diff --git a/.gitignore b/.gitignore
index b46da6e..3680446 100644
--- a/.gitignore
+++ b/.gitignore
@@ -5,3 +5,5 @@ build/
 dist/
 dsub_libs/
 dsub.egg-info/
+.ipynb_checkpoints/
+.DS_Store
diff --git a/README.md b/README.md
index 4df7b46..6eff054 100644
--- a/README.md
+++ b/README.md
@@ -56,7 +56,7 @@ While not used directly by `dsub` for the `google-batch` provider, you are likel
 Cloud SDK](https://cloud.google.com/sdk/).
 
 If you will be using the `local` provider for faster job development,
-you *will* need to install the Google Cloud SDK, which uses `gsutil` to ensure
+you *will* need to install the Google Cloud SDK, which uses `gcloud storage` to ensure
 file operation semantics consistent with the Google `dsub` providers.
 
 1. [Install the Google Cloud SDK](https://cloud.google.com/sdk/)
@@ -182,10 +182,10 @@ The steps for getting started differ slightly as indicated in the steps below:
 
     The dsub logs and output files will be written to a bucket. Create a
     bucket using the [storage browser](https://console.cloud.google.com/storage/browser?project=)
-    or run the command-line utility [gsutil](https://cloud.google.com/storage/docs/gsutil),
+    or run the command-line utility [gcloud storage](https://cloud.google.com/sdk/gcloud/reference/storage),
     included in the Cloud SDK.
 
-        gsutil mb gs://my-bucket
+        gcloud storage buckets create gs://my-bucket
 
     Change `my-bucket` to a unique name that follows the
     [bucket-naming conventions](https://cloud.google.com/storage/docs/bucket-naming).
@@ -215,7 +215,7 @@ The steps for getting started differ slightly as indicated in the steps below:
 
 1. View the output file.
 
-        gsutil cat gs://my-bucket/output/out.txt
+        gcloud storage cat gs://my-bucket/output/out.txt
 
 ## Backend providers
 
@@ -351,9 +351,8 @@ by:
 
 To upload the files to Google Cloud Storage, you can use the
 [storage browser](https://console.cloud.google.com/storage/browser?project=) or
-[gsutil](https://cloud.google.com/storage/docs/gsutil). You can also run on data
-that’s public or shared with your service account, an email address that you
-can find in the [Google Cloud Console](https://console.cloud.google.com).
+[gcloud storage](https://cloud.google.com/sdk/gcloud/reference/storage). 
+You can also run on data that’s public or shared with your service account, an email address that you can find in the [Google Cloud Console](https://console.cloud.google.com).
 
 #### Files
 
@@ -728,7 +727,9 @@ of the service account will be `sa-name@project-id.iam.gserviceaccount.com`.
 
 2. Grant IAM access on buckets, etc. to the service account.
 
-        gsutil iam ch serviceAccount:sa-name@project-id.iam.gserviceaccount.com:roles/storage.objectAdmin gs://bucket-name
+        gcloud storage buckets add-iam-policy-binding gs://bucket-name
+            --member=serviceAccount:sa-name@project-id.iam.gserviceaccount.com
+            --role=roles/storage.objectAdmin
 
 3. Update your `dsub` command to include `--service-account`
 
diff --git a/docs/code.md b/docs/code.md
index 2e465ec..80c1d29 100644
--- a/docs/code.md
+++ b/docs/code.md
@@ -187,7 +187,7 @@ To run the driver script, first copy `script1.sh` and `script2.sh` to
 cloud storage:
 
 ```
-gsutil cp my-code/script1.sh my-code/script2.sh gs://MY-BUCKET/my-code/
+gcloud storage cp my-code/script1.sh my-code/script2.sh gs://MY-BUCKET/my-code/
 ```
 
 Then launch a dsub job:
@@ -205,7 +205,7 @@ Extending the previous example, you could copy `script1.sh` and `script2.sh`
 to cloud storage with:
 
 ```
-gsutil rsync -r my-code gs://MY-BUCKET/my-code/
+gcloud storage rsync -r my-code gs://MY-BUCKET/my-code/
 ```
 
 and then launch a `dsub` job with:
diff --git a/docs/providers/README.md b/docs/providers/README.md
index c482dc8..131caa7 100644
--- a/docs/providers/README.md
+++ b/docs/providers/README.md
@@ -131,8 +131,8 @@ copying output files.
 
