diff --git a/beam/beam.sh b/beam/beam.sh
index 2cad639d8..2ce1640bf 100755
--- a/beam/beam.sh
+++ b/beam/beam.sh
@@ -18,14 +18,13 @@ readonly BEAM_IMAGE_VERSION_DEFAULT="master"
 readonly BEAM_IMAGE_REPOSITORY_KEY="beam-image-repository"
 readonly BEAM_IMAGE_REPOSITORY_DEFAULT="apache.bintray.io/beam"
 
-
 readonly START_FLINK_YARN_SESSION_METADATA_KEY='flink-start-yarn-session'
 # Set this to true to start a flink yarn session at initialization time.
 readonly START_FLINK_YARN_SESSION_DEFAULT=true
 
 function is_master() {
   local role="$(/usr/share/google/get_metadata_value attributes/dataproc-role)"
-  if [[ "$role" == 'Master' ]] ; then
+  if [[ "$role" == 'Master' ]]; then
     true
   else
     false
@@ -33,8 +32,8 @@ function is_master() {
 }
 
 function get_artifacts_dir() {
-  /usr/share/google/get_metadata_value "attributes/${ARTIFACTS_GCS_PATH_METADATA_KEY}" \
-    || echo "gs://$(/usr/share/google/get_metadata_value "attributes/dataproc-bucket")/beam-artifacts"
+  /usr/share/google/get_metadata_value "attributes/${ARTIFACTS_GCS_PATH_METADATA_KEY}" ||
+    echo "gs://$(/usr/share/google/get_metadata_value "attributes/dataproc-bucket")/beam-artifacts"
 }
 
 function download_snapshot() {
@@ -49,17 +48,17 @@ function download_snapshot() {
 
 function flink_master_url() {
   local start_flink_yarn_session="$(/usr/share/google/get_metadata_value \
-    "attributes/${START_FLINK_YARN_SESSION_METADATA_KEY}" \
-    || echo "${START_FLINK_YARN_SESSION_DEFAULT}")"
-  # TODO: delete this workaround when the beam job service is able to understand 
+    "attributes/${START_FLINK_YARN_SESSION_METADATA_KEY}" ||
+    echo "${START_FLINK_YARN_SESSION_DEFAULT}")"
+  # TODO: delete this workaround when the beam job service is able to understand
   # flink in yarn mode.
-  if ${start_flink_yarn_session} ; then
+  if ${start_flink_yarn_session}; then
     # grab final field from the first yarn application that contains 'flink'
-    yarn application -list \
-      | grep -i 'flink' \
-      | head -n1 \
-      | awk -F $'\t' '{print $9}' \
-      | cut -c8-
+    yarn application -list |
+      grep -i 'flink' |
+      head -n1 |
+      awk -F $'\t' '{print $9}' |
+      cut -c8-
   else
     echo "localhost:8081"
   fi
@@ -69,8 +68,8 @@ function install_job_service() {
   local master_url="$(/usr/share/google/get_metadata_value attributes/dataproc-master)"
   local artifacts_dir="$(get_artifacts_dir)"
   local release_snapshot_url="$(/usr/share/google/get_metadata_value \
-    "attributes/${RELEASE_SNAPSHOT_URL_METADATA_KEY}" \
-    || echo "${RELEASE_SNAPSHOT_URL_DEFAULT}")"
+    "attributes/${RELEASE_SNAPSHOT_URL_METADATA_KEY}" ||
+    echo "${RELEASE_SNAPSHOT_URL_DEFAULT}")"
 
   echo "Retrieving Beam Job Service snapshot from ${release_snapshot_url}"
 
@@ -84,7 +83,7 @@ function install_job_service() {
   mkdir -p "${SERVICE_WORKING_DIR}"
   chown -R "${SERVICE_WORKING_USER}" "${SERVICE_WORKING_DIR}"
 
-  cat > "/etc/systemd/system/beam-job-service.service" <<EOF
+  cat >"/etc/systemd/system/beam-job-service.service" <<EOF
 [Unit]
 Description=Beam Job Service
 After=default.target
@@ -103,7 +102,7 @@ Restart=on-failure
 [Install]
 WantedBy=multi-user.target
 EOF
-systemctl enable beam-job-service
+  systemctl enable beam-job-service
 }
 
 function run_job_service() {
@@ -112,11 +111,11 @@ function run_job_service() {
 
 function pull_beam_images() {
   local beam_image_version="$(/usr/share/google/get_metadata_value \
-    "attributes/${BEAM_IMAGE_VERSION_METADATA_KEY}" \
-    || echo "${BEAM_IMAGE_VERSION_DEFAULT}")"
+    "attributes/${BEAM_IMAGE_VERSION_METADATA_KEY}" ||
+    echo "${BEAM_IMAGE_VERSION_DEFAULT}")"
   local image_repo="$(/usr/share/google/get_metadata_value \
-    "attributes/${BEAM_IMAGE_REPOSITORY_KEY}" \
-    || echo "${BEAM_IMAGE_REPOSITORY_DEFAULT}")"
+    "attributes/${BEAM_IMAGE_REPOSITORY_KEY}" ||
+    echo "${BEAM_IMAGE_REPOSITORY_DEFAULT}")"
   # Pull beam images with `sudo -i` since if pulling from GCR, yarn will be
   # configured with GCR authorization
   sudo -u yarn -i docker pull "${image_repo}/go:${beam_image_version}"
@@ -131,9 +130,9 @@ function main() {
   fi
 
   local pull_images="$(/usr/share/google/get_metadata_value \
-    "attributes/${BEAM_IMAGE_ENABLE_PULL_METADATA_KEY}" \
-    || echo "${BEAM_IMAGE_ENABLE_PULL_DEFAULT}")"
-  if ${pull_images} ; then
+    "attributes/${BEAM_IMAGE_ENABLE_PULL_METADATA_KEY}" ||
+    echo "${BEAM_IMAGE_ENABLE_PULL_DEFAULT}")"
+  if ${pull_images}; then
     pull_beam_images
   fi
 }
diff --git a/beam/util/build-beam-artifacts.sh b/beam/util/build-beam-artifacts.sh
index c122f66a8..ed83f6ed9 100755
--- a/beam/util/build-beam-artifacts.sh
+++ b/beam/util/build-beam-artifacts.sh
@@ -2,7 +2,7 @@
 
 set -exuo pipefail
 
-if [ "$#" -lt 2 ] ; then
+if [ "$#" -lt 2 ]; then
   echo "Usage: $0 <BEAM_JOB_SERVICE_DESTINATION> <BEAM_CONTAINER_IMAGE_DESTINATION> [<BEAM_SOURCE_VERSION> [<BEAM_SOURCE_DIRECTORY>]]" >&2
   exit 1
 fi
@@ -20,10 +20,10 @@ function build_job_service() {
 
 function build_container() {
   ./gradlew docker
-  local images=($(docker images \
-    | grep '.*-docker-apache' \
-    | awk '{print $1}'))
-  for image in ${images} ; do
+  local images=($(docker images |
+    grep '.*-docker-apache' |
+    awk '{print $1}'))
+  for image in ${images}; do
     local image_destination="${BEAM_CONTAINER_IMAGE_DESTINATION}/$(basename ${image}):${BEAM_SOURCE_VERSION}"
     docker tag $image:latest ${image_destination}
     docker push ${image_destination}
@@ -31,7 +31,7 @@ function build_container() {
 }
 
 function main() {
-  if [[ $# -eq 4 ]] ; then
+  if [[ $# -eq 4 ]]; then
     # if there is a 4th argument, use it as the beam source directory
     pushd "$4"
   else
diff --git a/bigdl/bigdl.sh b/bigdl/bigdl.sh
index 30fd485de..e600aa314 100755
--- a/bigdl/bigdl.sh
+++ b/bigdl/bigdl.sh
@@ -17,10 +17,10 @@ cd /opt/intel-bigdl
 wget -nv --timeout=30 --tries=5 --retry-connrefused "${BIGDL_DOWNLOAD_URL}"
 unzip *.zip
 
-JAR=`realpath lib/*.jar`
-PYTHON_ZIP=`realpath lib/*.zip`
+JAR=$(realpath lib/*.jar)
+PYTHON_ZIP=$(realpath lib/*.zip)
 
-cat << EOF >> /etc/spark/conf/spark-env.sh
+cat <<EOF >>/etc/spark/conf/spark-env.sh
 SPARK_DIST_CLASSPATH="\$SPARK_DIST_CLASSPATH:$JAR"
 PYTHONPATH="\$PYTHONPATH:$PYTHON_ZIP"
 EOF
@@ -30,20 +30,20 @@ EOF
 
 if [[ "${ROLE}" == "Master" ]]; then
   NUM_NODEMANAGERS_TARGET="${WORKER_COUNT}"
-  if (( "${WORKER_COUNT}" == 0 )); then
+  if (("${WORKER_COUNT}" == 0)); then
     # Single node clusters have one node manager
     NUM_NODEMANAGERS_TARGET=1
   fi
   # Wait for 5 minutes for Node Managers to register and run.
   # Break early if the expected number of node managers have registered.
-  for (( i=0; i < 5*60; i++ )); do
+  for ((i = 0; i < 5 * 60; i++)); do
     CURRENTLY_RUNNING_NODEMANAGERS=$(yarn node -list | grep RUNNING | wc -l)
-    if (( CURRENTLY_RUNNING_NODEMANAGERS == NUM_NODEMANAGERS_TARGET )); then
+    if ((CURRENTLY_RUNNING_NODEMANAGERS == NUM_NODEMANAGERS_TARGET)); then
       break
     fi
     sleep 1
   done
-  if (( CURRENTLY_RUNNING_NODEMANAGERS == 0 )); then
+  if ((CURRENTLY_RUNNING_NODEMANAGERS == 0)); then
     echo "No node managers running. Cluster creation likely failed"
     exit 1
   fi
@@ -73,14 +73,14 @@ if [[ "${ROLE}" == "Master" ]]; then
 
   # Check if it BigDL conf or Zoo
   if [ -f conf/spark-bigdl.conf ]; then
-    cat conf/spark-bigdl.conf >> /etc/spark/conf/spark-defaults.conf
+    cat conf/spark-bigdl.conf >>/etc/spark/conf/spark-defaults.conf
   elif [ -f conf/spark-analytics-zoo.conf ]; then
-    cat conf/spark-analytics-zoo.conf >> /etc/spark/conf/spark-defaults.conf
+    cat conf/spark-analytics-zoo.conf >>/etc/spark/conf/spark-defaults.conf
   else
     err "Can't find any suitable spark config for Intel BigDL/Zoo"
   fi
 
-  cat << EOF >> /etc/spark/conf/spark-defaults.conf
+  cat <<EOF >>/etc/spark/conf/spark-defaults.conf
 
 spark.dynamicAllocation.enabled=false
 spark.executor.instances=${SPARK_NUM_EXECUTORS}
diff --git a/cloud-sql-proxy/cloud-sql-proxy.sh b/cloud-sql-proxy/cloud-sql-proxy.sh
index 787de9c38..e38605948 100755
--- a/cloud-sql-proxy/cloud-sql-proxy.sh
+++ b/cloud-sql-proxy/cloud-sql-proxy.sh
@@ -42,36 +42,36 @@ readonly kms_key_uri="$(/usr/share/google/get_metadata_value attributes/kms-key-
 readonly db_admin_password_uri="$(/usr/share/google/get_metadata_value attributes/db-admin-password-uri)"
 if [[ -n "${db_admin_password_uri}" ]]; then
   # Decrypt password
-  readonly db_admin_password="$(gsutil cat $db_admin_password_uri | \
+  readonly db_admin_password="$(gsutil cat $db_admin_password_uri |
     gcloud kms decrypt \
-    --ciphertext-file - \
-    --plaintext-file - \
-    --key $kms_key_uri)"
+      --ciphertext-file - \
+      --plaintext-file - \
+      --key $kms_key_uri)"
 else
   readonly db_admin_password=''
 fi
 if [ "${db_admin_password}" == "" ]; then
-    readonly db_admin_password_parameter=""
+  readonly db_admin_password_parameter=""
 else
-    readonly db_admin_password_parameter="-p${db_admin_password}"
+  readonly db_admin_password_parameter="-p${db_admin_password}"
 fi
 
 # Database password to use to access metastore.
 readonly db_hive_password_uri="$(/usr/share/google/get_metadata_value attributes/db-hive-password-uri)"
 if [[ -n "${db_hive_password_uri}" ]]; then
   # Decrypt password
-  readonly db_hive_password="$(gsutil cat $db_hive_password_uri | \
+  readonly db_hive_password="$(gsutil cat $db_hive_password_uri |
     gcloud kms decrypt \
-    --ciphertext-file - \
-    --plaintext-file - \
-    --key $kms_key_uri)"
+      --ciphertext-file - \
+      --plaintext-file - \
+      --key $kms_key_uri)"
 else
   readonly db_hive_password='hive-password'
 fi
 if [ "${db_hive_password}" == "" ]; then
-    readonly db_hive_password_parameter=""
+  readonly db_hive_password_parameter=""
 else
-    readonly db_hive_password_parameter="-p${db_hive_password}"
+  readonly db_hive_password_parameter="-p${db_hive_password}"
 fi
 
 readonly PROXY_DIR='/var/run/cloud_sql_proxy'
@@ -85,8 +85,8 @@ readonly DATAPROC_MASTER=$(/usr/share/google/get_metadata_value attributes/datap
 function get_java_property() {
   local property_file=$1
   local property_name=$2
-  local property_value=$(grep "^${property_name}=" "${property_file}" | \
-      tail -n 1 | cut -d '=' -f 2- | sed -r 's/\\([#!=:])/\1/g')
+  local property_value=$(grep "^${property_name}=" "${property_file}" |
+    tail -n 1 | cut -d '=' -f 2- | sed -r 's/\\([#!=:])/\1/g')
   echo "${property_value}"
 }
 
@@ -102,7 +102,7 @@ function is_component_selected() {
   local component=$1
 
   local activated_components=$(get_dataproc_property \
-      dataproc.components.activate)
+    dataproc.components.activate)
 
   if [[ ${activated_components} == *${component}* ]]; then
     return 0
@@ -123,7 +123,7 @@ function get_hive_principal() {
 
 function get_hiveserver_uri() {
   local base_connect_string="jdbc:hive2://localhost:10000"
-  if [[ "${KERBEROS_ENABLED}" == 'true' ]] ; then
+  if [[ "${KERBEROS_ENABLED}" == 'true' ]]; then
     local hive_principal=$(get_hive_principal)
     echo "${base_connect_string}/;principal=${hive_principal}"
   else
@@ -151,12 +151,12 @@ function run_with_retries() {
       break
     else
       local sleep_time=${retry_backoff[$i]}
-      echo "'${cmd[*]}' attempt $(( $i + 1 )) failed! Sleeping ${sleep_time}." >&2
+      echo "'${cmd[*]}' attempt $(($i + 1)) failed! Sleeping ${sleep_time}." >&2
       sleep ${sleep_time}
     fi
   done
 
