistent HDFS installed, won't start by default..."
 
-popd > /dev/null
+  popd > /dev/null
+
+else
+  echo "Skipping persistent hdfs setup on $DISTRIB_ID"
+fi
diff --git a/resolve-hostname.sh b/resolve-hostname.sh
index 1fe6deea..881954d8 100755
--- a/resolve-hostname.sh
+++ b/resolve-hostname.sh
@@ -21,7 +21,7 @@ PRIVATE_IP=`wget -q -O - http://169.254.169.254/latest/meta-data/local-ipv4`
 # do changes only if short hostname does not resolve
 ping -c 1 -q "${SHORT_HOSTNAME}" > /dev/null 2>&1
 if [ $? -ne 0  ]; then
-    echo -e "\n# fixed by resolve-hostname.sh \n${PRIVATE_IP} ${SHORT_HOSTNAME}\n" >> /etc/hosts
+    sudo echo -e "\n# fixed by resolve-hostname.sh \n${PRIVATE_IP} ${SHORT_HOSTNAME}\n" >> /etc/hosts
 
     # let's make sure that it got fixed
     ping -c 1 -q "${SHORT_HOSTNAME}" > /dev/null 2>&1
diff --git a/rstudio/init.sh b/rstudio/init.sh
old mode 100644
new mode 100755
index fd18d6be..184f2cc9
--- a/rstudio/init.sh
+++ b/rstudio/init.sh
@@ -1,17 +1,29 @@
 #!/usr/bin/env bash
 
-# download rstudio 
-wget http://download2.rstudio.org/rstudio-server-rhel-0.99.446-x86_64.rpm
-sudo yum install --nogpgcheck -y rstudio-server-rhel-0.99.446-x86_64.rpm
+#learn the linux distribution
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
 
-# restart rstudio 
-rstudio-server restart 
+if [[ $DISTRIB_ID = "Centos" ]]; then
 
-# add user for rstudio, user needs to supply password later on
-adduser rstudio
+  # download rstudio 
+  wget http://download2.rstudio.org/rstudio-server-rhel-0.99.446-x86_64.rpm
+  sudo yum install --nogpgcheck -y rstudio-server-rhel-0.99.446-x86_64.rpm
+
+  # restart rstudio 
+  rstudio-server restart 
+
+  # add user for rstudio, user needs to supply password later on
+  adduser rstudio
+
+  # create a Rscript that connects to Spark, to help starting user
+  cp /root/spark-ec2/rstudio/startSpark.R /home/rstudio
+
+elif [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  echo "WARNING: Skipping rstudio installation on Ubuntu"
+fi
 
-# create a Rscript that connects to Spark, to help starting user
-cp /root/spark-ec2/rstudio/startSpark.R /home/rstudio
 
 # make sure that the temp dirs exist and can be written to by any user
 # otherwise this will create a conflict for the rstudio user
@@ -23,7 +35,7 @@ function create_temp_dirs {
   chmod a+w $location
 }
 
-create_temp_dirs /mnt/spark
-create_temp_dirs /mnt2/spark
-create_temp_dirs /mnt3/spark
-create_temp_dirs /mnt4/spark
+if [[ -d /mnt ]]; then create_temp_dirs /mnt/spark; fi
+if [[ -d /mnt2 ]]; then create_temp_dirs /mnt2/spark; fi
+if [[ -d /mnt3 ]]; then create_temp_dirs /mnt3/spark; fi
+if [[ -d /mnt4 ]]; then create_temp_dirs /mnt4/spark; fi
diff --git a/scala/init.sh b/scala/init.sh
index 73a299f5..39a1f16b 100755
--- a/scala/init.sh
+++ b/scala/init.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 
-pushd /root > /dev/null
+pushd ~ > /dev/null
 
 if [ -d "scala" ]; then
   echo "Scala seems to be installed. Exiting."
diff --git a/scala/setup.sh b/scala/setup.sh
index 6aa5d27b..7cb3b3d1 100755
--- a/scala/setup.sh
+++ b/scala/setup.sh
@@ -1,3 +1,3 @@
 #!/bin/bash
 
-/root/spark-ec2/copy-dir /root/scala
+~/spark-ec2/copy-dir ~/scala
diff --git a/setup-slave.sh b/setup-slave.sh
index 76372d9a..e9d5a822 100755
--- a/setup-slave.sh
+++ b/setup-slave.sh
@@ -1,56 +1,86 @@
 #!/bin/bash
 
+#learn the current user
+USER=`whoami`
+
+#learn the linux distribution
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
+
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  sudo apt-get install -y -q git
+fi
+
 # Disable Transparent Huge Pages (THP)
 # THP can result in system thrashing (high sys usage) due to frequent defrags of memory.
 # Most systems recommends turning THP off.
 if [[ -e /sys/kernel/mm/transparent_hugepage/enabled ]]; then
-  echo never > /sys/kernel/mm/transparent_hugepage/enabled
+  sudo sh -c 'echo never > /sys/kernel/mm/transparent_hugepage/enabled'
 fi
 