 The copying of files is performed in the host environment, not inside the
 Docker container. This means that for copying to/from Google Cloud Storage,
-the host environment requires a copy of
-[gsutil](https://cloud.google.com/storage/docs/gsutil) to be installed.
+the host environment requires the
+[Google Cloud SDK](https://cloud.google.com/sdk/docs/install) to be installed.
 
 #### Container runtime environment
 
diff --git a/dsub/_dsub_version.py b/dsub/_dsub_version.py
index 050f045..7b12e0d 100644
--- a/dsub/_dsub_version.py
+++ b/dsub/_dsub_version.py
@@ -27,5 +27,5 @@
 """
 
 
-DSUB_VERSION = '0.5.2'
+DSUB_VERSION = '0.5.3.dev0'
 
diff --git a/dsub/lib/param_util.py b/dsub/lib/param_util.py
index c5b9dba..ccddf48 100644
--- a/dsub/lib/param_util.py
+++ b/dsub/lib/param_util.py
@@ -827,23 +827,23 @@ def directory_fmt(directory):
 
   Multiple files copy, works as intended in all cases:
       $ touch a.txt b.txt
-      $ gsutil cp ./*.txt gs://mybucket/text_dest
-      $ gsutil ls gs://mybucket/text_dest/
+      $ gcloud storage cp ./*.txt gs://mybucket/text_dest
+      $ gcloud storage ls gs://mybucket/text_dest/
             0  2017-07-19T21:44:36Z  gs://mybucket/text_dest/a.txt
             0  2017-07-19T21:44:36Z  gs://mybucket/text_dest/b.txt
       TOTAL: 2 objects, 0 bytes (0 B)
 
   Single file copy fails to copy into a directory:
       $ touch 1.bam
-      $ gsutil cp ./*.bam gs://mybucket/bad_dest
-      $ gsutil ls gs://mybucket/bad_dest
+      $ gcloud storage cp ./*.bam gs://mybucket/bad_dest
+      $ gcloud storage ls gs://mybucket/bad_dest
                0  2017-07-19T21:46:16Z  gs://mybucket/bad_dest
       TOTAL: 1 objects, 0 bytes (0 B)
 
   Adding a trailing forward slash fixes this:
       $ touch my.sam
-      $ gsutil cp ./*.sam gs://mybucket/good_folder
-      $ gsutil ls gs://mybucket/good_folder
+      $ gcloud storage cp ./*.sam gs://mybucket/good_folder
+      $ gcloud storage ls gs://mybucket/good_folder
                0  2017-07-19T21:46:16Z  gs://mybucket/good_folder/my.sam
       TOTAL: 1 objects, 0 bytes (0 B)
 
diff --git a/dsub/lib/providers_util.py b/dsub/lib/providers_util.py
index 5700ea5..4f17669 100644
--- a/dsub/lib/providers_util.py
+++ b/dsub/lib/providers_util.py
@@ -20,7 +20,7 @@
 from .._dsub_version import DSUB_VERSION
 
 _LOCALIZE_COMMAND_MAP = {
-    job_model.P_GCS: 'gsutil -m rsync -r',
+    job_model.P_GCS: 'gcloud storage rsync -r',
     job_model.P_LOCAL: 'rsync -r',
 }
 
diff --git a/dsub/providers/google_batch.py b/dsub/providers/google_batch.py
index 73a35d4..2938730 100644
--- a/dsub/providers/google_batch.py
+++ b/dsub/providers/google_batch.py
@@ -145,11 +145,11 @@ def copy_log_to_staging(glob_str: str, staging_path: str, filter_str: str = None
       "${{LOGGING_DIR}}/stderr.txt" \
       "{user_action}"
 