-  if ! (( ${update_succeeded} )); then
+  if ! ((${update_succeeded})); then
     echo "Final attempt of '${cmd[*]}'..."
     # Let any final error propagate all the way out to any error traps.
     "${cmd[@]}"
@@ -165,10 +165,10 @@ function run_with_retries() {
 
 function configure_proxy_flags() {
   # If a cloud sql instance has both public and private IP, use private IP.
-  if [[ $use_cloud_sql_private_ip = "true" ]]; then
+  if [[ $use_cloud_sql_private_ip == "true" ]]; then
     proxy_instances_flags+=" --ip_address_types=PRIVATE"
   fi
-  if [[ $enable_cloud_sql_metastore = "true" ]]; then
+  if [[ $enable_cloud_sql_metastore == "true" ]]; then
     if [[ -z "${metastore_instance}" ]]; then
       err 'Must specify hive-metastore-instance VM metadata'
     elif ! [[ "${metastore_instance}" =~ .+:.+:.+ ]]; then
@@ -197,7 +197,7 @@ function install_cloud_sql_proxy() {
   mkdir -p ${PROXY_DIR}
 
   # Install proxy as systemd service for reboot tolerance.
-  cat << EOF > ${INIT_SCRIPT}
+  cat <<EOF >${INIT_SCRIPT}
 [Unit]
 Description=Google Cloud SQL Proxy
 After=local-fs.target network-online.target
@@ -215,18 +215,18 @@ WantedBy=multi-user.target
 EOF
   chmod a+rw ${INIT_SCRIPT}
   systemctl enable cloud-sql-proxy
-  systemctl start cloud-sql-proxy \
-    || err 'Unable to start cloud-sql-proxy service'
-    
-  if [[ $enable_cloud_sql_metastore = "true" ]]; then
+  systemctl start cloud-sql-proxy ||
+    err 'Unable to start cloud-sql-proxy service'
+
+  if [[ $enable_cloud_sql_metastore == "true" ]]; then
     run_with_retries nc -zv localhost ${metastore_proxy_port}
   fi
 
   echo 'Cloud SQL Proxy installation succeeded' >&2
 
-  if [[ $enable_cloud_sql_metastore = "true" ]]; then
+  if [[ $enable_cloud_sql_metastore == "true" ]]; then
     # Update hive-site.xml
-    cat << EOF > hive-template.xml
+    cat <<EOF >hive-template.xml
 <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
 <configuration>
   <property>
@@ -245,17 +245,16 @@ EOF
 </configuration>
 EOF
 
-  bdconfig merge_configurations \
-    --configuration_file /etc/hive/conf/hive-site.xml \
-    --source_configuration_file hive-template.xml \
-    --clobber
-fi
+    bdconfig merge_configurations \
+      --configuration_file /etc/hive/conf/hive-site.xml \
+      --source_configuration_file hive-template.xml \
+      --clobber
+  fi
 }
 
-
-function configure_sql_client(){
+function configure_sql_client() {
   # Configure mysql client to talk to metastore
-  cat << EOF > /etc/mysql/conf.d/cloud-sql-proxy.cnf
+  cat <<EOF >/etc/mysql/conf.d/cloud-sql-proxy.cnf
 [client]
 protocol = tcp
 port = ${metastore_proxy_port}
@@ -280,29 +279,29 @@ EOF
     mysql -u "${db_admin_user}" "${db_admin_password_parameter}" -e \
       "CREATE DATABASE ${metastore_db}; \
        GRANT ALL PRIVILEGES ON ${metastore_db}.* TO '${db_hive_user}';"
-    /usr/lib/hive/bin/schematool -dbType mysql -initSchema \
-      || err 'Failed to set mysql schema.'
+    /usr/lib/hive/bin/schematool -dbType mysql -initSchema ||
+      err 'Failed to set mysql schema.'
   fi
 
   run_with_retries run_validation
 }
 
 function run_validation() {
-  if ( systemctl is-enabled --quiet hive-metastore ); then
+  if (systemctl is-enabled --quiet hive-metastore); then
     # Start metastore back up.
-    systemctl restart hive-metastore \
-      || err 'Unable to start hive-metastore service'
+    systemctl restart hive-metastore ||
+      err 'Unable to start hive-metastore service'
   else
     echo "Service hive-metastore is not loaded"
   fi
 
   # Check that metastore schema is compatible.
-  /usr/lib/hive/bin/schematool -dbType mysql -info || \
-      err 'Run /usr/lib/hive/bin/schematool -dbType mysql -upgradeSchemaFrom <schema-version> to upgrade the schema. Note that this may break Hive metastores that depend on the old schema'
+  /usr/lib/hive/bin/schematool -dbType mysql -info ||
+    err 'Run /usr/lib/hive/bin/schematool -dbType mysql -upgradeSchemaFrom <schema-version> to upgrade the schema. Note that this may break Hive metastores that depend on the old schema'
 
   # Validate it's functioning.
   local hiveserver_uri=$(get_hiveserver_uri)
-  if ! timeout 60s beeline -u ${hiveserver_uri} -e 'SHOW TABLES;' >& /dev/null; then
+  if ! timeout 60s beeline -u ${hiveserver_uri} -e 'SHOW TABLES;' >&/dev/null; then
     err 'Failed to bring up Cloud SQL Metastore'
   else
     echo 'Cloud SQL Hive Metastore initialization succeeded' >&2
@@ -310,15 +309,14 @@ function run_validation() {
 
 }
 
-
-function configure_hive_warehouse_dir(){
+function configure_hive_warehouse_dir() {
   # Wait for master 0 to create the metastore db if necessary.
   run_with_retries run_validation
 
   local hiveserver_uri=$(get_hiveserver_uri)
   HIVE_WAREHOURSE_URI=$(beeline -u ${hiveserver_uri} \
-    -e "describe database default;" \
-    | sed '4q;d' | cut -d "|" -f4 | tr -d '[:space:]')
+    -e "describe database default;" |
+    sed '4q;d' | cut -d "|" -f4 | tr -d '[:space:]')
 
   echo "Hive warehouse uri: $HIVE_WAREHOURSE_URI"
 
@@ -368,14 +366,14 @@ function main() {
 
   if [[ "${role}" == 'Master' ]]; then
     # Disable Hive Metastore and MySql Server.
-    if [[ $enable_cloud_sql_metastore = "true" ]]; then
-      if ( systemctl is-enabled --quiet hive-metastore ); then
+    if [[ $enable_cloud_sql_metastore == "true" ]]; then
+      if (systemctl is-enabled --quiet hive-metastore); then
         # Stop hive-metastore if it is enabled
         systemctl stop hive-metastore
       else
         echo "Service hive-metastore is not enabled"
       fi
-      if ( systemctl is-enabled --quiet mysql ); then
+      if (systemctl is-enabled --quiet mysql); then
         systemctl stop mysql
         systemctl disable mysql
       else
@@ -383,7 +381,7 @@ function main() {
       fi
     fi
     install_cloud_sql_proxy
-    if [[ $enable_cloud_sql_metastore = "true" ]]; then
+    if [[ $enable_cloud_sql_metastore == "true" ]]; then
       if [[ "${HOSTNAME}" == "${DATAPROC_MASTER}" ]]; then
         # Initialize metastore db instance and set hive.metastore.warehouse.dir
         # on master 0.
@@ -396,7 +394,7 @@ function main() {
   else
     # This part runs on workers.
     # Run installation on workers when enable_proxy_on_workers is set.
-    if [[ $enable_proxy_on_workers = "true" ]]; then
+    if [[ $enable_proxy_on_workers == "true" ]]; then
       install_cloud_sql_proxy
     fi
   fi
diff --git a/conda/bootstrap-conda.sh b/conda/bootstrap-conda.sh
index 63391ae69..06f501b1b 100755
--- a/conda/bootstrap-conda.sh
+++ b/conda/bootstrap-conda.sh
@@ -17,105 +17,105 @@ NEW_MINICONDA_VERSION="4.7.10"
 MIN_SPARK_VERSION_FOR_NEWER_MINICONDA="2.2.0"
 
 if [[ -f /etc/profile.d/effective-python.sh ]]; then
-    PROFILE_SCRIPT_PATH=/etc/profile.d/effective-python.sh
+  PROFILE_SCRIPT_PATH=/etc/profile.d/effective-python.sh
 else
-    PROFILE_SCRIPT_PATH=/etc/profile.d/conda.sh
+  PROFILE_SCRIPT_PATH=/etc/profile.d/conda.sh
 fi
 
 if [[ ! -v CONDA_INSTALL_PATH ]]; then
-    echo "CONDA_INSTALL_PATH not set, setting ..."
-    CONDA_INSTALL_PATH="/opt/conda"
-    echo "Set CONDA_INSTALL_PATH to $CONDA_INSTALL_PATH"
+  echo "CONDA_INSTALL_PATH not set, setting ..."
+  CONDA_INSTALL_PATH="/opt/conda"
+  echo "Set CONDA_INSTALL_PATH to $CONDA_INSTALL_PATH"
 fi
 
 # Check if Conda is already installed at the expected location. This will allow
 # this init action to override the default Miniconda in 1.4+ or Anaconda
 # optional component in 1.3+ which is installed at /opt/conda/default.
 if [[ -f "${CONDA_INSTALL_PATH}/bin/conda" ]]; then
-    echo "Dataproc has installed Conda previously at ${CONDA_INSTALL_PATH}. Skipping install!"
-    exit 0
+  echo "Dataproc has installed Conda previously at ${CONDA_INSTALL_PATH}. Skipping install!"
+  exit 0
 else
-    # 0. Specify Miniconda version
-    ## 0.1 A few parameters
-    ## specify base operating system
-    if [[ ! -v OS_TYPE ]]; then
-        echo "OS_TYPE not set, setting  ..."
-        OS_TYPE="Linux-x86_64.sh"
-        echo "Set OS_TYPE to $OS_TYPE"
-    fi
-    ## Python 2 or 3 based miniconda?
-    if [[ -z "${MINICONDA_VARIANT}" ]]; then
-        echo "MINICONDA_VARIANT not set, setting ... "
-        MINICONDA_VARIANT="3"  #for Python 3.5.x
-        echo "Set MINICONDA_VARIANT to $MINICONDA_VARIANT"
-    fi
-    ## specify Miniconda release (e.g., MINICONDA_VERSION='4.0.5')
-    if [[ -z "${MINICONDA_VERSION}" ]]; then
-      # Pin to 4.2.12 by default until Spark default is 2.2.0, then use newer
-      # one (https://issues.apache.org/jira/browse/SPARK-19019)
-      SPARK_VERSION=`spark-submit --version 2>&1  | sed -n 's/.*version[[:blank:]]\+\([0-9]\+\.[0-9]\.[0-9]\+\+\).*/\1/p' | head -n1`
-      if dpkg --compare-versions ${SPARK_VERSION} ge ${MIN_SPARK_VERSION_FOR_NEWER_MINICONDA}; then
-        MINICONDA_VERSION="${NEW_MINICONDA_VERSION}"
-      else
-        MINICONDA_VERSION="${OLD_MINICONDA_VERSION}"
-      fi
-    fi
-
-    ## 0.2 Compute Miniconda version
-    MINICONDA_FULL_NAME="Miniconda$MINICONDA_VARIANT-$MINICONDA_VERSION-$OS_TYPE"
-    echo "Complete Miniconda version resolved to: $MINICONDA_FULL_NAME"
-    ## 0.3 Set MD5 hash for check (if desired)
-    #expectedHash="b1b15a3436bb7de1da3ccc6e08c7a5df"
-
-    # 1. Setup Miniconda Install
-    ## 1.1 Define Miniconda install directory
-    echo "Working directory: $PWD"
-    if [[ ! -v $PROJ_DIR ]]; then
-        echo "No path argument specified, setting install directory as working directory: $PWD."
-        PROJ_DIR=$PWD
-    fi
-
-    ## 1.2 Setup Miniconda
-    cd $PROJ_DIR
-    MINICONDA_SCRIPT_PATH="$PROJ_DIR/$MINICONDA_FULL_NAME"
-    echo "Defined Miniconda script path: $MINICONDA_SCRIPT_PATH"
-
-    if [[ -f "$MINICONDA_SCRIPT_PATH" ]]; then
-      echo "Found existing Miniconda script at: $MINICONDA_SCRIPT_PATH"
+  # 0. Specify Miniconda version
+  ## 0.1 A few parameters
+  ## specify base operating system
+  if [[ ! -v OS_TYPE ]]; then
+    echo "OS_TYPE not set, setting  ..."
+    OS_TYPE="Linux-x86_64.sh"
+    echo "Set OS_TYPE to $OS_TYPE"
+  fi
+  ## Python 2 or 3 based miniconda?
+  if [[ -z "${MINICONDA_VARIANT}" ]]; then
+    echo "MINICONDA_VARIANT not set, setting ... "
+    MINICONDA_VARIANT="3" #for Python 3.5.x
+    echo "Set MINICONDA_VARIANT to $MINICONDA_VARIANT"
+  fi
+  ## specify Miniconda release (e.g., MINICONDA_VERSION='4.0.5')
+  if [[ -z "${MINICONDA_VERSION}" ]]; then
+    # Pin to 4.2.12 by default until Spark default is 2.2.0, then use newer
+    # one (https://issues.apache.org/jira/browse/SPARK-19019)
+    SPARK_VERSION=$(spark-submit --version 2>&1 | sed -n 's/.*version[[:blank:]]\+\([0-9]\+\.[0-9]\.[0-9]\+\+\).*/\1/p' | head -n1)
+    if dpkg --compare-versions ${SPARK_VERSION} ge ${MIN_SPARK_VERSION_FOR_NEWER_MINICONDA}; then
+      MINICONDA_VERSION="${NEW_MINICONDA_VERSION}"
     else
-      echo "Downloading Miniconda script to: $MINICONDA_SCRIPT_PATH ..."
-      wget -nv --timeout=30 --tries=5 --retry-connrefused \
-        https://repo.continuum.io/miniconda/$MINICONDA_FULL_NAME -P "$PROJ_DIR"
-      echo "Downloaded $MINICONDA_FULL_NAME!"
-      ls -al $MINICONDA_SCRIPT_PATH
-      chmod 755 $MINICONDA_SCRIPT_PATH
-    fi
-
-    ## 1.3 #md5sum hash check of miniconda installer
-    if [[ -v expectedHash ]]; then
-        md5Output=$(md5sum $MINICONDA_SCRIPT_PATH | awk '{print $1}')
-        if [ "$expectedHash" != "$md5Output" ]; then
-            echo "Unexpected md5sum $md5Output for $MINICONDA_FULL_NAME"
-            exit 1
-        fi
+      MINICONDA_VERSION="${OLD_MINICONDA_VERSION}"
     fi
-
-    # 2. Install conda
-    ## 2.1 Via bootstrap
-    LOCAL_CONDA_PATH="$PROJ_DIR/miniconda"
-    if [[ ! -d $LOCAL_CONDA_PATH ]]; then
-        #blow away old symlink / default Miniconda install
-        rm -rf "$PROJ_DIR/miniconda"
-        # Install Miniconda
-        echo "Installing $MINICONDA_FULL_NAME to $CONDA_INSTALL_PATH..."
-        bash $MINICONDA_SCRIPT_PATH -b -p $CONDA_INSTALL_PATH -f
-        chmod 755 $CONDA_INSTALL_PATH
-        #create symlink
-        ln -sf $CONDA_INSTALL_PATH "$PROJ_DIR/miniconda"
-        chmod 755 "$PROJ_DIR/miniconda"
-    else
-        echo "Existing directory at path: $LOCAL_CONDA_PATH, skipping install!"
+  fi
+
+  ## 0.2 Compute Miniconda version
+  MINICONDA_FULL_NAME="Miniconda$MINICONDA_VARIANT-$MINICONDA_VERSION-$OS_TYPE"
+  echo "Complete Miniconda version resolved to: $MINICONDA_FULL_NAME"
+  ## 0.3 Set MD5 hash for check (if desired)
+  #expectedHash="b1b15a3436bb7de1da3ccc6e08c7a5df"
+
+  # 1. Setup Miniconda Install
+  ## 1.1 Define Miniconda install directory
+  echo "Working directory: $PWD"
+  if [[ ! -v $PROJ_DIR ]]; then
+    echo "No path argument specified, setting install directory as working directory: $PWD."
+    PROJ_DIR=$PWD
+  fi
+
+  ## 1.2 Setup Miniconda
+  cd $PROJ_DIR
+  MINICONDA_SCRIPT_PATH="$PROJ_DIR/$MINICONDA_FULL_NAME"
+  echo "Defined Miniconda script path: $MINICONDA_SCRIPT_PATH"
+
+  if [[ -f "$MINICONDA_SCRIPT_PATH" ]]; then
+    echo "Found existing Miniconda script at: $MINICONDA_SCRIPT_PATH"
+  else
+    echo "Downloading Miniconda script to: $MINICONDA_SCRIPT_PATH ..."
+    wget -nv --timeout=30 --tries=5 --retry-connrefused \
+      https://repo.continuum.io/miniconda/$MINICONDA_FULL_NAME -P "$PROJ_DIR"
+    echo "Downloaded $MINICONDA_FULL_NAME!"
+    ls -al $MINICONDA_SCRIPT_PATH
+    chmod 755 $MINICONDA_SCRIPT_PATH
+  fi
+
+  ## 1.3 #md5sum hash check of miniconda installer
+  if [[ -v expectedHash ]]; then
+    md5Output=$(md5sum $MINICONDA_SCRIPT_PATH | awk '{print $1}')
+    if [ "$expectedHash" != "$md5Output" ]; then
+      echo "Unexpected md5sum $md5Output for $MINICONDA_FULL_NAME"
+      exit 1
     fi
+  fi
+
+  # 2. Install conda
+  ## 2.1 Via bootstrap
+  LOCAL_CONDA_PATH="$PROJ_DIR/miniconda"
+  if [[ ! -d $LOCAL_CONDA_PATH ]]; then
+    #blow away old symlink / default Miniconda install
+    rm -rf "$PROJ_DIR/miniconda"
+    # Install Miniconda
+    echo "Installing $MINICONDA_FULL_NAME to $CONDA_INSTALL_PATH..."
+    bash $MINICONDA_SCRIPT_PATH -b -p $CONDA_INSTALL_PATH -f
+    chmod 755 $CONDA_INSTALL_PATH
+    #create symlink
+    ln -sf $CONDA_INSTALL_PATH "$PROJ_DIR/miniconda"
+    chmod 755 "$PROJ_DIR/miniconda"
+  else
+    echo "Existing directory at path: $LOCAL_CONDA_PATH, skipping install!"
+  fi
 fi
 