 # Make sure we are in the spark-ec2 directory
-pushd /root/spark-ec2 > /dev/null
+pushd ~/spark-ec2 > /dev/null
+
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  ./install-java-on-ubuntu.sh
+  source /etc/environment
+fi
 
 source ec2-variables.sh
 
 # Set hostname based on EC2 private DNS name, so that it is set correctly
 # even if the instance is restarted with a different private DNS name
 PRIVATE_DNS=`wget -q -O - http://169.254.169.254/latest/meta-data/local-hostname`
-hostname $PRIVATE_DNS
-echo $PRIVATE_DNS > /etc/hostname
+sudo hostname $PRIVATE_DNS
+sudo sh -c "echo $PRIVATE_DNS > /etc/hostname"
 HOSTNAME=$PRIVATE_DNS  # Fix the bash built-in hostname variable too
 
 echo "checking/fixing resolution of hostname"
-bash /root/spark-ec2/resolve-hostname.sh
+bash ~/spark-ec2/resolve-hostname.sh
 
 # Work around for R3 or I2 instances without pre-formatted ext3 disks
 instance_type=$(curl http://169.254.169.254/latest/meta-data/instance-type 2> /dev/null)
 
 echo "Setting up slave on `hostname`... of type $instance_type"
 
+create_ephemeral_blkdev_links() {
+  device_letter=$1
+  devx=/dev/xvd${device_letter}
+  devs=/dev/sd${device_letter}
+  if [[ -e $devx && ! -e $devs ]]; then sudo ln -s $devx $devs; fi
+}
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  create_ephemeral_blkdev_links b
+  create_ephemeral_blkdev_links c
+  create_ephemeral_blkdev_links d
+fi
+
 if [[ $instance_type == r3* || $instance_type == i2* || $instance_type == hi1* ]]; then
   # Format & mount using ext4, which has the best performance among ext3, ext4, and xfs based
   # on our shuffle heavy benchmark
   EXT4_MOUNT_OPTS="defaults,noatime,nodiratime"
-  rm -rf /mnt*
-  mkdir /mnt
+  sudo rm -rf /mnt*
+  sudo mkdir /mnt
   # To turn TRIM support on, uncomment the following line.
   #echo '/dev/sdb /mnt  ext4  defaults,noatime,nodiratime,discard 0 0' >> /etc/fstab
-  mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/sdb
-  mount -o $EXT4_MOUNT_OPTS /dev/sdb /mnt
+  sudo mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/sdb
+  sudo mount -o $EXT4_MOUNT_OPTS /dev/sdb /mnt
+  echo "Changing slave /mnt mode to 777"
 
   if [[ $instance_type == "r3.8xlarge" || $instance_type == "hi1.4xlarge" ]]; then
-    mkdir /mnt2
+    sudo mkdir /mnt2
     # To turn TRIM support on, uncomment the following line.
     #echo '/dev/sdc /mnt2  ext4  defaults,noatime,nodiratime,discard 0 0' >> /etc/fstab
     if [[ $instance_type == "r3.8xlarge" ]]; then
-      mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/sdc      
-      mount -o $EXT4_MOUNT_OPTS /dev/sdc /mnt2
+      sudo mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/sdc      
+      sudo mount -o $EXT4_MOUNT_OPTS /dev/sdc /mnt2
     fi
     # To turn TRIM support on, uncomment the following line.
     #echo '/dev/sdf /mnt2  ext4  defaults,noatime,nodiratime,discard 0 0' >> /etc/fstab
     if [[ $instance_type == "hi1.4xlarge" ]]; then
-      mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/sdf      
-      mount -o $EXT4_MOUNT_OPTS /dev/sdf /mnt2
+      sudo mkfs.ext4 -E lazy_itable_init=0,lazy_journal_init=0 /dev/sdf      
+      sudo mount -o $EXT4_MOUNT_OPTS /dev/sdf /mnt2
     fi    
   fi
 fi
@@ -59,29 +89,33 @@ fi
 # are ext3, but we use xfs for EBS volumes to format them faster)
 XFS_MOUNT_OPTS="defaults,noatime,nodiratime,allocsize=8m"
 
-function setup_ebs_volume {
+setup_ebs_volume() {
   device=$1
   mount_point=$2
   if [[ -e $device ]]; then
     # Check if device is already formatted
     if ! blkid $device; then
-      mkdir $mount_point
-      yum install -q -y xfsprogs
-      if mkfs.xfs -q $device; then
-        mount -o $XFS_MOUNT_OPTS $device $mount_point
-        chmod -R a+w $mount_point
+      sudo mkdir $mount_point
+      if [[ DISTRIB_ID = "Centos" ]]; then
+	yum install -q -y xfsprogs
+      elif [[ DISTRIB_ID = "Ubuntu" ]]; then
+	sudo apt-get install -y -q xfsprogs
+      fi
+      if sudo mkfs.xfs -q $device; then
+        sudo mount -o $XFS_MOUNT_OPTS $device $mount_point
+        sudo chmod -R a+w $mount_point
       else
         # mkfs.xfs is not installed on this machine or has failed;
         # delete /vol so that the user doesn't think we successfully
         # mounted the EBS volume
-        rmdir $mount_point
+        sudo rmdir $mount_point
       fi
     else
       # EBS volume is already formatted. Mount it if its not mounted yet.
       if ! grep -qs '$mount_point' /proc/mounts; then
-        mkdir $mount_point
-        mount -o $XFS_MOUNT_OPTS $device $mount_point
-        chmod -R a+w $mount_point
+        sudo mkdir $mount_point
+        sudo mount -o $XFS_MOUNT_OPTS $device $mount_point
+        sudo chmod -R a+w $mount_point
       fi
     fi
   fi
@@ -100,35 +134,47 @@ setup_ebs_volume /dev/sdz /vol7
 # Alias vol to vol3 for backward compatibility: the old spark-ec2 script supports only attaching
 # one EBS volume at /dev/sdv.
 if [[ -e /vol3 && ! -e /vol ]]; then
-  ln -s /vol3 /vol
+  sudo ln -s /vol3 /vol
 fi
 