-  gsutil_cp "${{LOGGING_DIR}}/stdout.txt" "${{STDOUT_PATH}}" "text/plain" "${{USER_PROJECT}}" &
+  gcloud_cp "${{LOGGING_DIR}}/stdout.txt" "${{STDOUT_PATH}}" "text/plain" "${{USER_PROJECT}}" &
   STDOUT_PID=$!
-  gsutil_cp "${{LOGGING_DIR}}/stderr.txt" "${{STDERR_PATH}}" "text/plain"  "${{USER_PROJECT}}" &
+  gcloud_cp "${{LOGGING_DIR}}/stderr.txt" "${{STDERR_PATH}}" "text/plain"  "${{USER_PROJECT}}" &
   STDERR_PID=$!
-  gsutil_cp "${{LOGGING_DIR}}/log.txt" "${{LOGGING_PATH}}" "text/plain" "${{USER_PROJECT}}" &
+  gcloud_cp "${{LOGGING_DIR}}/log.txt" "${{LOGGING_PATH}}" "text/plain" "${{USER_PROJECT}}" &
   LOG_PID=$!
 
   wait "${{STDOUT_PID}}"
@@ -168,7 +168,7 @@ def copy_log_to_staging(glob_str: str, staging_path: str, filter_str: str = None
   touch "${{LOGGING_DIR}}/.stop_logging"
 
   {log_msg_fn}
-  {gsutil_cp_fn}
+  {gcloud_cp_fn}
 
   {log_cp}
 """)
@@ -182,14 +182,14 @@ def copy_log_to_staging(glob_str: str, staging_path: str, filter_str: str = None
   readonly LOGGING_DIR="{logging_dir}"
 
   {log_msg_fn}
-  {gsutil_cp_fn}
+  {gcloud_cp_fn}
 
   # Make sure the logging work directory exists
   mkdir -p "${{LOGGING_DIR}}"
 
   # Prep the log filter script
   echo "${{{log_filter_var}}}" \
-    | python -c '{python_decode_script}' \
+    | python3 -c '{python_decode_script}' \
     > "{log_filter_script_path}"
   chmod a+x "{log_filter_script_path}"
 
@@ -609,7 +609,7 @@ def _create_batch_request(
 
     continuous_logging_cmd = _CONTINUOUS_LOGGING_CMD.format(
         log_msg_fn=google_utils.LOG_MSG_FN,
-        gsutil_cp_fn=google_utils.GSUTIL_CP_FN,
+        gcloud_cp_fn=google_utils.GCLOUD_CP_FN,
         log_filter_var=_LOG_FILTER_VAR,
         log_filter_script_path=_LOG_FILTER_SCRIPT_PATH,
         python_decode_script=google_utils.PYTHON_DECODE_SCRIPT,
@@ -623,7 +623,7 @@ def _create_batch_request(
 
     logging_cmd = _FINAL_LOGGING_CMD.format(
         log_msg_fn=google_utils.LOG_MSG_FN,
-        gsutil_cp_fn=google_utils.GSUTIL_CP_FN,
+        gcloud_cp_fn=google_utils.GCLOUD_CP_FN,
         log_filter_var=_LOG_FILTER_VAR,
         log_filter_script_path=_LOG_FILTER_SCRIPT_PATH,
         python_decode_script=google_utils.PYTHON_DECODE_SCRIPT,
@@ -729,8 +729,8 @@ def _create_batch_request(
                 '-c',
                 google_utils.LOCALIZATION_CMD.format(
                     log_msg_fn=google_utils.LOG_MSG_FN,
-                    recursive_cp_fn=google_utils.GSUTIL_RSYNC_FN,
-                    cp_fn=google_utils.GSUTIL_CP_FN,
+                    recursive_cp_fn=google_utils.GCLOUD_RSYNC_FN,
+                    cp_fn=google_utils.GCLOUD_CP_FN,
                     cp_loop=google_utils.LOCALIZATION_LOOP,
                 ),
             ],
@@ -778,8 +778,8 @@ def _create_batch_request(
                 '-c',
                 google_utils.LOCALIZATION_CMD.format(
                     log_msg_fn=google_utils.LOG_MSG_FN,
-                    recursive_cp_fn=google_utils.GSUTIL_RSYNC_FN,
-                    cp_fn=google_utils.GSUTIL_CP_FN,
+                    recursive_cp_fn=google_utils.GCLOUD_RSYNC_FN,
+                    cp_fn=google_utils.GCLOUD_CP_FN,
                     cp_loop=google_utils.DELOCALIZATION_LOOP,
                 ),
             ],
diff --git a/dsub/providers/google_utils.py b/dsub/providers/google_utils.py
index cea1196..a73a901 100644
--- a/dsub/providers/google_utils.py
+++ b/dsub/providers/google_utils.py
@@ -68,9 +68,9 @@ def make_runtime_dirs_command(script_dir: str, tmp_dir: str,
 # pylint: enable=g-complex-comprehension
 