 ## 2.2 Update PATH and conda...
@@ -138,8 +138,8 @@ conda info -a
 echo "Updating global profiles to export miniconda bin location to PATH..."
 echo "Adding path definition to profiles..."
 echo "# Environment varaibles set by Conda init action." | tee -a "${PROFILE_SCRIPT_PATH}" #/etc/*bashrc /etc/profile
-echo "export CONDA_BIN_PATH=$CONDA_BIN_PATH" | tee -a "${PROFILE_SCRIPT_PATH}" #/etc/*bashrc /etc/profile
-echo 'export PATH=$CONDA_BIN_PATH:$PATH' | tee -a "${PROFILE_SCRIPT_PATH}"  #/etc/*bashrc /etc/profile
+echo "export CONDA_BIN_PATH=$CONDA_BIN_PATH" | tee -a "${PROFILE_SCRIPT_PATH}"             #/etc/*bashrc /etc/profile
+echo 'export PATH=$CONDA_BIN_PATH:$PATH' | tee -a "${PROFILE_SCRIPT_PATH}"                 #/etc/*bashrc /etc/profile
 
 # 2.3 Update global profiles to add the miniconda location to PATH
 echo "Updating global profiles to export miniconda bin location to PATH and set PYTHONHASHSEED ..."
@@ -147,16 +147,16 @@ echo "Updating global profiles to export miniconda bin location to PATH and set
 # Issue here: https://issues.apache.org/jira/browse/SPARK-13330 (fixed in Spark 2.2.0 release)
 # Fix here: http://blog.stuart.axelbrooke.com/python-3-on-spark-return-of-the-pythonhashseed/
 echo "Adding PYTHONHASHSEED=0 to profiles and spark-defaults.conf..."
-echo "export PYTHONHASHSEED=0" | tee -a  "${PROFILE_SCRIPT_PATH}"  #/etc/*bashrc  /usr/lib/spark/conf/spark-env.sh
-echo "spark.executorEnv.PYTHONHASHSEED=0" >> /etc/spark/conf/spark-defaults.conf
+echo "export PYTHONHASHSEED=0" | tee -a "${PROFILE_SCRIPT_PATH}" #/etc/*bashrc  /usr/lib/spark/conf/spark-env.sh
+echo "spark.executorEnv.PYTHONHASHSEED=0" >>/etc/spark/conf/spark-defaults.conf
 
 ## 3. Ensure that Anaconda Python and PySpark play nice
 ### http://blog.cloudera.com/blog/2015/09/how-to-prepare-your-apache-hadoop-cluster-for-pyspark-jobs/
 echo "Ensure that Anaconda Python and PySpark play nice by all pointing to same Python distro..."
-echo "export PYSPARK_PYTHON=$CONDA_BIN_PATH/python" | tee -a  "${PROFILE_SCRIPT_PATH}" /etc/environment /usr/lib/spark/conf/spark-env.sh
+echo "export PYSPARK_PYTHON=$CONDA_BIN_PATH/python" | tee -a "${PROFILE_SCRIPT_PATH}" /etc/environment /usr/lib/spark/conf/spark-env.sh
 
 # CloudSDK libraries are installed in system python
-echo 'export CLOUDSDK_PYTHON=/usr/bin/python' | tee -a "${PROFILE_SCRIPT_PATH}"  #/etc/*bashrc /etc/profile
+echo 'export CLOUDSDK_PYTHON=/usr/bin/python' | tee -a "${PROFILE_SCRIPT_PATH}" #/etc/*bashrc /etc/profile
 
 echo "Finished bootstrapping via Miniconda, sourcing /etc/profile ..."
 source /etc/profile
diff --git a/conda/install-conda-env.sh b/conda/install-conda-env.sh
index ab67037ef..da81cb89e 100755
--- a/conda/install-conda-env.sh
+++ b/conda/install-conda-env.sh
@@ -2,14 +2,14 @@
 set -exo pipefail
 
 if [[ -f /etc/profile.d/effective-python.sh ]]; then
-    PROFILE_SCRIPT_PATH=/etc/profile.d/effective-python.sh
+  PROFILE_SCRIPT_PATH=/etc/profile.d/effective-python.sh
 elif [[ -f /etc/profile.d/conda.sh ]]; then
-    PROFILE_SCRIPT_PATH=/etc/profile.d/conda.sh
+  PROFILE_SCRIPT_PATH=/etc/profile.d/conda.sh
 fi
 
 # 0.1 Ensure we have conda installed and available on the PATH
 if [[ -f "${PROFILE_SCRIPT_PATH}" ]]; then
-    source "${PROFILE_SCRIPT_PATH}"
+  source "${PROFILE_SCRIPT_PATH}"
 fi
 
 echo "echo \$USER: $USER"
@@ -27,18 +27,17 @@ fi
 
 # 0.2. Specify conda environment name (recommend leaving as root)
 if [[ ! -v CONDA_ENV_NAME ]]; then
-    echo "No conda environment name specified, setting to 'root' env..."
-    CONDA_ENV_NAME='root'
+  echo "No conda environment name specified, setting to 'root' env..."
+  CONDA_ENV_NAME='root'
 # Force conda env name to be set to root for now, until a braver soul manages the complexity of environment activation
 # across the cluster.
 else
-    echo "conda environment name is set to $CONDA_ENV_NAME"
-    if [[ ! $CONDA_ENV_NAME == 'root' ]]
-        then
-        echo "Custom conda environment names not supported at this time."
-        echo "Force setting conda env to 'root'..."
-    fi
-    CONDA_ENV_NAME='root'
+  echo "conda environment name is set to $CONDA_ENV_NAME"
+  if [[ ! $CONDA_ENV_NAME == 'root' ]]; then
+    echo "Custom conda environment names not supported at this time."
+    echo "Force setting conda env to 'root'..."
+  fi
+  CONDA_ENV_NAME='root'
 fi
 
 #conda update --all
@@ -46,35 +45,32 @@ fi
 # 1.1 Update conda env from conda environment.yml (if specified)
 # For Dataproc provisioning, we should install to root conda env.
 if [[ -v CONDA_ENV_YAML ]]; then
-    #CONDA_ENV_NAME=$(grep 'name: ' $CONDA_ENV_YAML | awk '{print $2}')
-    # if conda environment name is root, we *update* the root environment with env yaml
-    if [[ $CONDA_ENV_NAME == 'root' ]]
-        then
-        echo "Updating root environment with file $CONDA_ENV_YAML"
-        conda env update --name=$CONDA_ENV_NAME --file=$CONDA_ENV_YAML
-        echo "Root environment updated..."
-    # otherwise, perform a typical environment creation via install
-    else
-        echo "Creating $CONDA_ENV_NAME environment with file $CONDA_ENV_YAML"
-        conda env create --name=$CONDA_ENV_NAME --file=$CONDA_ENV_YAML
-        echo "conda environment $CONDA_ENV_NAME created..."
-    fi
+  #CONDA_ENV_NAME=$(grep 'name: ' $CONDA_ENV_YAML | awk '{print $2}')
+  # if conda environment name is root, we *update* the root environment with env yaml
+  if [[ $CONDA_ENV_NAME == 'root' ]]; then
+    echo "Updating root environment with file $CONDA_ENV_YAML"
+    conda env update --name=$CONDA_ENV_NAME --file=$CONDA_ENV_YAML
+    echo "Root environment updated..."
+  # otherwise, perform a typical environment creation via install
+  else
+    echo "Creating $CONDA_ENV_NAME environment with file $CONDA_ENV_YAML"
+    conda env create --name=$CONDA_ENV_NAME --file=$CONDA_ENV_YAML
+    echo "conda environment $CONDA_ENV_NAME created..."
+  fi
 fi
 
 # 1. Or create conda env manually.
 echo "Attempting to create conda environment: $CONDA_ENV_NAME"
-if conda info --envs | grep -q $CONDA_ENV_NAME
-    then
-    echo "conda environment $CONDA_ENV_NAME detected, skipping env creation..."
+if conda info --envs | grep -q $CONDA_ENV_NAME; then
+  echo "conda environment $CONDA_ENV_NAME detected, skipping env creation..."
 else
-    echo "Creating conda environment directly..."
-    conda create --quiet --yes --name=$CONDA_ENV_NAME python || true
-    echo "conda environment $CONDA_ENV_NAME created..."
+  echo "Creating conda environment directly..."
+  conda create --quiet --yes --name=$CONDA_ENV_NAME python || true
+  echo "conda environment $CONDA_ENV_NAME created..."
 fi
-if [[ ! $CONDA_ENV_NAME == 'root' ]]
-    then
-    echo "Activating $CONDA_ENV_NAME environment..."
-    source activate $CONDA_ENV_NAME
+if [[ ! $CONDA_ENV_NAME == 'root' ]]; then
+  echo "Activating $CONDA_ENV_NAME environment..."
+  source activate $CONDA_ENV_NAME
 fi
 
 # Pin base conda and Python versions to minor version to prevent unexpected upgrades
@@ -89,43 +85,40 @@ CONDA_PINNED_FILE="${CONDA_BASE_PATH}/conda-meta/pinned"
 function pin_component_version() {
   local component=$1
 
-  version=$(conda list "${component}" \
-      | grep -E "^${component}\s+" | sed -E "s/[ ]+/ /g" \
-      | cut -f2 -d' ' | cut -f1,2 -d'.')
-  echo "${component} ${version}.*" >> "${CONDA_PINNED_FILE}"
+  version=$(conda list "${component}" |
+    grep -E "^${component}\s+" | sed -E "s/[ ]+/ /g" |
+    cut -f2 -d' ' | cut -f1,2 -d'.')
+  echo "${component} ${version}.*" >>"${CONDA_PINNED_FILE}"
 }
 pin_component_version conda
 pin_component_version python
 
 # 3. Install conda and pip packages (if specified)
 if [[ ! -z "${CONDA_PACKAGES}" ]]; then
-    echo "Installing conda packages for $CONDA_ENV_NAME..."
-    echo "conda packages requested: $CONDA_PACKAGES"
-    conda install $CONDA_PACKAGES
+  echo "Installing conda packages for $CONDA_ENV_NAME..."
+  echo "conda packages requested: $CONDA_PACKAGES"
+  conda install $CONDA_PACKAGES
 fi
 if [[ ! -z "${PIP_PACKAGES}" ]]; then
-    echo "Installing pip packages for $CONDA_ENV_NAME..."
-    echo "conda packages requested: $PIP_PACKAGES"
-    pip install $PIP_PACKAGES
+  echo "Installing pip packages for $CONDA_ENV_NAME..."
+  echo "conda packages requested: $PIP_PACKAGES"
+  pip install $PIP_PACKAGES
 fi
 