 # Make data dirs writable by non-root users, such as CDH's hadoop user
-chmod -R a+w /mnt*
+sudo chmod -R a+w /mnt*
 
 # Remove ~/.ssh/known_hosts because it gets polluted as you start/stop many
 # clusters (new machines tend to come up under old hostnames)
-rm -f /root/.ssh/known_hosts
+rm -f ~/.ssh/known_hosts
 
 # Create swap space on /mnt
-/root/spark-ec2/create-swap.sh $SWAP_MB
+sudo ~/spark-ec2/create-swap.sh $SWAP_MB
 
 # Allow memory to be over committed. Helps in pyspark where we fork
-echo 1 > /proc/sys/vm/overcommit_memory
+sudo sh -c 'echo 1 > /proc/sys/vm/overcommit_memory'
 
 # Add github to known hosts to get git@github.com clone to work
 # TODO(shivaram): Avoid duplicate entries ?
-cat /root/spark-ec2/github.hostkey >> /root/.ssh/known_hosts
+cat ~/spark-ec2/github.hostkey >> ~/.ssh/known_hosts
 
 # Create /usr/bin/realpath which is used by R to find Java installations
 # NOTE: /usr/bin/realpath is missing in CentOS AMIs. See
 # http://superuser.com/questions/771104/usr-bin-realpath-not-found-in-centos-6-5
-echo '#!/bin/bash' > /usr/bin/realpath
-echo 'readlink -e "$@"' >> /usr/bin/realpath
-chmod a+x /usr/bin/realpath
+if [[ $DISTRIB_ID = "CentOS" ]]; then
+  echo '#!/bin/bash' > /usr/bin/realpath
+  echo 'readlink -e "$@"' >> /usr/bin/realpath
+  chmod a+x /usr/bin/realpath
+elif [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  sudo apt-get install -y -q realpath
+fi
+
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  [[ -d /mnt ]] && sudo chmod 777 /mnt
+  [[ -d /mnt2 ]] && sudo chmod 777 /mnt2
+  [[ -d /mnt3 ]] && sudo chmod 777 /mnt3
+  [[ -d /mnt4 ]] && sudo chmod 777 /mnt4
+fi
+
 
 popd > /dev/null
 
 # this is to set the ulimit for root and other users
-echo '* soft nofile 1000000' >> /etc/security/limits.conf
-echo '* hard nofile 1000000' >> /etc/security/limits.conf
\ No newline at end of file
+sudo sh -c "echo '* soft nofile 1000000' >> /etc/security/limits.conf"
+sudo sh -c "echo '* hard nofile 1000000' >> /etc/security/limits.conf"
diff --git a/setup.sh b/setup.sh
index 5a3beea2..c97cacfa 100755
--- a/setup.sh
+++ b/setup.sh
@@ -1,6 +1,19 @@
 #!/bin/bash
 
-sudo yum install -y -q pssh
+#learn the current user
+USER=`whoami`
+
+#learn the linux distribution
+DISTRIB_ID=Centos
+if [[ -e /etc/lsb-release ]]; then source /etc/lsb-release; fi
+echo "DISTRIB_ID=$DISTRIB_ID"
+
+if [[ $DISTRIB_ID = "Centos" ]]; then
+  sudo yum install -y -q pssh
+elif [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  sudo apt-get install -y -q pssh
+  sudo apt-get install -y -q git
+fi
 
 # usage: echo_time_diff name start_time end_time
 echo_time_diff () {
@@ -11,20 +24,24 @@ echo_time_diff () {
 }
 
 # Make sure we are in the spark-ec2 directory
-pushd /root/spark-ec2 > /dev/null
+pushd ~/spark-ec2 > /dev/null
+
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  ./install-java-on-ubuntu.sh
+  source /etc/environment
+fi
 
 # Load the environment variables specific to this AMI
-source /root/.bash_profile
+if [[ -e ~/.bash_profile ]]; then source ~/.bash_profile; fi
 