 # 2. Append profiles with conda env source activate
 echo "Attempting to append ${PROFILE_SCRIPT_PATH} to activate conda env at login..."
-if [[ -f "${PROFILE_SCRIPT_PATH}"  ]]  && [[ ! $CONDA_ENV_NAME == 'root' ]]
-    then
-    if grep -ir "source activate $CONDA_ENV_NAME" "${PROFILE_SCRIPT_PATH}"
-        then
-        echo "conda env activation found in ${PROFILE_SCRIPT_PATH}, skipping..."
-    else
-        echo "Appending ${PROFILE_SCRIPT_PATH} to activate conda env $CONDA_ENV_NAME for shell..."
-        sudo echo "source activate $CONDA_ENV_NAME" | tee -a "${PROFILE_SCRIPT_PATH}"
-        echo "${PROFILE_SCRIPT_PATH} successfully appended!"
-    fi
-elif [[ $CONDA_ENV_NAME == 'root' ]]
-    then
-    echo "The conda env specified is 'root', the default environment, no need to activate, skipping..."
+if [[ -f "${PROFILE_SCRIPT_PATH}" ]] && [[ ! $CONDA_ENV_NAME == 'root' ]]; then
+  if grep -ir "source activate $CONDA_ENV_NAME" "${PROFILE_SCRIPT_PATH}"; then
+    echo "conda env activation found in ${PROFILE_SCRIPT_PATH}, skipping..."
+  else
+    echo "Appending ${PROFILE_SCRIPT_PATH} to activate conda env $CONDA_ENV_NAME for shell..."
+    sudo echo "source activate $CONDA_ENV_NAME" | tee -a "${PROFILE_SCRIPT_PATH}"
+    echo "${PROFILE_SCRIPT_PATH} successfully appended!"
+  fi
+elif [[ $CONDA_ENV_NAME == 'root' ]]; then
+  echo "The conda env specified is 'root', the default environment, no need to activate, skipping..."
 else
-    echo "No file detected at ${PROFILE_SCRIPT_PATH}..."
-    echo "Are you sure you installed conda?"
-    exit 1
+  echo "No file detected at ${PROFILE_SCRIPT_PATH}..."
+  echo "Are you sure you installed conda?"
+  exit 1
 fi
diff --git a/docker/docker.sh b/docker/docker.sh
index 0afbbf7a1..e3e6b4668 100755
--- a/docker/docker.sh
+++ b/docker/docker.sh
@@ -8,10 +8,9 @@ readonly OS_CODE=$(lsb_release -cs)
 readonly DOCKER_VERSION="18.06.0~ce~3-0~${OS_ID}"
 readonly CREDENTIAL_HELPER_VERSION='1.5.0'
 
-
 function is_master() {
   local role="$(/usr/share/google/get_metadata_value attributes/dataproc-role)"
-  if [[ "$role" == 'Master' ]] ; then
+  if [[ "$role" == 'Master' ]]; then
     true
   else
     false
@@ -23,7 +22,7 @@ function get_docker_gpg() {
 }
 
 function update_apt_get() {
-  for ((i = 0; i < 10; i++)) ; do
+  for ((i = 0; i < 10; i++)); do
     if apt-get update; then
       return 0
     fi
@@ -44,9 +43,9 @@ function install_docker() {
 function configure_gcr() {
   # this standalone method is recommended here:
   # https://cloud.google.com/container-registry/docs/advanced-authentication#standalone_docker_credential_helper
-  curl -fsSL "https://github.com/GoogleCloudPlatform/docker-credential-gcr/releases/download/v${CREDENTIAL_HELPER_VERSION}/docker-credential-gcr_linux_amd64-${CREDENTIAL_HELPER_VERSION}.tar.gz" \
-    | tar xz --to-stdout ./docker-credential-gcr \
-    > /usr/local/bin/docker-credential-gcr && chmod +x /usr/local/bin/docker-credential-gcr
+  curl -fsSL "https://github.com/GoogleCloudPlatform/docker-credential-gcr/releases/download/v${CREDENTIAL_HELPER_VERSION}/docker-credential-gcr_linux_amd64-${CREDENTIAL_HELPER_VERSION}.tar.gz" |
+    tar xz --to-stdout ./docker-credential-gcr \
+      >/usr/local/bin/docker-credential-gcr && chmod +x /usr/local/bin/docker-credential-gcr
 
   # this command configures docker on a per-user basis. Therefore we configure
   # the root user, as well as the yarn user which is part of the docker group.
@@ -64,7 +63,7 @@ function configure_docker() {
 
   systemctl enable docker
   # Restart YARN daemons to pick up new group without restarting nodes.
-  if is_master ; then
+  if is_master; then
     systemctl restart hadoop-yarn-resourcemanager
   else
     systemctl restart hadoop-yarn-nodemanager
diff --git a/dr-elephant/dr-elephant.sh b/dr-elephant/dr-elephant.sh
index 4efd23a44..989f8cca7 100755
--- a/dr-elephant/dr-elephant.sh
+++ b/dr-elephant/dr-elephant.sh
@@ -130,9 +130,9 @@ the License.
 EOF
 
   bdconfig set_property \
-      --configuration_file "/opt/dr-elephant/app-conf/GeneralConf.xml" \
-      --name 'drelephant.analysis.backfill.enabled' --value 'true' \
-      --clobber
+    --configuration_file "/opt/dr-elephant/app-conf/GeneralConf.xml" \
+    --name 'drelephant.analysis.backfill.enabled' --value 'true' \
+    --clobber
 
   # Enable compression to make metrics accessible by Dr. Elephant
   echo "spark.eventLog.compress = true" >>"/usr/lib/spark/conf/spark-defaults.conf"
diff --git a/drill/drill.sh b/drill/drill.sh
index f0599f8ea..38c825137 100755
--- a/drill/drill.sh
+++ b/drill/drill.sh
@@ -15,17 +15,16 @@ function err() {
 }
 
 function print_err_logs() {
-  for i in ${DRILL_LOG_DIR}/*;
-  do
+  for i in ${DRILL_LOG_DIR}/*; do
     echo ">>> $i"
-    cat "$i";
+    cat "$i"
   done
   return 1
 }
 
 function create_hive_storage_plugin() {
   # Create the hive storage plugin
-  cat > /tmp/hive_plugin.json <<EOF
+  cat >/tmp/hive_plugin.json <<EOF
 {
 "name": "hive",
 "config": {
@@ -45,7 +44,7 @@ EOF
 
 function create_gcs_storage_plugin() {
   # Create GCS storage plugin
-  cat > /tmp/gcs_plugin.json <<EOF
+  cat >/tmp/gcs_plugin.json <<EOF
 {
     "config": {
         "connection": "${gs_plugin_bucket}",
@@ -116,7 +115,7 @@ EOF
 
 function create_hdfs_storage_plugin() {
   # Create/Update hdfs storage plugin
-  cat > /tmp/hdfs_plugin.json <<EOF
+  cat >/tmp/hdfs_plugin.json <<EOF
 {
     "config": {
         "connection": "${hdfs}",
@@ -192,8 +191,8 @@ EOF
 
 function start_drillbit() {
   # Start drillbit
-  sudo -u ${DRILL_USER} ${DRILL_HOME}/bin/drillbit.sh status \
-    || sudo -u ${DRILL_USER} ${DRILL_HOME}/bin/drillbit.sh start && sleep 60
+  sudo -u ${DRILL_USER} ${DRILL_HOME}/bin/drillbit.sh status ||
+    sudo -u ${DRILL_USER} ${DRILL_HOME}/bin/drillbit.sh start && sleep 60
   create_hive_storage_plugin
   create_gcs_storage_plugin
   create_hdfs_storage_plugin
@@ -215,17 +214,17 @@ function main() {
 
   # intelligently generate the zookeeper string
   readonly zookeeper_cfg="/etc/zookeeper/conf/zoo.cfg"
-  readonly zookeeper_client_port=$(grep 'clientPort' ${zookeeper_cfg} \
-    | tail -n 1 \
-    | cut -d '=' -f 2)
-  readonly zookeeper_list=$(grep '^server\.' ${zookeeper_cfg} \
-    | tac \
-    | sort -u -t '='  -k1,1 \
-    | cut -d '=' -f 2 \
-    | cut -d ':' -f 1 \
-    | sed "s/$/:${zookeeper_client_port}/" \
-    | xargs echo \
-    | sed "s/ /,/g")
+  readonly zookeeper_client_port=$(grep 'clientPort' ${zookeeper_cfg} |
+    tail -n 1 |
+    cut -d '=' -f 2)
+  readonly zookeeper_list=$(grep '^server\.' ${zookeeper_cfg} |
+    tac |
+    sort -u -t '=' -k1,1 |
+    cut -d '=' -f 2 |
+    cut -d ':' -f 1 |
+    sed "s/$/:${zookeeper_client_port}/" |
+    xargs echo |
+    sed "s/ /,/g")
 
   # Get hive metastore thrift and HDFS URIs
   local hivemeta=$(bdconfig get_property_value \
@@ -256,7 +255,7 @@ function main() {
   mkdir -p /etc/drill && ln -sf ${DRILL_HOME}/conf /etc/drill/
 
   # Point drill logs to $DRILL_LOG_DIR
-  echo DRILL_LOG_DIR=${DRILL_LOG_DIR} >> ${DRILL_HOME}/conf/drill-env.sh
+  echo DRILL_LOG_DIR=${DRILL_LOG_DIR} >>${DRILL_HOME}/conf/drill-env.sh
 
   # Link GCS connector to drill 3rdparty jars
   local connector_dir
@@ -276,7 +275,7 @@ function main() {
   # Set ZK PStore to use a GCS Bucket
   # Using GCS makes all Drill profiles available from any drillbit, and also
   # persists the profiles past the lifetime of a cluster.
-  cat >> ${DRILL_HOME}/conf/drill-override.conf <<EOF
+  cat >>${DRILL_HOME}/conf/drill-override.conf <<EOF
 drill.exec: { sys.store.provider.zk.blobroot: "${profile_store}" }
 EOF
   chown -R drill:drill /etc/drill/conf/*
diff --git a/flink/validate.sh b/flink/validate.sh
index 32f73357c..556be15d9 100755
--- a/flink/validate.sh
+++ b/flink/validate.sh
@@ -12,7 +12,7 @@ if [[ ${num_workers} == 0 ]] && [[ "${hostname}" == "${master_hostname}" ]]; the
   echo "  Running flink job"
   HADOOP_CONF_DIR=/etc/hadoop/conf /usr/lib/flink/bin/flink run \
     -m yarn-cluster \
-    -yn 1  -yjm 1024 -ytm 1024 \
+    -yn 1 -yjm 1024 -ytm 1024 \
     /usr/lib/flink/examples/batch/WordCount.jar
   flink_return_code=$?
   echo "  Done. Flink return code: ${flink_return_code}"
@@ -21,25 +21,25 @@ else
     echo "  Running flink job because detected yarn session start false"
     HADOOP_CONF_DIR=/etc/hadoop/conf /usr/lib/flink/bin/flink run \
       -m yarn-cluster \
-      -yn 1  -yjm 1024 -ytm 1024 \
+      -yn 1 -yjm 1024 -ytm 1024 \
       -p 2 \
       /usr/lib/flink/examples/batch/WordCount.jar
     yarn_result_code=$?
-  echo "Yarn test result is ${yarn_result_code}"
-else
-  echo 'flink-start-yarn-session set to true, so skipped validation for this case'
-  yarn_result_code=0
-   echo "  Using existing YARN session"
-  application_id=$(yarn --loglevel ERROR application -list | grep -oP "(application_\d+_\d+)") &> /dev/null
-  echo "  Running flink job"
-  HADOOP_CONF_DIR=/etc/hadoop/conf /usr/lib/flink/bin/flink run \
-    -m yarn-cluster \
-    -yid ${application_id} \
-    /usr/lib/flink/examples/batch/WordCount.jar
+    echo "Yarn test result is ${yarn_result_code}"
+  else
+    echo 'flink-start-yarn-session set to true, so skipped validation for this case'
+    yarn_result_code=0
+    echo "  Using existing YARN session"
+    application_id=$(yarn --loglevel ERROR application -list | grep -oP "(application_\d+_\d+)") &>/dev/null
+    echo "  Running flink job"
+    HADOOP_CONF_DIR=/etc/hadoop/conf /usr/lib/flink/bin/flink run \
+      -m yarn-cluster \
+      -yid ${application_id} \
+      /usr/lib/flink/examples/batch/WordCount.jar
 
-  flink_return_code=$?
-  echo "  Done. Flink return code: ${flink_return_code}"
-fi
+    flink_return_code=$?
+    echo "  Done. Flink return code: ${flink_return_code}"
+  fi
 fi
 
 if [[ ${flink_return_code} -eq 0 && ${yarn_result_code} -eq 0 ]]; then
diff --git a/gobblin/gobblin.sh b/gobblin/gobblin.sh
index 868ed5136..d3e4e0dc6 100755
--- a/gobblin/gobblin.sh
+++ b/gobblin/gobblin.sh
@@ -27,7 +27,7 @@ readonly JAR_NAME_CANONICALIZER="s/([-a-zA-Z0-9]+?)[-]([0-9][0-9.]+?)([-.].*?)?.
 
 function maybe_symlink() {
   local jar=$1
-  if [[ ! -f "${HADOOP_LIB}/${jar}" ]] ; then
+  if [[ ! -f "${HADOOP_LIB}/${jar}" ]]; then
     ln -s "${INSTALL_LIB}/${jar}" "${HADOOP_LIB}/${jar}"
   fi
 }
@@ -37,14 +37,14 @@ function configure_env() {
   # Use hdfs:/// so we don't have to disambiguate between Highly Available
   # and regular cluster types.
   sed -E "s/(fs.uri)=(.+)$/\1=hdfs:\/\/\//" \
-      -i "${INSTALL_CONF}/gobblin-mapreduce.properties"
+    -i "${INSTALL_CONF}/gobblin-mapreduce.properties"
 
   sed -E "s/env:GOBBLIN_WORK_DIR/fs.uri/g" \
-      -i "${INSTALL_CONF}/gobblin-mapreduce.properties"
+    -i "${INSTALL_CONF}/gobblin-mapreduce.properties"
 
-  echo "export HADOOP_USER_CLASSPATH_FIRST=true" >> "/etc/hadoop/conf/hadoop-env.sh"
+  echo "export HADOOP_USER_CLASSPATH_FIRST=true" >>"/etc/hadoop/conf/hadoop-env.sh"
 
-  cat << EOF >> "${INSTALL_BIN}/gobblin-env.sh"
+  cat <<EOF >>"${INSTALL_BIN}/gobblin-env.sh"
 export JAVA_HOME=${JAVA_HOME}
 export HADOOP_BIN_DIR=/usr/lib/hadoop/bin
 EOF
@@ -82,7 +82,7 @@ EOF
   rm -f "${HADOOP_LIB}/guava"*
 
   for prefix in "${lib_prefixes[@]}"; do
-    for jar in `ls ${INSTALL_LIB}/${prefix}* | sed 's#.*/##'`; do
+    for jar in $(ls ${INSTALL_LIB}/${prefix}* | sed 's#.*/##'); do
       maybe_symlink "${jar}"
     done
   done
@@ -108,4 +108,3 @@ function main() {
 }
 
 main
-
diff --git a/hive-hcatalog/hive-hcatalog.sh b/hive-hcatalog/hive-hcatalog.sh
index c0999c863..99333d6cb 100755
--- a/hive-hcatalog/hive-hcatalog.sh
+++ b/hive-hcatalog/hive-hcatalog.sh
@@ -39,7 +39,6 @@ function update_apt_get() {
   return 1
 }
 
-
 update_apt_get
 
 # Install the hive-hcatalog package
diff --git a/jupyter/internal/bootstrap-jupyter-ext.sh b/jupyter/internal/bootstrap-jupyter-ext.sh
index 3f16a9932..0560ab068 100755
--- a/jupyter/internal/bootstrap-jupyter-ext.sh
+++ b/jupyter/internal/bootstrap-jupyter-ext.sh
@@ -55,4 +55,3 @@ if [[ ! -d "${RISE_PATH}" ]]; then
 else
   echo "Existing directory at path: ${RISE_PATH}, skipping install!"
 fi
-
diff --git a/jupyter/internal/launch-jupyter-kernel.sh b/jupyter/internal/launch-jupyter-kernel.sh
index 7df0fa17d..63f32dada 100755
--- a/jupyter/internal/launch-jupyter-kernel.sh
+++ b/jupyter/internal/launch-jupyter-kernel.sh
@@ -11,7 +11,7 @@ echo "Installing Jupyter service..."
 readonly JUPYTER_LAUNCHER='/usr/local/bin/launch_jupyter.sh'
 readonly INIT_SCRIPT='/usr/lib/systemd/system/jupyter-notebook.service'
 
-cat << EOF > "${JUPYTER_LAUNCHER}"
+cat <<EOF >"${JUPYTER_LAUNCHER}"
 #!/bin/bash
 
 source /etc/profile.d/conda.sh
@@ -19,7 +19,7 @@ source /etc/profile.d/conda.sh
 EOF
 chmod 750 "${JUPYTER_LAUNCHER}"
 
-cat << EOF > "${INIT_SCRIPT}"
+cat <<EOF >"${INIT_SCRIPT}"
 [Unit]
 Description=Jupyter Notebook Server
 
diff --git a/jupyter/kernels/generate-pyspark.sh b/jupyter/kernels/generate-pyspark.sh
index 68fcff185..c53bc60b1 100755
--- a/jupyter/kernels/generate-pyspark.sh
+++ b/jupyter/kernels/generate-pyspark.sh
@@ -6,7 +6,7 @@
 
 set -e
 
-SPARK_MAJOR_VERSION=$(spark-submit --version |& \
+SPARK_MAJOR_VERSION=$(spark-submit --version |&
   grep 'version' | head -n 1 | sed 's/.*version //' | cut -d '.' -f 1)
 echo "Determined SPARK_MAJOR_VERSION to be '${SPARK_MAJOR_VERSION}'" >&2
 
@@ -18,13 +18,13 @@ PY4J_ZIP=$(ls /usr/lib/spark/python/lib/py4j-*.zip)
 PY4J_ZIP=$(echo ${PY4J_ZIP} | cut -d ' ' -f 1)
 echo "Found PY4J_ZIP: '${PY4J_ZIP}'" >&2
 
-if (( "${SPARK_MAJOR_VERSION}" >= 2 )); then
+if (("${SPARK_MAJOR_VERSION}" >= 2)); then
   PACKAGES_ARG=''
 else
   PACKAGES_ARG='--packages com.databricks:spark-csv_2.10:1.3.0'
 fi
 
-cat << EOF
+cat <<EOF
 {
  "argv": [
     "python", "-m", "ipykernel", "-f", "{connection_file}"],
diff --git a/jupyter/kernels/install-toree.sh b/jupyter/kernels/install-toree.sh
index 5cf0c3763..3e8b7dc16 100755
--- a/jupyter/kernels/install-toree.sh
+++ b/jupyter/kernels/install-toree.sh
@@ -8,10 +8,10 @@ set -e
 
 REPO_URL="https://dist.apache.org/repos/dist"
 /opt/conda/bin/pip \
-    install \
-    ${REPO_URL}/release/incubator/toree/0.2.0-incubating/toree-pip/toree-0.2.0.tar.gz
+  install \
+  ${REPO_URL}/release/incubator/toree/0.2.0-incubating/toree-pip/toree-0.2.0.tar.gz
 
-SPARK_MAJOR_VERSION=$(spark-submit --version |& \
+SPARK_MAJOR_VERSION=$(spark-submit --version |&
   grep 'version' | head -n 1 | sed 's/.*version //' | cut -d '.' -f 1)
 echo "Determined SPARK_MAJOR_VERSION to be '${SPARK_MAJOR_VERSION}'" >&2
 
@@ -24,7 +24,7 @@ PY4J_ZIP=$(echo ${PY4J_ZIP} | cut -d ' ' -f 1)
 echo "Found PY4J_ZIP: '${PY4J_ZIP}'" >&2
 
 COMMON_PACKAGES='org.vegas-viz:vegas_2.11:0.3.11,org.vegas-viz:vegas-spark_2.11:0.3.11'
-if (( "${SPARK_MAJOR_VERSION}" >= 2 )); then
+if (("${SPARK_MAJOR_VERSION}" >= 2)); then
   PACKAGES_ARG="--packages ${COMMON_PACKAGES}"
 else
   PACKAGES_ARG="--packages com.databricks:spark-csv_2.10:1.3.0,${COMMON_PACKAGES}"
@@ -32,6 +32,6 @@ fi
 
 SPARK_OPTS="--master yarn --deploy-mode client ${PACKAGES_ARG}"
 /opt/conda/bin/jupyter toree install \
-    --spark_opts="${SPARK_OPTS}" \
-    --spark_home="/usr/lib/spark" \
-    --kernel_name="Toree"
+  --spark_opts="${SPARK_OPTS}" \
+  --spark_home="/usr/lib/spark" \
+  --kernel_name="Toree"
diff --git a/jupyter/launch-jupyter-interface.sh b/jupyter/launch-jupyter-interface.sh
index 89b57fe5b..cfe3457f5 100755
--- a/jupyter/launch-jupyter-interface.sh
+++ b/jupyter/launch-jupyter-interface.sh
@@ -11,8 +11,8 @@ readonly DIR
 
 source "${DIR}/../util/utils.sh"
 
-function usage {
-  cat << EOF
+function usage() {
+  cat <<EOF
 Creates an SSH tunnel and socks proxy and launches Chrome, using the environment
 variable DATAPROC_CLUSTER_NAME for the unique cluster name. The cluster metadata
 must contain a value for the key 'JUPYTER_PORT'.
@@ -29,22 +29,21 @@ EOF
   exit 1
 }
 
-for i in "$@"
-do
+for i in "$@"; do
   case $i in
-    -z=*)
-      ZONE="${i#*=}"
-      shift # past argument=value
-      ;;
-    -c=*)
-      DATAPROC_CLUSTER_NAME="${i#*=}"
-      shift # past argument=value
-      ;;
-    -h)
-      usage
-      ;;
-    *)
-      ;;
+  -z=*)
+    ZONE="${i#*=}"
+    shift # past argument=value
+    ;;
+  -c=*)
+    DATAPROC_CLUSTER_NAME="${i#*=}"
+    shift # past argument=value
+    ;;
+  -h)
+    usage
+    ;;
+  *) ;;
+
   esac
 done
 
diff --git a/jupyter2/jupyter2.sh b/jupyter2/jupyter2.sh
index a7636e37f..c4656ec80 100755
--- a/jupyter2/jupyter2.sh
+++ b/jupyter2/jupyter2.sh
@@ -46,7 +46,7 @@ function configure_jupyter() {
   echo "Creating Jupyter config..."
   jupyter notebook --allow-root --generate-config -y
 
-  cat << EOF >> ~/.jupyter/jupyter_notebook_config.py
+  cat <<EOF >>~/.jupyter/jupyter_notebook_config.py
 
 ## Configs generated in Dataproc init action
 c.Application.log_level = 'DEBUG'
@@ -65,7 +65,7 @@ EOF
   # {connection_file} is a magic variable that Jupyter fills in for us
   # Note: we can only use it in argv, so cannot use env to set those
   # environment variables.
-  cat << EOF > "${KERNELSPEC_FILE}"
+  cat <<EOF >"${KERNELSPEC_FILE}"
 {
  "argv": [
     "bash",
@@ -78,7 +78,7 @@ EOF
   # Ensure Jupyter has picked up the new kernel
   jupyter kernelspec list | grep pyspark || err "Failed to create kernelspec"
 
-  cat << EOF > "${INIT_SCRIPT}"
+  cat <<EOF >"${INIT_SCRIPT}"
 [Unit]
 Description=Jupyter Notebook Server
 After=hadoop-yarn-resourcemanager.service
@@ -116,4 +116,3 @@ function main() {
 }
 
 main
-
diff --git a/kafka/cruise-control.sh b/kafka/cruise-control.sh
index 60ef0b586..e94e00306 100755
--- a/kafka/cruise-control.sh
+++ b/kafka/cruise-control.sh
@@ -57,7 +57,7 @@ function update_kafka_metrics_reporter() {
   fi
 
   cp ${CRUISE_CONTROL_HOME}/cruise-control-metrics-reporter/build/libs/cruise-control-metrics-reporter-${CRUISE_CONTROL_VERSION}.jar \
-      ${KAFKA_HOME}/libs
+    ${KAFKA_HOME}/libs
   cat >>${KAFKA_CONFIG_FILE} <<EOF
 
 # Properties added by Cruise Control init action.
@@ -86,7 +86,7 @@ EOF
 
 function start_cruise_control_server() {
   # Wait for the metrics topic to be created.
-  for ((i=1; i<=20; i++)); do
+  for ((i = 1; i <= 20; i++)); do
     local metrics_topic=$(/usr/lib/kafka/bin/kafka-topics.sh --list --zookeeper localhost:2181 | grep __CruiseControlMetrics)
     if [[ -n "${metrics_topic}" ]]; then
       break
diff --git a/kafka/kafka-manager.sh b/kafka/kafka-manager.sh
index 3a5d432bf..99af7f3ba 100755
--- a/kafka/kafka-manager.sh
+++ b/kafka/kafka-manager.sh
@@ -52,17 +52,17 @@ function build_kafka_manager() {
 
 # Returns list of zookeeper servers configured in the zoo.cfg file.
 function get_zookeeper_list() {
-  local zookeeper_client_port=$(grep 'clientPort' "${ZOOKEEPER_CONFIG}" \
-      | tail -n 1 \
-      | cut -d '=' -f 2)
-  local zookeeper_list=$(grep '^server.' "${ZOOKEEPER_CONFIG}" \
-      | tac \
-      | sort -u -t '=' -k1,1 \
-      | cut -d '=' -f 2 \
-      | cut -d ':' -f 1 \
-      | sed "s/$/:${zookeeper_client_port}/" \
-      | xargs echo \
-      | sed "s/ /,/g")
+  local zookeeper_client_port=$(grep 'clientPort' "${ZOOKEEPER_CONFIG}" |
+    tail -n 1 |
+    cut -d '=' -f 2)
+  local zookeeper_list=$(grep '^server.' "${ZOOKEEPER_CONFIG}" |
+    tac |
+    sort -u -t '=' -k1,1 |
+    cut -d '=' -f 2 |
+    cut -d ':' -f 1 |
+    sed "s/$/:${zookeeper_client_port}/" |
+    xargs echo |
+    sed "s/ /,/g")
   echo "${zookeeper_list}"
 }
 
@@ -149,9 +149,9 @@ function start_kafka_manager() {
 
   echo "Starting Kafka Manager server on ${HOSTNAME}:${KAFKA_MANAGER_HTTP_PORT}."
   ${KAFKA_MANAGER_HOME}/bin/kafka-manager \
-      -Dconfig.file=${KAFKA_MANAGER_CONFIG} \
-      -Dapplication.home=${KAFKA_MANAGER_HOME} \
-      -Dhttp.port=${KAFKA_MANAGER_HTTP_PORT} &
+    -Dconfig.file=${KAFKA_MANAGER_CONFIG} \
+    -Dapplication.home=${KAFKA_MANAGER_HOME} \
+    -Dhttp.port=${KAFKA_MANAGER_HTTP_PORT} &
 
   # Wait until znode /kafka-manager/configs is created by Kafka Manager.
   wait_for_zookeeper /kafka-manager/configs
diff --git a/kafka/kafka.sh b/kafka/kafka.sh
index 83938d3d7..c6ab0401d 100755
--- a/kafka/kafka.sh
+++ b/kafka/kafka.sh
@@ -59,10 +59,10 @@ function err() {
 # Returns the list of broker IDs registered in ZooKeeper, e.g., " 0, 2, 1,".
 function get_broker_list() {
   ${KAFKA_HOME}/bin/zookeeper-shell.sh "${ZOOKEEPER_ADDRESS}" \
-      <<< "ls /brokers/ids" \
-      | grep '\[.*\]' \
-      | sed 's/\[/ /' \
-      | sed 's/\]/,/'
+    <<<"ls /brokers/ids" |
+    grep '\[.*\]' |
+    sed 's/\[/ /' |
+    sed 's/\]/,/'
 }
 