 # Load the cluster variables set by the deploy script
-source ec2-variables.sh
-
+source ec2-variables.sh 
 # Set hostname based on EC2 private DNS name, so that it is set correctly
 # even if the instance is restarted with a different private DNS name
 PRIVATE_DNS=`wget -q -O - http://169.254.169.254/latest/meta-data/local-hostname`
 PUBLIC_DNS=`wget -q -O - http://169.254.169.254/latest/meta-data/hostname`
-hostname $PRIVATE_DNS
-echo $PRIVATE_DNS > /etc/hostname
+sudo hostname $PRIVATE_DNS
+sudo sh -c "echo $PRIVATE_DNS > /etc/hostname"
 export HOSTNAME=$PRIVATE_DNS  # Fix the bash built-in hostname variable too
 
 echo "Setting up Spark on `hostname`..."
@@ -52,35 +69,66 @@ fi
 echo "Setting executable permissions on scripts..."
 find . -regex "^.+.\(sh\|py\)" | xargs chmod a+x
 
-echo "RSYNC'ing /root/spark-ec2 to other cluster nodes..."
+echo "RSYNC'ing ~/spark-ec2 to other cluster nodes..."
 rsync_start_time="$(date +'%s')"
 for node in $SLAVES $OTHER_MASTERS; do
   echo $node
-  rsync -e "ssh $SSH_OPTS" -az /root/spark-ec2 $node:/root &
+  rsync -e "ssh $SSH_OPTS" -az ~/spark-ec2 $node:~ &
   scp $SSH_OPTS ~/.ssh/id_rsa $node:.ssh &
   sleep 0.1
 done
 wait
 rsync_end_time="$(date +'%s')"
-echo_time_diff "rsync /root/spark-ec2" "$rsync_start_time" "$rsync_end_time"
+echo_time_diff "rsync ~/spark-ec2" "$rsync_start_time" "$rsync_end_time"
+
+create_ephemeral_blkdev_links() {
+  device_letter=$1
+  devx=/dev/xvd${device_letter}
+  devs=/dev/sd${device_letter}
+  if [[ -e $devx && ! -e $devs ]]; then sudo ln -s $devx $devs; fi
+}
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  create_ephemeral_blkdev_links b
+  create_ephemeral_blkdev_links c
+  create_ephemeral_blkdev_links d
+fi
+
+if [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  [[ -d /mnt ]] && sudo chmod 777 /mnt
+  [[ -d /mnt2 ]] && sudo chmod 777 /mnt2
+  [[ -d /mnt3 ]] && sudo chmod 777 /mnt3
+  [[ -d /mnt4 ]] && sudo chmod 777 /mnt4
+fi
 
 echo "Running setup-slave on all cluster nodes to mount filesystems, etc..."
 setup_slave_start_time="$(date +'%s')"
-pssh --inline \
+if [[ $DISTRIB_ID = "Centos" ]]; then
+  pssh --inline \
+    --host "$MASTERS $SLAVES" \
+    --user $USER \
+    --extra-args "-t -t $SSH_OPTS" \
+    --timeout 0 \
+    "spark-ec2/setup-slave.sh"
+elif [[ $DISTRIB_ID = "Ubuntu" ]]; then
+  parallel-ssh --inline \
     --host "$MASTERS $SLAVES" \
-    --user root \
+    --user $USER \
     --extra-args "-t -t $SSH_OPTS" \
     --timeout 0 \
     "spark-ec2/setup-slave.sh"
+fi
+
 setup_slave_end_time="$(date +'%s')"
 echo_time_diff "setup-slave" "$setup_slave_start_time" "$setup_slave_end_time"
 
+
 # Always include 'scala' module if it's not defined as a work around
 # for older versions of the scripts.
 if [[ ! $MODULES =~ *scala* ]]; then
   MODULES=$(printf "%s\n%s\n" "scala" $MODULES)
 fi
 
+
 # Install / Init module
 for module in $MODULES; do
   echo "Initializing $module"
@@ -90,18 +138,32 @@ for module in $MODULES; do
   fi
   module_init_end_time="$(date +'%s')"
   echo_time_diff "$module init" "$module_init_start_time" "$module_init_end_time"
-  cd /root/spark-ec2  # guard against init.sh changing the cwd
+  cd ~/spark-ec2  # guard against init.sh changing the cwd
 done
 
+
 # Deploy templates
 # TODO: Move configuring templates to a per-module ?
 echo "Creating local config files..."
+
+export DEPLOY_ROOT_DIR=~
+export TMP_TEMPLATE_DIR="/tmp/templates/"
+mkdir $TMP_TEMPLATE_DIR
+
 ./deploy_templates.py
+find $TMP_TEMPLATE_DIR -type f > conflist
+for f in `cat conflist`; do
+  outf=`echo $f | sed "s/\/tmp\/templates//"`
+  dir=`dirname $outf`
+  [[ ! -e $dir ]] && sudo mkdir -p $dir
+  sudo mv $f $outf
+done
+
 
 # Copy spark conf by default
 echo "Deploying Spark config files..."
-chmod u+x /root/spark/conf/spark-env.sh
-/root/spark-ec2/copy-dir /root/spark/conf
+chmod u+x ~/spark/conf/spark-env.sh
+~/spark-ec2/copy-dir ~/spark/conf
 
 # Setup each module
 for module in $MODULES; do
@@ -111,7 +173,7 @@ for module in $MODULES; do
   sleep 0.1
   module_setup_end_time="$(date +'%s')"
   echo_time_diff "$module setup" "$module_setup_start_time" "$module_setup_end_time"
-  cd /root/spark-ec2  # guard against setup.sh changing the cwd
+  cd ~/spark-ec2  # guard against setup.sh changing the cwd
 done
 
 popd > /dev/null
diff --git a/spark-standalone/setup.sh b/spark-standalone/setup.sh
index e9c04c1f..4a252718 100755
--- a/spark-standalone/setup.sh
+++ b/spark-standalone/setup.sh
@@ -1,33 +1,31 @@
 #!/bin/bash
 
-BIN_FOLDER="/root/spark/sbin"
-
 if [[ "0.7.3 0.8.0 0.8.1" =~ $SPARK_VERSION ]]; then
-  BIN_FOLDER="/root/spark/bin"
+  BIN_FOLDER="~/spark/bin"
 fi
 
 # Copy the slaves to spark conf
-cp /root/spark-ec2/slaves /root/spark/conf/
-/root/spark-ec2/copy-dir /root/spark/conf
+cp ~/spark-ec2/slaves ~/spark/conf/
+~/spark-ec2/copy-dir ~/spark/conf
 
 # Set cluster-url to standalone master
-echo "spark://""`cat /root/spark-ec2/masters`"":7077" > /root/spark-ec2/cluster-url
-/root/spark-ec2/copy-dir /root/spark-ec2
+echo "spark://""`cat ~/spark-ec2/masters`"":7077" > ~/spark-ec2/cluster-url
+~/spark-ec2/copy-dir ~/spark-ec2
 
 # The Spark master seems to take time to start and workers crash if
 # they start before the master. So start the master first, sleep and then start
 # workers.
 
 # Stop anything that is running
-$BIN_FOLDER/stop-all.sh
+~/spark/sbin/stop-all.sh
 
 sleep 2
 
 # Start Master
-$BIN_FOLDER/start-master.sh
+~/spark/sbin/start-master.sh
 
 # Pause
 sleep 20
 
 # Start Workers
-$BIN_FOLDER/start-slaves.sh
+~/spark/sbin/start-slaves.sh
diff --git a/spark/init.sh b/spark/init.sh
index 71fbc7bf..6497b868 100755
--- a/spark/init.sh
+++ b/spark/init.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 
-pushd /root > /dev/null
+pushd ~ > /dev/null
 
 if [ -d "spark" ]; then
   echo "Spark seems to be installed. Exiting."
diff --git a/spark/setup.sh b/spark/setup.sh
index 9035d438..8e70f710 100755
--- a/spark/setup.sh
+++ b/spark/setup.sh
@@ -1,3 +1,3 @@
 #!/bin/bash
 
-/root/spark-ec2/copy-dir /root/spark
+~/spark-ec2/copy-dir ~/spark
diff --git a/spark_ec2.py b/spark_ec2.py
old mode 100644
new mode 100755
index 28d72f43..6120bf09
--- a/spark_ec2.py
+++ b/spark_ec2.py
@@ -42,6 +42,7 @@
 from datetime import datetime
 from optparse import OptionParser
 from sys import stderr
+from collections import namedtuple
 
 if sys.version < "3":
     from urllib2 import urlopen, Request, HTTPError
@@ -177,7 +178,7 @@ def parse_args():
         prog="spark-ec2",
         version="%prog {v}".format(v=SPARK_EC2_VERSION),
         usage="%prog [options] <action> <cluster_name>\n\n"
-        + "<action> can be: launch, destroy, login, stop, start, get-master, reboot-slaves")
+        + "<action> can be: create, launch, destroy, login, stop, start, get-master, reboot-slaves")
 
     parser.add_option(
         "-s", "--slaves", type="int", default=1,
@@ -210,6 +211,11 @@ def parse_args():
         help="Availability zone to launch instances in, or 'all' to spread " +
              "slaves across multiple (an additional $0.01/Gb for bandwidth" +
              "between zones applies) (default: a single zone chosen at random)")
+    parser.add_option(
+        "--ami-type", default="spark",
+        help="Type of ami to use (default: %default). " +
+             "Valid options are %default and ubuntu. " +
+	     "If you specify an ami, the ami-type option will be ignored.")
     parser.add_option(
         "-a", "--ami",
         help="Amazon Machine Image ID to use")
@@ -331,6 +337,12 @@ def parse_args():
     parser.add_option(
         "--instance-profile-name", default=None,
         help="IAM profile name to launch instances under")
+    parser.add_option(
+        "--elastic-ip", default=None,
+        help="Elastic IP to associate with the master")
+    parser.add_option(
+        "--no-setup", action="store_true", default=False,
+        help="Do not run the usual setup commands")
 
     (opts, args) = parser.parse_args()
     if len(args) != 2:
@@ -466,7 +478,14 @@ def get_spark_ami(opts):
         r=opts.spark_ec2_git_repo.replace("https://github.com", "https://raw.github.com", 1),
         b=opts.spark_ec2_git_branch)
 