 # Waits for zookeeper to be up or time out.
@@ -97,19 +97,19 @@ function wait_for_kafka() {
 function install_and_configure_kafka_server() {
   # Find zookeeper list first, before attempting any installation.
   local zookeeper_client_port
-  zookeeper_client_port=$(grep 'clientPort' /etc/zookeeper/conf/zoo.cfg \
-    | tail -n 1 \
-    | cut -d '=' -f 2)
+  zookeeper_client_port=$(grep 'clientPort' /etc/zookeeper/conf/zoo.cfg |
+    tail -n 1 |
+    cut -d '=' -f 2)
 
   local zookeeper_list
-  zookeeper_list=$(grep '^server\.' /etc/zookeeper/conf/zoo.cfg \
-    | cut -d '=' -f 2 \
-    | cut -d ':' -f 1 \
-    | sort \
-    | uniq \
-    | sed "s/$/:${zookeeper_client_port}/" \
-    | xargs echo  \
-    | sed "s/ /,/g")
+  zookeeper_list=$(grep '^server\.' /etc/zookeeper/conf/zoo.cfg |
+    cut -d '=' -f 2 |
+    cut -d ':' -f 1 |
+    sort |
+    uniq |
+    sed "s/$/:${zookeeper_client_port}/" |
+    xargs echo |
+    sed "s/ /,/g")
 
   if [[ -z "${zookeeper_list}" ]]; then
     # Didn't find zookeeper quorum in zoo.cfg, but possibly workers just didn't
@@ -127,8 +127,8 @@ function install_and_configure_kafka_server() {
   ZOOKEEPER_ADDRESS="${zookeeper_list%%,*}"
 
   # Install Kafka from Dataproc distro.
-  install_apt_get kafka-server || dpkg -l kafka-server \
-    || err 'Unable to install and find kafka-server.'
+  install_apt_get kafka-server || dpkg -l kafka-server ||
+    err 'Unable to install and find kafka-server.'
 
   mkdir -p /var/lib/kafka-logs
   chown kafka:kafka -R /var/lib/kafka-logs
@@ -152,8 +152,8 @@ function install_and_configure_kafka_server() {
     "${KAFKA_PROP_FILE}"
   sed -i 's,^\(broker\.id=\).*,\1'${BROKER_ID}',' \
     "${KAFKA_PROP_FILE}"
-  echo -e '\nreserved.broker.max.id=100000' >> "${KAFKA_PROP_FILE}"
-  echo -e '\ndelete.topic.enable=true' >> "${KAFKA_PROP_FILE}"
+  echo -e '\nreserved.broker.max.id=100000' >>"${KAFKA_PROP_FILE}"
+  echo -e '\ndelete.topic.enable=true' >>"${KAFKA_PROP_FILE}"
 
   if [[ "${KAFKA_ENABLE_JMX}" == "true" ]]; then
     sed -i '/kafka-run-class.sh/i export KAFKA_JMX_OPTS="-Dcom.sun.management.jmxremote=true -Dcom.sun.management.jmxremote.authenticate=false -Dcom.sun.management.jmxremote.ssl=false -Djava.rmi.server.hostname=localhost -Djava.net.preferIPv4Stack=true"' /usr/lib/kafka/bin/kafka-server-start.sh
@@ -173,16 +173,16 @@ function main() {
 
   # Only run the installation on workers; verify zookeeper on master(s).
   if [[ "${ROLE}" == 'Master' ]]; then
-    service zookeeper-server status \
-      || err 'Required zookeeper-server not running on master!'
+    service zookeeper-server status ||
+      err 'Required zookeeper-server not running on master!'
     if [[ "${RUN_ON_MASTER}" == "true" ]]; then
       # Run installation on masters.
       install_and_configure_kafka_server
     else
       # On master nodes, just install kafka command-line tools and libs but not
       # kafka-server.
-      install_apt_get kafka \
-        || err 'Unable to install kafka libraries on master!'
+      install_apt_get kafka ||
+        err 'Unable to install kafka libraries on master!'
     fi
   else
     # Run installation on workers.
diff --git a/kafka/validate.sh b/kafka/validate.sh
index 196e7b109..062ee65d5 100755
--- a/kafka/validate.sh
+++ b/kafka/validate.sh
@@ -7,25 +7,27 @@ hostname="$(hostname)"
 echo "---------------------------------"
 echo "Starting validation on ${hostname}"
 
-
 # Create a test topic, just talking to the local master's zookeeper server.
 /usr/lib/kafka/bin/kafka-topics.sh --zookeeper localhost:2181 --create \
-    --replication-factor 1 --partitions 1 --topic ${hostname}
+  --replication-factor 1 --partitions 1 --topic ${hostname}
 /usr/lib/kafka/bin/kafka-topics.sh --zookeeper localhost:2181 --list
 
 # Use worker 0 as broker to publish 100 messages over 100 seconds
 # asynchronously.
 echo "Testing worker-0 as producer and worker-1 as consumer"
 CLUSTER_NAME=$(/usr/share/google/get_metadata_value attributes/dataproc-cluster-name)
-for i in {0..50}; do echo "message${i}"; sleep 1; done | \
-    /usr/lib/kafka/bin/kafka-console-producer.sh \
+for i in {0..50}; do
+  echo "message${i}"
+  sleep 1
+done |
+  /usr/lib/kafka/bin/kafka-console-producer.sh \
     --broker-list ${CLUSTER_NAME}-w-0:9092 --topic ${hostname} &
 
 # User worker 1 as broker to consume those 100 messages as they come.
 # This can also be run in any other master or worker node of the cluster.
 /usr/lib/kafka/bin/kafka-console-consumer.sh \
-    --bootstrap-server ${CLUSTER_NAME}-w-1:9092 \
-    --topic ${hostname} --from-beginning > /tmp/messages &
+  --bootstrap-server ${CLUSTER_NAME}-w-1:9092 \
+  --topic ${hostname} --from-beginning >/tmp/messages &
 
 received_messages=$(wc -l /tmp/messages | awk '{print $1}')
 while [[ ${received_messages} -ne 51 ]]; do
@@ -35,13 +37,16 @@ while [[ ${received_messages} -ne 51 ]]; do
 done
 
 echo "Testing worker-0 as consumer and worker-1 as producer"
-for i in {0..50}; do echo "message${i}"; sleep 1; done | \
-    /usr/lib/kafka/bin/kafka-console-producer.sh \
+for i in {0..50}; do
+  echo "message${i}"
+  sleep 1
+done |
+  /usr/lib/kafka/bin/kafka-console-producer.sh \
     --broker-list ${CLUSTER_NAME}-w-1:9092 --topic ${hostname} &
 
 /usr/lib/kafka/bin/kafka-console-consumer.sh \
-    --bootstrap-server ${CLUSTER_NAME}-w-0:9092 \
-    --topic ${hostname} --from-beginning > /tmp/messages1 &
+  --bootstrap-server ${CLUSTER_NAME}-w-0:9092 \
+  --topic ${hostname} --from-beginning >/tmp/messages1 &
 
 received_messages=$(wc -l /tmp/messages1 | awk '{print $1}')
 while [[ ${received_messages} -ne 102 ]]; do
diff --git a/oozie/oozie.sh b/oozie/oozie.sh
index cb5f6fcf0..c5dd1b753 100755
--- a/oozie/oozie.sh
+++ b/oozie/oozie.sh
@@ -62,7 +62,7 @@ function main() {
   fi
 }
 
-function install_oozie(){
+function install_oozie() {
   local master_node=$(/usr/share/google/get_metadata_value attributes/dataproc-master)
   local node_name=${HOSTNAME}
 
@@ -71,7 +71,7 @@ function install_oozie(){
   install_apt_get oozie oozie-client || err 'Unable to install oozie-client'
   # Remove Log4j 2 jar not compatible with Log4j 1 that was brought by Hive 2
   # TODO: remove after upgrade to Oozie 5.1
-  if compgen -G "/usr/lib/oozie/lib/log4j-1.2.*.jar" > /dev/null; then
+  if compgen -G "/usr/lib/oozie/lib/log4j-1.2.*.jar" >/dev/null; then
     rm -f /usr/lib/oozie/lib/log4j-1.2-api*.jar
   fi
 
@@ -100,36 +100,35 @@ function install_oozie(){
   # Hadoop must allow impersonation for Oozie to work properly
 
   bdconfig set_property \
-      --configuration_file "/etc/hadoop/conf/core-site.xml" \
-      --name 'hadoop.proxyuser.oozie.hosts' --value '*' \
-      --clobber
+    --configuration_file "/etc/hadoop/conf/core-site.xml" \
+    --name 'hadoop.proxyuser.oozie.hosts' --value '*' \
+    --clobber
 
   bdconfig set_property \
     --configuration_file "/etc/hadoop/conf/core-site.xml" \
     --name 'hadoop.proxyuser.oozie.groups' --value '*' \
     --clobber
 
-
   # Detect if current node configuration is HA and then set oozie servers
   local additional_nodes=$(/usr/share/google/get_metadata_value attributes/dataproc-master-additional | sed 's/,/\n/g' | wc -l)
   if [[ ${additional_nodes} -ge 2 ]]; then
     echo 'Starting configuration for HA'
     # List of servers is used for proper zookeeper configuration. It is needed to replace original ports range with specific one
-    local servers=$(cat /usr/lib/zookeeper/conf/zoo.cfg \
-      | grep 'server.' \
-      | sed 's/server.//g' \
-      | sed 's/:2888:3888//g' \
-      | cut -d'=' -f2- \
-      | sed 's/\n/,/g' \
-      | head -n 3 \
-      | sed 's/$/:2181,/g' \
-      | xargs -L3 \
-      | sed 's/.$//g')
+    local servers=$(cat /usr/lib/zookeeper/conf/zoo.cfg |
+      grep 'server.' |
+      sed 's/server.//g' |
+      sed 's/:2888:3888//g' |
+      cut -d'=' -f2- |
+      sed 's/\n/,/g' |
+      head -n 3 |
+      sed 's/$/:2181,/g' |
+      xargs -L3 |
+      sed 's/.$//g')
 
     bdconfig set_property \
       --configuration_file "/etc/oozie/conf/oozie-site.xml" \
       --name 'oozie.services.ext' --value \
-        'org.apache.oozie.service.ZKLocksService,
+      'org.apache.oozie.service.ZKLocksService,
         org.apache.oozie.service.ZKXLogStreamingService,
         org.apache.oozie.service.ZKJobsConcurrencyService,
         org.apache.oozie.service.ZKUUIDService' \
@@ -145,7 +144,7 @@ function install_oozie(){
   /usr/lib/zookeeper/bin/zkServer.sh restart
   # HDFS and YARN must be cycled; restart to clean things up
   for service in hadoop-hdfs-namenode hadoop-hdfs-secondarynamenode hadoop-yarn-resourcemanager oozie; do
-    if [[ $(systemctl list-unit-files | grep ${service}) != '' ]] && \
+    if [[ $(systemctl list-unit-files | grep ${service}) != '' ]] &&
       [[ $(systemctl is-enabled ${service}) == 'enabled' ]]; then
       systemctl restart ${service}
     fi
diff --git a/oozie/validate.sh b/oozie/validate.sh
index fe5309082..01af40dc5 100755
--- a/oozie/validate.sh
+++ b/oozie/validate.sh
@@ -17,7 +17,6 @@ fi
 rm -f job.properties
 hdfs dfs -get "oozie-examples/apps/map-reduce/job.properties" job.properties
 
-
 echo "---------------------------------"
 echo "Starting validation on ${HOSTNAME}"
 sudo -u hdfs hadoop dfsadmin -safemode leave &>/dev/null
diff --git a/post-init/master-post-init.sh b/post-init/master-post-init.sh
index 61428caf2..1ef2f7282 100755
--- a/post-init/master-post-init.sh
+++ b/post-init/master-post-init.sh
@@ -25,12 +25,12 @@ if [[ "${ROLE}" != 'Master' ]]; then
 fi
 
 CLUSTER_NAME=$(curl -f -s -H Metadata-Flavor:Google \
-    ${METADATA_ROOT}/dataproc-cluster-name)
+  ${METADATA_ROOT}/dataproc-cluster-name)
 
 # Fetch the actual command we want to run once the cluster is healthy.
 # The command is specified with the 'post-init-command' key.
 POST_INIT_COMMAND=$(curl -f -s -H Metadata-Flavor:Google \
-    ${METADATA_ROOT}/post-init-command)
+  ${METADATA_ROOT}/post-init-command)
 
 if [ -z ${POST_INIT_COMMAND} ]; then
   echo "Failed to find metadata key 'post-init-command'"
@@ -39,7 +39,7 @@ fi
 
 # We must put the bulk of the login in a separate helper script so that we can
 # 'nohup' it.
-cat << EOF > /usr/local/bin/await_cluster_and_run_command.sh
+cat <<EOF >/usr/local/bin/await_cluster_and_run_command.sh
 #!/bin/bash
 
 # Helper to get current cluster state.
@@ -73,5 +73,5 @@ chmod 750 /usr/local/bin/await_cluster_and_run_command.sh
 # Uncomment this following line and comment out the line after it to throw away
 # the stdout/stderr of the command instead of logging it.
 #nohup /usr/local/bin/await_cluster_and_run_command.sh &>> /dev/null &
-nohup /usr/local/bin/await_cluster_and_run_command.sh &>> \
-    /var/log/master-post-init.log &
+nohup /usr/local/bin/await_cluster_and_run_command.sh &>>/var/log/master-post-init.log \
+  ;
diff --git a/prometheus/prometheus.sh b/prometheus/prometheus.sh
index 04e3e9361..1938afc2a 100755
--- a/prometheus/prometheus.sh
+++ b/prometheus/prometheus.sh
@@ -32,7 +32,7 @@ readonly KAFKA_JMX_JAVAAGENT_URI="http://central.maven.org/maven2/io/prometheus/
 readonly KAFKA_JMX_EXPORTER_CONFIG_NAME="kafka-0-8-2.yml"
 readonly KAFKA_JMX_EXPORTER_CONFIG_URI="https://raw.githubusercontent.com/prometheus/jmx_exporter/master/example_configs/${KAFKA_JMX_EXPORTER_CONFIG_NAME}"
 
-function is_kafka_installed {
+function is_kafka_installed() {
   local result="$(cat ${KAFKA_CONFIG_FILE} | grep broker.id | tail -1)"
   if [[ "${result}" == "broker.id=0" ]]; then
     return 1
@@ -41,7 +41,7 @@ function is_kafka_installed {
   fi
 }
 
-function install_prometheus {
+function install_prometheus() {
   mkdir -p /etc/prometheus /var/lib/prometheus
   wget -nv --timeout=30 --tries=5 --retry-connrefused \
     "https://github.com/prometheus/prometheus/releases/download/v${PROMETHEUS_VER}/prometheus-${PROMETHEUS_VER}.linux-amd64.tar.gz"
@@ -53,7 +53,7 @@ function install_prometheus {
 
   rm -rf "prometheus-${PROMETHEUS_VER}.linux-amd64.tar.gz" "prometheus-${PROMETHEUS_VER}.linux-amd64"
 
-cat << EOF > /etc/systemd/system/prometheus.service
+  cat <<EOF >/etc/systemd/system/prometheus.service
 [Unit]
 Description=Prometheus
 Wants=network-online.target
@@ -75,9 +75,9 @@ WantedBy=multi-user.target
 EOF
 }
 