-    ami_path = "%s/%s/%s" % (ami_prefix, opts.region, instance_type)
+    if opts.ami_type == "spark":
+      ami_path = "%s/%s/%s" % (ami_prefix, opts.region, instance_type)
+    elif opts.ami_type == "ubuntu":
+      ami_path = "%s-ubuntu/ami-list-ubuntu-amd64" % ami_prefix
+    else:
+      print("Bad ami type")
+      sys.exit(1)
+
     reader = codecs.getreader("ascii")
     try:
         ami = reader(urlopen(ami_path)).read().strip()
@@ -474,6 +493,26 @@ def get_spark_ami(opts):
         print("Could not resolve AMI at: " + ami_path, file=stderr)
         sys.exit(1)
 
+    if opts.ami_type == "ubuntu":
+      storage_type = "ebs"	#"", "ebs", "ebs-io1", "ebs-ssd"
+      AMItup = namedtuple('AMI', ['region','release','version','arch','hwtype','date','ami','hyper'])
+      amil = ami.split('\n')
+      amill = [l.split(',') for l in amil]
+      amilt = [AMItup._make(l) for l in amill]
+      if instance_type == 'hvm':
+	hwtype = 'hvm:' + storage_type
+	amilt = [t for t in amilt if t.region == opts.region and t.hyper == 'hvm' and t.hwtype == hwtype]
+      else:
+	hwtype = storage_type
+	amilt = [t for t in amilt if t.region == opts.region and t.hyper != 'hvm' and t.hwtype == hwtype]
+      if len(amilt) != 1:
+        print("%d AMIs found, needed exactly one" % len(amilt))
+	for t in amilt:
+	  print(t)
+	sys.exit(1)
+      ami = amilt[0].ami
+
+
     print("Spark AMI: " + ami)
     return ami
 
@@ -483,6 +522,7 @@ def get_spark_ami(opts):
 # Returns a tuple of EC2 reservation objects for the master and slaves
 # Fails if there already instances running in the cluster's groups.
 def launch_cluster(conn, opts, cluster_name):
+
     if opts.identity_file is None:
         print("ERROR: Must provide an identity file (-i) for ssh connections.", file=stderr)
         sys.exit(1)
@@ -599,6 +639,16 @@ def launch_cluster(conn, opts, cluster_name):
             device.delete_on_termination = True
             block_map["/dev/sd" + chr(ord('s') + i)] = device
 
+    # for vanilla ubuntu AMIs, list out ephemeral block devices so they
+    # attached to be mounted on /mnt*
+    if opts.user == "ubuntu":
+        for i in range(get_num_disks(opts.instance_type)):
+            dev = BlockDeviceType()
+            dev.ephemeral_name = 'ephemeral%d' % i
+            # The first ephemeral drive is /dev/sdb.
+            name = '/dev/sd' + string.ascii_letters[i + 1]
+            block_map[name] = dev
+
     # AWS ignores the AMI-specified block device mapping for M3 (see SPARK-3342).
     if opts.instance_type.startswith('m3.'):
         for i in range(get_num_disks(opts.instance_type)):
@@ -810,6 +860,11 @@ def setup_cluster(conn, master_nodes, slave_nodes, opts, deploy_ssh_key):
             print(slave_address)
             ssh_write(slave_address, opts, ['tar', 'x'], dot_ssh_tar)
 
+    #install git on vanilla ubuntu ami
+    if opts.user == "ubuntu":
+        git_install = "sudo apt-get install -y -q git"
+        ssh(master, opts, git_install)
+
     modules = ['spark', 'ephemeral-hdfs', 'persistent-hdfs',
                'mapreduce', 'spark-standalone', 'tachyon', 'rstudio']
 
@@ -836,10 +891,17 @@ def setup_cluster(conn, master_nodes, slave_nodes, opts, deploy_ssh_key):
                                                   b=opts.spark_ec2_git_branch)
     )
 
+
     print("Deploying files to master...")
+    if opts.user == "root":
+        root_dir=SPARK_EC2_DIR + "/" + "deploy.generic"
+    elif opts.user == "ubuntu":
+        root_dir=SPARK_EC2_DIR + "/" + "deploy.ubuntu"
+    else:
+        root_dir=SPARK_EC2_DIR + "/" + "deploy.generic"
     deploy_files(
         conn=conn,
-        root_dir=SPARK_EC2_DIR + "/" + "deploy.generic",
+        root_dir=root_dir,
         opts=opts,
         master_nodes=master_nodes,
         slave_nodes=slave_nodes,
@@ -1346,20 +1408,22 @@ def real_main():
         opts.zone = random.choice(conn.get_all_zones()).name
 