-function configure_prometheus {
+function configure_prometheus() {
   # Statsd for Hadoop and Spark.
-  cat << EOF > /etc/prometheus/prometheus.yml
+  cat <<EOF >/etc/prometheus/prometheus.yml
 global:
   scrape_interval: 10s
   evaluation_interval: 10s
@@ -91,7 +91,7 @@ EOF
 
   # Kafka JMX exporter.
   if [[ "${MONITOR_KAFKA}" == "true" ]] && is_kafka_installed; then
-    cat << EOF >> /etc/prometheus/prometheus.yml
+    cat <<EOF >>/etc/prometheus/prometheus.yml
   - job_name: 'kafka'
     static_configs:
       - targets: ['localhost:${KAFKA_JMX_EXPORTER_PORT}']
@@ -99,7 +99,7 @@ EOF
   fi
 }
 
-function install_statsd_exporter {
+function install_statsd_exporter() {
   mkdir -p /var/lib/statsd
   wget -nv --timeout=30 --tries=5 --retry-connrefused \
     "https://github.com/prometheus/statsd_exporter/releases/download/v${STATSD_EXPORTER_VER}/statsd_exporter-${STATSD_EXPORTER_VER}.linux-amd64.tar.gz"
@@ -108,7 +108,7 @@ function install_statsd_exporter {
 
   rm -rf "statsd_exporter-${STATSD_EXPORTER_VER}.linux-amd64.tar.gz" "statsd_exporter-${STATSD_EXPORTER_VER}.linux-amd64"
 
-cat << EOF > /etc/systemd/system/statsd-exporter.service
+  cat <<EOF >/etc/systemd/system/statsd-exporter.service
 [Unit]
 Description=Statsd
 Wants=network-online.target
@@ -126,23 +126,23 @@ WantedBy=multi-user.target
 EOF
 }
 
-function install_jmx_exporter {
+function install_jmx_exporter() {
   wget -nv --timeout=30 --tries=5 --retry-connrefused \
-     "${KAFKA_JMX_JAVAAGENT_URI}" -P "${KAFKA_LIBS_DIR}"
+    "${KAFKA_JMX_JAVAAGENT_URI}" -P "${KAFKA_LIBS_DIR}"
   wget -nv --timeout=30 --tries=5 --retry-connrefused \
-     "${KAFKA_JMX_EXPORTER_CONFIG_URI}" -P "${KAFKA_CONFIG_DIR}"
+    "${KAFKA_JMX_EXPORTER_CONFIG_URI}" -P "${KAFKA_CONFIG_DIR}"
   sed -i "/kafka-run-class.sh/i export KAFKA_OPTS=\"\${KAFKA_OPTS} -javaagent:${KAFKA_LIBS_DIR}/${KAFKA_JMX_JAVAAGENT_NAME}=${KAFKA_JMX_EXPORTER_PORT}:${KAFKA_CONFIG_DIR}/${KAFKA_JMX_EXPORTER_CONFIG_NAME}\"" \
-        /usr/lib/kafka/bin/kafka-server-start.sh
+    /usr/lib/kafka/bin/kafka-server-start.sh
 }
 
-function start_services {
+function start_services() {
   systemctl daemon-reload
   systemctl start statsd-exporter
   systemctl start prometheus
 }
 
-function configure_hadoop {
-  cat << EOF > /etc/hadoop/conf/hadoop-metrics2.properties
+function configure_hadoop() {
+  cat <<EOF >/etc/hadoop/conf/hadoop-metrics2.properties
 resourcemanager.sink.statsd.class=org.apache.hadoop.metrics2.sink.StatsDSink
 resourcemanager.sink.statsd.server.host=${HOSTNAME}
 resourcemanager.sink.statsd.server.port=9125
@@ -163,20 +163,20 @@ EOF
   fi
 }
 
-function configure_spark {
-  cat << EOF > /etc/spark/conf/metrics.properties
+function configure_spark() {
+  cat <<EOF >/etc/spark/conf/metrics.properties
 *.sink.statsd.class=org.apache.spark.metrics.sink.StatsdSink
 *.sink.statsd.prefix=spark
 *.sink.statsd.port=9125
 EOF
 }
 
-function configure_kafka {
+function configure_kafka() {
   install_jmx_exporter
   systemctl restart kafka-server.service
 }
 
-function configure_components {
+function configure_components() {
   if [[ "${MONITOR_HADOOP}" == "true" || "${MONITOR_SPARK}" == "true" ]]; then
     install_statsd_exporter
 
@@ -194,13 +194,13 @@ function configure_components {
   fi
 }
 
-function restart_service_gracefully {
+function restart_service_gracefully() {
   while true; do
-    if systemctl status "$1"| grep -q 'Active: active (running)'; then
+    if systemctl status "$1" | grep -q 'Active: active (running)'; then
       systemctl restart "$1"
-      break;
+      break
     fi
-    sleep 5;
+    sleep 5
   done
 }
 
diff --git a/python/conda-install.sh b/python/conda-install.sh
index 007a35d60..1053b1a40 100755
--- a/python/conda-install.sh
+++ b/python/conda-install.sh
@@ -6,11 +6,10 @@ readonly PACKAGES=$(/usr/share/google/get_metadata_value attributes/CONDA_PACKAG
 
 function main() {
   if [[ -z "${PACKAGES}" ]]; then
-    echo "ERROR: Must specify CONDA_PACKAGES metadata key" 
+    echo "ERROR: Must specify CONDA_PACKAGES metadata key"
     exit 1
   fi
   conda install ${PACKAGES}
 }
 
 main
-
diff --git a/ranger/ranger.sh b/ranger/ranger.sh
index f4dab7c6b..fb0ac76f2 100755
--- a/ranger/ranger.sh
+++ b/ranger/ranger.sh
@@ -60,14 +60,14 @@ function configure_admin() {
   sed -i 's/^audit_solr_user=/audit_solr_user=solr/' \
     "${RANGER_INSTALL_DIR}/ranger-admin/install.properties"
   bdconfig set_property \
-      --configuration_file "${RANGER_INSTALL_DIR}/ranger-admin/ews/webapp/WEB-INF/classes/conf.dist/ranger-admin-site.xml" \
-      --name 'ranger.service.http.port' --value "${RANGER_ADMIN_PORT}" \
-      --clobber
+    --configuration_file "${RANGER_INSTALL_DIR}/ranger-admin/ews/webapp/WEB-INF/classes/conf.dist/ranger-admin-site.xml" \
+    --name 'ranger.service.http.port' --value "${RANGER_ADMIN_PORT}" \
+    --clobber
   mysql -u root -proot-password -e "CREATE USER 'rangeradmin'@'localhost' IDENTIFIED BY 'rangerpass';"
   mysql -u root -proot-password -e "CREATE DATABASE ranger;"
   mysql -u root -proot-password -e "GRANT ALL PRIVILEGES ON ranger.* TO 'rangeradmin'@'localhost';"
 
-  if [[ "${MASTER_ADDITIONAL}" != "" ]] ; then
+  if [[ "${MASTER_ADDITIONAL}" != "" ]]; then
     sed -i "s/^audit_solr_zookeepers=/audit_solr_zookeepers=${CLUSTER_NAME}-m-0:2181,${CLUSTER_NAME}-m-1:2181,${CLUSTER_NAME}-m-2:2181\/solr/" \
       "${RANGER_INSTALL_DIR}/ranger-admin/install.properties"
     sed -i 's/^audit_solr_urls=/audit_solr_urls=none/' \
@@ -88,8 +88,8 @@ function run_ranger_admin() {
 }
 
 function add_usersync_plugin() {
-  mkdir -p /var/log/ranger-usersync && chown ranger /var/log/ranger-usersync \
-    && chgrp ranger /var/log/ranger-usersync
+  mkdir -p /var/log/ranger-usersync && chown ranger /var/log/ranger-usersync &&
+    chgrp ranger /var/log/ranger-usersync
 
   sed -i 's/^logdir=logs/logdir=\/var\/log\/ranger-usersync/' \
     "${RANGER_INSTALL_DIR}/ranger-usersync/install.properties"
@@ -144,12 +144,11 @@ function add_hdfs_plugin() {
     systemctl start hadoop-hdfs-namenode.service
 
     # Notify cluster that plugin is installed on master.
-    until hadoop fs -touchz /tmp/ranger-hdfs-plugin-ready  &> /dev/null
-    do
+    until hadoop fs -touchz /tmp/ranger-hdfs-plugin-ready &>/dev/null; do
       sleep 10
     done
 
-    cat << EOF > service-hdfs.json
+    cat <<EOF >service-hdfs.json
 {
     "configs": {
         "username": "admin",
@@ -165,12 +164,11 @@ function add_hdfs_plugin() {
     "version": 1
 }
 EOF
-    curl --user "admin:${RANGER_ADMIN_PASS}" -H "Content-Type: application/json"  \
+    curl --user "admin:${RANGER_ADMIN_PASS}" -H "Content-Type: application/json" \
       -X POST -d @service-hdfs.json "http://localhost:${RANGER_ADMIN_PORT}/service/public/v2/api/service"
   elif [[ "${NODE_NAME}" =~ ^.*(-m-1)$ ]]; then
     # Waiting until hdfs plugin will be configured on m-0
-    until hadoop fs -ls /tmp/ranger-hdfs-plugin-ready &> /dev/null
-    do
+    until hadoop fs -ls /tmp/ranger-hdfs-plugin-ready &>/dev/null; do
       sleep 10
     done
     systemctl stop hadoop-hdfs-namenode.service
@@ -180,9 +178,9 @@ EOF
 
 function add_hive_plugin() {
   apply_common_plugin_configuration "ranger-hive-plugin" "hive-dataproc"
-  mkdir -p hive \
-    && ln -s /etc/hive/conf hive \
-    && ln -s /usr/lib/hive/lib hive
+  mkdir -p hive &&
+    ln -s /etc/hive/conf hive &&
+    ln -s /usr/lib/hive/lib hive
   pushd ranger-hive-plugin && ./enable-hive-plugin.sh && popd
 
   if [[ "${NODE_NAME}" =~ ^.*(-m|-m-0)$ ]]; then
@@ -190,12 +188,11 @@ function add_hive_plugin() {
     systemctl start hive-server2.service
 
     # Notify cluster that hive plugin is installed on master.
-    until hadoop fs -touchz /tmp/ranger-hive-plugin-ready  &> /dev/null
-    do
+    until hadoop fs -touchz /tmp/ranger-hive-plugin-ready &>/dev/null; do
       sleep 10
     done
 
-    cat << EOF > service-hive.json
+    cat <<EOF >service-hive.json
 {
     "configs": {
         "username": "admin",
@@ -210,12 +207,11 @@ function add_hive_plugin() {
     "version": 1
 }
 EOF
-    curl --user "admin:${RANGER_ADMIN_PASS}" -H "Content-Type: application/json"  \
+    curl --user "admin:${RANGER_ADMIN_PASS}" -H "Content-Type: application/json" \
       -X POST -d @service-hive.json "http://localhost:${RANGER_ADMIN_PORT}/service/public/v2/api/service"
   elif [[ "${NODE_NAME}" =~ ^.*(-m-1|-m-2)$ ]]; then
     # Waiting until hive plugin will be configured on m-0
-    until hadoop fs -ls /tmp/ranger-hive-plugin-ready &> /dev/null
-    do
+    until hadoop fs -ls /tmp/ranger-hive-plugin-ready &>/dev/null; do
       sleep 10
     done
     systemctl stop hive-server2.service
@@ -232,12 +228,11 @@ function add_yarn_plugin() {
     systemctl start hadoop-yarn-resourcemanager.service
 
     # Notify cluster that yarn plugin is installed on master.
-    until hadoop fs -touchz /tmp/ranger-yarn-plugin-ready  &> /dev/null
-    do
+    until hadoop fs -touchz /tmp/ranger-yarn-plugin-ready &>/dev/null; do
       sleep 10
     done
 
-    cat << EOF > service-yarn.json
+    cat <<EOF >service-yarn.json
 {
     "configs": {
         "username": "admin",
@@ -251,12 +246,11 @@ function add_yarn_plugin() {
     "version": 1
 }
 EOF
-    curl --user "admin:${RANGER_ADMIN_PASS}" -H "Content-Type: application/json"  \
+    curl --user "admin:${RANGER_ADMIN_PASS}" -H "Content-Type: application/json" \
       -X POST -d @service-yarn.json "http://localhost:${RANGER_ADMIN_PORT}/service/public/v2/api/service"
   elif [[ "${NODE_NAME}" =~ ^.*(-m-1|-m-2)$ ]]; then
     # Waiting until yarn plugin will be configured on m-0
-    until hadoop fs -ls /tmp/ranger-yarn-plugin-ready &> /dev/null
-    do
+    until hadoop fs -ls /tmp/ranger-yarn-plugin-ready &>/dev/null; do
       sleep 10
     done
     systemctl stop hadoop-yarn-resourcemanager.service
diff --git a/rapids/internal/launch-dask.sh b/rapids/internal/launch-dask.sh
index b262f547d..2f985446f 100755
--- a/rapids/internal/launch-dask.sh
+++ b/rapids/internal/launch-dask.sh
@@ -14,7 +14,7 @@ readonly INIT_SCRIPT='/usr/lib/systemd/system/dask-cluster.service'
 readonly PREFIX='/opt/conda/anaconda/envs/RAPIDS/bin'
 
 if [[ "${ROLE}" == 'Master' ]]; then
-    cat << EOF > "${DASK_LAUNCHER}"
+  cat <<EOF >"${DASK_LAUNCHER}"
 #!/bin/bash
 if [[ "${RUN_CUDA_WORKER_ON_MASTER}" == true ]]; then
   echo "dask-scheduler starting, logging to /var/log/dask-scheduler.log.."
@@ -28,14 +28,14 @@ else
 fi
 EOF
 else
-    cat << EOF > "${DASK_LAUNCHER}"
+  cat <<EOF >"${DASK_LAUNCHER}"
 #!/bin/bash
 $PREFIX/dask-cuda-worker --memory-limit 0 ${MASTER}:8786 > /var/log/dask-cuda-workers.log 2>&1
 EOF
 fi
 chmod 750 "${DASK_LAUNCHER}"
 
-cat << EOF > "${INIT_SCRIPT}"
+cat <<EOF >"${INIT_SCRIPT}"
 [Unit]
 Description=Dask Cluster Service
 [Service]
diff --git a/rstudio/rstudio.sh b/rstudio/rstudio.sh
index 25b06542e..191b2411f 100755
--- a/rstudio/rstudio.sh
+++ b/rstudio/rstudio.sh
@@ -42,12 +42,12 @@ function run_with_retries() {
       break
     else
       local sleep_time=${retry_backoff[$i]}
-      echo "'${cmd[*]}' attempt $(( $i + 1 )) failed! Sleeping ${sleep_time}." >&2
+      echo "'${cmd[*]}' attempt $(($i + 1)) failed! Sleeping ${sleep_time}." >&2
       sleep ${sleep_time}
     fi
   done
 
-  if ! (( ${succeeded} )); then
+  if ! ((${succeeded})); then
     echo "Final attempt of '${cmd[*]}'..."
     # Let any final error propagate all the way out to any error traps.
     "${cmd[@]}"
@@ -63,20 +63,20 @@ OS_ID=$(lsb_release -is | tr '[:upper:]' '[:lower:]')
 OS_CODE=$(lsb_release -cs)
 
 function get_apt_key_for_debian() {
-  apt-key adv --no-tty --keyserver keys.gnupg.net --recv-key E19F5F87128899B192B1A2C2AD5F960A256A04AF || \
-      apt-key adv --no-tty --keyserver pgp.mit.edu --recv-key E19F5F87128899B192B1A2C2AD5F960A256A04AF
+  apt-key adv --no-tty --keyserver keys.gnupg.net --recv-key E19F5F87128899B192B1A2C2AD5F960A256A04AF ||
+    apt-key adv --no-tty --keyserver pgp.mit.edu --recv-key E19F5F87128899B192B1A2C2AD5F960A256A04AF
 }
 
 if [[ "${ROLE}" == 'Master' ]]; then
-  if [[ -n ${USER_PASSWORD} ]] && (( ${#USER_PASSWORD} < 7 )) ; then
+  if [[ -n ${USER_PASSWORD} ]] && ((${#USER_PASSWORD} < 7)); then
     echo "You must specify a password of at least 7 characters for user '$USER_NAME' through metadata 'rstudio-password'."
     exit 1
   fi
-  if [[ -z "${USER_NAME}" ]] ; then
+  if [[ -z "${USER_NAME}" ]]; then
     echo "RStudio user name must not be empty."
     exit 2
   fi
-  if [[ "${USER_NAME}" == "${USER_PASSWORD}" ]] ; then
+  if [[ "${USER_NAME}" == "${USER_PASSWORD}" ]]; then
     echo "RStudio user name and password must not be the same."
     exit 3
   fi
@@ -104,11 +104,11 @@ if [[ "${ROLE}" == 'Master' ]]; then
   fi
   if ! [ $(id -u "${USER_NAME}") ]; then
     useradd --create-home --gid "${USER_NAME}" "${USER_NAME}"
-    if [[ -n "${USER_PASSWORD}" ]] ; then
+    if [[ -n "${USER_PASSWORD}" ]]; then
       echo "${USER_NAME}:${USER_PASSWORD}" | chpasswd
     fi
   fi
-  if [[ -z "${USER_PASSWORD}" ]] ; then
+  if [[ -z "${USER_PASSWORD}" ]]; then
     sed -i 's:ExecStart=\(.*\):Environment=USER=rstudio\nExecStart=\1 --auth-none 1:1' /etc/systemd/system/rstudio-server.service
     systemctl daemon-reload
     systemctl restart rstudio-server
diff --git a/solr/solr.sh b/solr/solr.sh
index 6db95be81..7585799bf 100755
--- a/solr/solr.sh
+++ b/solr/solr.sh
@@ -51,10 +51,10 @@ function err() {
 function install_and_configure_solr() {
   local solr_home_dir
   local zookeeper_nodes
-  zookeeper_nodes="$(grep '^server\.' /etc/zookeeper/conf/zoo.cfg \
-    | uniq | cut -d '=' -f 2 | cut -d ':' -f 1 | xargs echo | sed "s/ /,/g")"
+  zookeeper_nodes="$(grep '^server\.' /etc/zookeeper/conf/zoo.cfg |
+    uniq | cut -d '=' -f 2 | cut -d ':' -f 1 | xargs echo | sed "s/ /,/g")"
 
-# Install deb packages from GS
+  # Install deb packages from GS
   update_apt_get
   install_apt_get solr
 
@@ -75,12 +75,12 @@ function install_and_configure_solr() {
   else
     solr_home_dir="hdfs://${CLUSTER_NAME}-m:8020/solr"
   fi
-  cat << EOF >> "${SOLR_CONF_FILE}"
+  cat <<EOF >>"${SOLR_CONF_FILE}"
 SOLR_OPTS="\${SOLR_OPTS} -Dsolr.directoryFactory=HdfsDirectoryFactory -Dsolr.lock.type=hdfs \
  -Dsolr.hdfs.home=${solr_home_dir}"
 EOF
 
-  cat << EOF > /etc/systemd/system/solr.service
+  cat <<EOF >/etc/systemd/system/solr.service
 [Unit]
 Description=Apache SOLR
 ConditionPathExists=/usr/lib/solr/bin
diff --git a/tez/tez.sh b/tez/tez.sh
index af2032e02..8929d05c4 100755
--- a/tez/tez.sh
+++ b/tez/tez.sh
@@ -57,15 +57,15 @@ function err() {
 
 function configure_master_node() {
   update_apt_get || err 'Unable to update packages lists.'
-  install_apt_get tez hadoop-yarn-timelineserver \
-    || err 'Failed to install required packages.'
+  install_apt_get tez hadoop-yarn-timelineserver ||
+    err 'Failed to install required packages.'
 
   # Copy to hdfs from one master only to avoid race
   if [[ "${HOSTNAME}" == "${master_hostname}" ]]; then
     # Stage Tez
     hadoop fs -mkdir -p ${TEZ_HDFS_PATH}
-    hadoop fs -copyFromLocal ${TEZ_JARS}/* ${TEZ_HDFS_PATH}/ \
-      || err 'Unable to copy tez jars to hdfs destination.'
+    hadoop fs -copyFromLocal ${TEZ_JARS}/* ${TEZ_HDFS_PATH}/ ||
+      err 'Unable to copy tez jars to hdfs destination.'
   fi
 
   # Update the hadoop-env.sh
@@ -73,7 +73,7 @@ function configure_master_node() {
     echo 'export TEZ_CONF_DIR=/etc/tez/'
     echo "export TEZ_JARS=${TEZ_JARS}"
     echo "HADOOP_CLASSPATH=\$HADOOP_CLASSPATH:${TEZ_CONF_DIR}:${TEZ_JARS}/*:${TEZ_JARS}/lib/*"
-  } >> /etc/hadoop/conf/hadoop-env.sh
+  } >>/etc/hadoop/conf/hadoop-env.sh
 
   # Configure YARN to enable the Application Timeline Server.
   bdconfig set_property \
@@ -140,18 +140,18 @@ function configure_master_node() {
 
   # Restart resource manager
   systemctl restart hadoop-yarn-resourcemanager
-  systemctl status hadoop-yarn-resourcemanager  # Ensure it started successfully
+  systemctl status hadoop-yarn-resourcemanager # Ensure it started successfully
 
   # Enable timeline server
   systemctl enable hadoop-yarn-timelineserver
   systemctl restart hadoop-yarn-timelineserver
-  systemctl status hadoop-yarn-timelineserver  # Ensure it started successfully
+  systemctl status hadoop-yarn-timelineserver # Ensure it started successfully
 
   # Check hive-server2 status
-  if ( systemctl is-enabled --quiet hive-server2 ); then
+  if (systemctl is-enabled --quiet hive-server2); then
     # Restart hive server2 if it is enabled
     systemctl restart hive-server2
-    systemctl status hive-server2  # Ensure it started successfully
+    systemctl status hive-server2 # Ensure it started successfully
   else
     echo "Service hive-server2 is not enabled"
   fi
diff --git a/user-environment/user-environment.sh b/user-environment/user-environment.sh
index 33b3a6c6f..16c583970 100755
--- a/user-environment/user-environment.sh
+++ b/user-environment/user-environment.sh
@@ -38,7 +38,7 @@ update-alternatives --set editor /usr/bin/vim.basic
 #apt-get install -y tmux sl
 
 ## The following script will get run as each user in their home directory.
-cat << 'EOF' > /tmp/customize_home_dir.sh
+cat <<'EOF' >/tmp/customize_home_dir.sh
 set -o errexit
 set -o nounset
 set -o xtrace
diff --git a/util/utils.sh b/util/utils.sh
index e27d80d40..77ba064d0 100755
--- a/util/utils.sh
+++ b/util/utils.sh
@@ -1,21 +1,20 @@
 #!/usr/bin/env bash
 
-function_exists () {
-    declare -f -F $1 > /dev/null
-    return $?
+function_exists() {
+  declare -f -F $1 >/dev/null
+  return $?
 }
 
-throw () {
-    echo "$*" >&2
-    echo
-    function_exists usage && usage
-    exit 1
+throw() {
+  echo "$*" >&2
+  echo
+  function_exists usage && usage
+  exit 1
 }
 
-get_metadata_property () {
-    [[ -z $1 ]] && throw "missing function param for DATAPROC_CLUSTER_NAME" || DATAPROC_CLUSTER_NAME=$1
-    [[ -z $2 ]] && throw "missing function param for METADATA_KEY" || METADATA_KEY=$2
-    # Get $DATAPROC_CLUSTER_NAME metadata value for key $METADATA_KEY...
-    gcloud dataproc clusters describe $DATAPROC_CLUSTER_NAME | python -c "import sys,yaml; cluster = yaml.load(sys.stdin); print(cluster['config']['gceClusterConfig']['metadata']['$METADATA_KEY'])"
+get_metadata_property() {
+  [[ -z $1 ]] && throw "missing function param for DATAPROC_CLUSTER_NAME" || DATAPROC_CLUSTER_NAME=$1
+  [[ -z $2 ]] && throw "missing function param for METADATA_KEY" || METADATA_KEY=$2
+  # Get $DATAPROC_CLUSTER_NAME metadata value for key $METADATA_KEY...
+  gcloud dataproc clusters describe $DATAPROC_CLUSTER_NAME | python -c "import sys,yaml; cluster = yaml.load(sys.stdin); print(cluster['config']['gceClusterConfig']['metadata']['$METADATA_KEY'])"
 }
-
diff --git a/zeppelin/zeppelin.sh b/zeppelin/zeppelin.sh
index 2b14c6c10..8ae1d1a55 100755
--- a/zeppelin/zeppelin.sh
+++ b/zeppelin/zeppelin.sh
@@ -19,7 +19,6 @@
 
 set -euxo pipefail
 
-
 readonly ROLE="$(/usr/share/google/get_metadata_value attributes/dataproc-role)"
 readonly INTERPRETER_FILE='/etc/zeppelin/conf/interpreter.json'
 readonly INIT_SCRIPT='/usr/lib/systemd/system/zeppelin-notebook.service'
@@ -49,7 +48,7 @@ function err() {
   return 1
 }
 
-function install_zeppelin(){
+function install_zeppelin() {
   # Install zeppelin. Don't mind if it fails to start the first time.
   retry_apt_command "apt-get install -t $(lsb_release -sc)-backports -y zeppelin" || dpkg -l zeppelin
   if [ $? != 0 ]; then
@@ -68,8 +67,8 @@ function install_zeppelin(){
   fi
 }
 
-function configure_zeppelin(){
-  local zeppelin_version;
+function configure_zeppelin() {
+  local zeppelin_version
   zeppelin_version="$(dpkg-query --showformat='${Version}' --show zeppelin)"
 
   # Only use sed to modify interpreter.json prior to Zeppelin 0.8.0.
@@ -101,7 +100,7 @@ function configure_zeppelin(){
     sed -i '/spark\.executor\.memory/d' "${INTERPRETER_FILE}"
 
     # Set BigQuery project ID if present.
-    local project_id;
+    local project_id
     project_id="$(/usr/share/google/get_metadata_value ../project/project-id)"
     sed -i "s/\(\"zeppelin.bigquery.project_id\"\)[^,}]*/\1: \"${project_id}\"/" \
       "${INTERPRETER_FILE}"
@@ -110,11 +109,11 @@ function configure_zeppelin(){
   # Link in hive configuration.
   ln -s /etc/hive/conf/hive-site.xml /etc/zeppelin/conf
 
-  local zeppelin_port;
+  local zeppelin_port
   zeppelin_port="$(/usr/share/google/get_metadata_value attributes/zeppelin-port || true)"
   if [[ -n "${zeppelin_port}" ]]; then
     echo "export ZEPPELIN_PORT=${zeppelin_port}" \
-      >> /etc/zeppelin/conf/zeppelin-env.sh
+      >>/etc/zeppelin/conf/zeppelin-env.sh
   fi
 
   # Install matplotlib. Note that this will work in Zeppelin, but not
diff --git a/zookeeper/zookeeper.sh b/zookeeper/zookeeper.sh
index c2b2092d5..503a7c1eb 100755
--- a/zookeeper/zookeeper.sh
+++ b/zookeeper/zookeeper.sh
@@ -36,7 +36,7 @@ function install_apt_get() {
 }
 
 function write_config() {
-  cat >> /etc/zookeeper/conf/zoo.cfg <<EOF
+  cat >>/etc/zookeeper/conf/zoo.cfg <<EOF
 
 # Properties from Zookeeper init action.
 tickTime=2000
@@ -48,7 +48,7 @@ server.0=${CLUSTER_NAME}-m:2888:3888
 EOF
 
   if [[ ${WORKER_COUNT} -gt 0 ]]; then
-    cat >> /etc/zookeeper/conf/zoo.cfg <<EOF
+    cat >>/etc/zookeeper/conf/zoo.cfg <<EOF
 server.1=${CLUSTER_NAME}-w-0:2888:3888
 server.2=${CLUSTER_NAME}-w-1:2888:3888
 EOF
@@ -69,12 +69,12 @@ fi
 
 # Configure ZooKeeper node ID, master has ID 0, workers start from 1.
 if [[ "${ROLE}" == 'Worker' ]]; then
-  NODE_NUMBER=$((`hostname | sed 's/.*-w-\([0-9]\)*.*/\1/g'`+1))
+  NODE_NUMBER=$(($(hostname | sed 's/.*-w-\([0-9]\)*.*/\1/g') + 1))
 else
   NODE_NUMBER=0
 fi
 
-if (( $NODE_NUMBER > 2 )); then
+if (($NODE_NUMBER > 2)); then
   write_config
   echo "Skip running ZooKeeper on this node."
   exit 0
@@ -86,7 +86,7 @@ install_apt_get zookeeper-server
 
 # Write ZooKeeper node ID.
 mkdir -p /var/lib/zookeeper
-echo ${NODE_NUMBER} >| /var/lib/zookeeper/myid
+echo ${NODE_NUMBER} >|/var/lib/zookeeper/myid
 
 # Write ZooKeeper configuration file
 write_config