     if action == "launch":
-        if opts.slaves <= 0:
-            print("ERROR: You have to start at least 1 slave", file=sys.stderr)
-            sys.exit(1)
-        if opts.resume:
-            (master_nodes, slave_nodes) = get_existing_cluster(conn, opts, cluster_name)
-        else:
-            (master_nodes, slave_nodes) = launch_cluster(conn, opts, cluster_name)
-        wait_for_cluster_state(
-            conn=conn,
-            opts=opts,
-            cluster_instances=(master_nodes + slave_nodes),
-            cluster_state='ssh-ready'
-        )
-        setup_cluster(conn, master_nodes, slave_nodes, opts, True)
+	if opts.slaves <= 0:
+	    print("ERROR: You have to start at least 1 slave", file=sys.stderr)
+	    sys.exit(1)
+	if opts.resume:
+	    (master_nodes, slave_nodes) = get_existing_cluster(conn, opts, cluster_name)
+	else:
+	    (master_nodes, slave_nodes) = launch_cluster(conn, opts, cluster_name)
+	wait_for_cluster_state(
+	    conn=conn,
+	    opts=opts,
+	    cluster_instances=(master_nodes + slave_nodes),
+	    cluster_state='ssh-ready'
+	)
+
+	if not opts.no_setup:
+	  setup_cluster(conn, master_nodes, slave_nodes, opts, True)
 
     elif action == "destroy":
         (master_nodes, slave_nodes) = get_existing_cluster(
@@ -1488,6 +1552,9 @@ def real_main():
                         inst.stop()
 
     elif action == "start":
+	if opts.elastic_ip:
+	    pdn = "ec2-" + "-".join(opts.elastic_ip.split(".")) + "." + opts.region + ".compute.amazonaws.com"
+	    print("will set master's public dns name to %s" % pdn)
         (master_nodes, slave_nodes) = get_existing_cluster(conn, opts, cluster_name)
         print("Starting slaves...")
         for inst in slave_nodes:
@@ -1503,6 +1570,29 @@ def real_main():
             cluster_instances=(master_nodes + slave_nodes),
             cluster_state='ssh-ready'
         )
+        if opts.elastic_ip:
+            print("setting master's public dns name to %s" % pdn)
+            conn.associate_address(master_nodes[0].id, opts.elastic_ip)
+            master_nodes[0].ip_address=opts.elastic_ip
+            master_nodes[0].public_dns_name=pdn
+            
+            print("Attachment made")
+            
+        # Determine types of running instances
+        existing_master_type = master_nodes[0].instance_type
+        existing_slave_type = slave_nodes[0].instance_type
+        # Setting opts.master_instance_type to the empty string indicates we
+        # have the same instance type for the master and the slaves
+        if existing_master_type == existing_slave_type:
+            existing_master_type = ""
+        opts.master_instance_type = existing_master_type
+        opts.instance_type = existing_slave_type
+
+	if not opts.no_setup:
+	  setup_cluster(conn, master_nodes, slave_nodes, opts, False)
+
+    elif action == "setup":
+        (master_nodes, slave_nodes) = get_existing_cluster(conn, opts, cluster_name)
 
         # Determine types of running instances
         existing_master_type = master_nodes[0].instance_type
@@ -1514,7 +1604,7 @@ def real_main():
         opts.master_instance_type = existing_master_type
         opts.instance_type = existing_slave_type
 
-        setup_cluster(conn, master_nodes, slave_nodes, opts, False)
+	setup_cluster(conn, master_nodes, slave_nodes, opts, True)
 
     else:
         print("Invalid action: %s" % action, file=stderr)
diff --git a/tachyon/init.sh b/tachyon/init.sh
index d5f1e481..ab71566d 100755
--- a/tachyon/init.sh
+++ b/tachyon/init.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 
-pushd /root > /dev/null
+pushd ~ > /dev/null
 
 if [ -d "tachyon" ]; then
   echo "Tachyon seems to be installed. Exiting."
diff --git a/tachyon/setup.sh b/tachyon/setup.sh
index 8d946abc..88708cbb 100755
--- a/tachyon/setup.sh
+++ b/tachyon/setup.sh
@@ -1,9 +1,9 @@
 #!/bin/bash
 
-/root/spark-ec2/copy-dir /root/tachyon
+~/spark-ec2/copy-dir ~/tachyon
 
-/root/tachyon/bin/tachyon format
+~/tachyon/bin/tachyon format
 
 sleep 1
 
-/root/tachyon/bin/tachyon-start.sh all Mount
+~/tachyon/bin/tachyon-start.sh all Mount
diff --git a/templates/root/ephemeral-hdfs/conf/hadoop-env.sh b/templates/root/ephemeral-hdfs/conf/hadoop-env.sh
index f4e5d7e9..30ddbac2 100755
--- a/templates/root/ephemeral-hdfs/conf/hadoop-env.sh
+++ b/templates/root/ephemeral-hdfs/conf/hadoop-env.sh
@@ -11,8 +11,8 @@ export JAVA_HOME={{java_home}}
 # Extra Java CLASSPATH elements.  Optional.
 # export HADOOP_CLASSPATH=
 
-export HADOOP_HOME="/root/ephemeral-hdfs"
-export HADOOP_MAPREDUCE_HOME="/root/mapreduce"
+export HADOOP_HOME="~/ephemeral-hdfs"
+export HADOOP_MAPREDUCE_HOME="~/mapreduce"
 
 # The maximum amount of heap to use, in MB. Default is 1000.
 export HADOOP_HEAPSIZE=1000
@@ -60,10 +60,14 @@ export HADOOP_PID_DIR=/var/hadoop/ephemeral-hdfs/pids
 # export HADOOP_NICENESS=10
 
 # Set hadoop user for CDH (which doesn't allow running as root)
-export HADOOP_NAMENODE_USER=hadoop
-export HADOOP_DATANODE_USER=hadoop
-export HADOOP_SECONDARYNAMENODE_USER=hadoop
-export HADOOP_JOBTRACKER_USER=hadoop
-export HADOOP_TASKTRACKER_USER=hadoop
+USER=`whoami`
+if [[ $USER == "root" ]]; then
+  USER=hadoop
+fi
+export HADOOP_NAMENODE_USER=$USER
+export HADOOP_DATANODE_USER=$USER
+export HADOOP_SECONDARYNAMENODE_USER=$USER
+export HADOOP_JOBTRACKER_USER=$USER
+export HADOOP_TASKTRACKER_USER=$USER
 
 ulimit -n 16000
diff --git a/templates/root/ephemeral-hdfs/conf/yarn-env.sh b/templates/root/ephemeral-hdfs/conf/yarn-env.sh
old mode 100644
new mode 100755
index 77e62194..cac6bc9c
--- a/templates/root/ephemeral-hdfs/conf/yarn-env.sh
+++ b/templates/root/ephemeral-hdfs/conf/yarn-env.sh
@@ -18,7 +18,7 @@ export HADOOP_YARN_USER=${HADOOP_YARN_USER:-yarn}
 
 # resolve links - $0 may be a softlink
 #export YARN_CONF_DIR="${YARN_CONF_DIR:-$HADOOP_YARN_HOME/conf}"
-export YARN_CONF_DIR="/root/ephemeral-hdfs/conf"
+export YARN_CONF_DIR="~/ephemeral-hdfs/conf"
 
 # some Java parameters
 # export JAVA_HOME=/home/y/libexec/jdk1.6.0/
diff --git a/templates/root/spark/conf/spark-defaults.conf b/templates/root/spark/conf/spark-defaults.conf
index c63994a5..2387bfd1 100644
--- a/templates/root/spark/conf/spark-defaults.conf
+++ b/templates/root/spark/conf/spark-defaults.conf
@@ -1,6 +1,6 @@
 spark.executor.memory	{{spark_worker_mem}}
-spark.executor.extraLibraryPath	/root/ephemeral-hdfs/lib/native/
-spark.executor.extraClassPath	/root/ephemeral-hdfs/conf
+spark.executor.extraLibraryPath	~/ephemeral-hdfs/lib/native/
+spark.executor.extraClassPath	~/ephemeral-hdfs/conf
 
 # for spark version < 1.4.0
 spark.tachyonStore.url tachyon://{{active_master}}:19998
diff --git a/templates/root/spark/conf/spark-env.sh b/templates/root/spark/conf/spark-env.sh
index aa4490b8..dc40d612 100755
--- a/templates/root/spark/conf/spark-env.sh
+++ b/templates/root/spark/conf/spark-env.sh
@@ -9,12 +9,12 @@ if [ -n "{{spark_worker_instances}}" ]; then
 fi
 export SPARK_WORKER_CORES={{spark_worker_cores}}
 
-export HADOOP_HOME="/root/ephemeral-hdfs"
+export HADOOP_HOME="~/ephemeral-hdfs"
 export SPARK_MASTER_IP={{active_master}}
-export MASTER=`cat /root/spark-ec2/cluster-url`
+export MASTER=`cat ~/spark-ec2/cluster-url`
 
-export SPARK_SUBMIT_LIBRARY_PATH="$SPARK_SUBMIT_LIBRARY_PATH:/root/ephemeral-hdfs/lib/native/"
-export SPARK_SUBMIT_CLASSPATH="$SPARK_CLASSPATH:$SPARK_SUBMIT_CLASSPATH:/root/ephemeral-hdfs/conf"
+export SPARK_SUBMIT_LIBRARY_PATH="$SPARK_SUBMIT_LIBRARY_PATH:~/ephemeral-hdfs/lib/native/"
+export SPARK_SUBMIT_CLASSPATH="$SPARK_CLASSPATH:$SPARK_SUBMIT_CLASSPATH:~/ephemeral-hdfs/conf"
 
 # Bind Spark's web UIs to this machine's public EC2 hostname otherwise fallback to private IP:
 export SPARK_PUBLIC_DNS=`
@@ -22,10 +22,10 @@ wget -q -O - http://169.254.169.254/latest/meta-data/public-hostname ||\
 wget -q -O - http://169.254.169.254/latest/meta-data/local-ipv4`
 
 # Used for YARN model
-export YARN_CONF_DIR="/root/ephemeral-hdfs/conf"
+export YARN_CONF_DIR="~/ephemeral-hdfs/conf"
 
 # Set a high ulimit for large shuffles, only root can do this
 if [ $(id -u) == "0" ]
 then
-    ulimit -n 1000000
+    sudo ulimit -n 1000000
 fi
diff --git a/templates/root/tachyon/conf/tachyon-env.sh b/templates/root/tachyon/conf/tachyon-env.sh
old mode 100644
new mode 100755