 
-# Action steps that interact with GCS need gsutil and Python.
+# Action steps that interact with GCS need gcloud and Python.
 # Use the 'slim' variant of the cloud-sdk image as it is much smaller.
-CLOUD_SDK_IMAGE = 'gcr.io/google.com/cloudsdktool/cloud-sdk:294.0.0-slim'
+CLOUD_SDK_IMAGE = 'gcr.io/google.com/cloudsdktool/cloud-sdk:499.0.0-slim'
 
 # Name of the data disk
 DATA_DISK_NAME = 'datadisk'
@@ -94,10 +94,10 @@ def make_runtime_dirs_command(script_dir: str, tmp_dir: str,
   }
 """)
 
-# Define a bash function for "gsutil cp" to be used by the logging,
+# Define a bash function for "gcloud storage cp" to be used by the logging,
 # localization, and delocalization actions.
-GSUTIL_CP_FN = textwrap.dedent("""\
-  function gsutil_cp() {
+GCLOUD_CP_FN = textwrap.dedent("""\
+  function gcloud_cp() {
     local src="${1}"
     local dst="${2}"
     local content_type="${3}"
@@ -105,33 +105,33 @@ def make_runtime_dirs_command(script_dir: str, tmp_dir: str,
 
     local headers=""
     if [[ -n "${content_type}" ]]; then
-      headers="-h Content-Type:${content_type}"
+      headers="--content-type=${content_type}"
     fi
 
     local user_project_flag=""
     if [[ -n "${user_project_name}" ]]; then
-      user_project_flag="-u ${user_project_name}"
+      user_project_flag="--billing-project=${user_project_name}"
     fi
 
     local attempt
     for ((attempt = 0; attempt < 4; attempt++)); do
-      log_info "gsutil ${headers} ${user_project_flag} -mq cp \"${src}\" \"${dst}\""
-      if gsutil ${headers} ${user_project_flag} -mq cp "${src}" "${dst}"; then
+      log_info "gcloud storage cp ${headers} ${user_project_flag} \"${src}\" \"${dst}\""
+      if gcloud storage cp ${headers} ${user_project_flag} "${src}" "${dst}"; then
         return
       fi
       if (( attempt < 3 )); then
-        log_warning "Sleeping 10s before the next attempt of failed gsutil command"
-        log_warning "gsutil ${headers} ${user_project_flag} -mq cp \"${src}\" \"${dst}\""
+        log_warning "Sleeping 10s before the next attempt of failed gcloud command"
+        log_warning "gcloud storage cp ${headers} ${user_project_flag} \"${src}\" \"${dst}\""
         sleep 10s
       fi
     done
 
-    log_error "gsutil ${headers} ${user_project_flag} -mq cp \"${src}\" \"${dst}\""
+    log_error "gcloud storage cp ${headers} ${user_project_flag} \"${src}\" \"${dst}\""
     exit 1
   }
 """)
 
-LOG_CP_FN = GSUTIL_CP_FN + textwrap.dedent("""\
+LOG_CP_FN = GCLOUD_CP_FN + textwrap.dedent("""\
 
   function log_cp() {
     local src="${1}"
@@ -144,43 +144,43 @@ def make_runtime_dirs_command(script_dir: str, tmp_dir: str,
       return
     fi
 
-    # Copy the log files to a local temporary location so that our "gsutil cp" is never
+    # Copy the log files to a local temporary location so that our "gcloud storage cp" is never
     # executed on a file that is changing.
 
     local tmp_path="${tmp}/$(basename ${src})"
     cp "${src}" "${tmp_path}"
 
-    gsutil_cp "${tmp_path}" "${dst}" "text/plain" "${user_project_name}"
+    gcloud_cp "${tmp_path}" "${dst}" "text/plain" "${user_project_name}"
   }
 """)
 
-# Define a bash function for "gsutil rsync" to be used by the logging,
+# Define a bash function for "gcloud rsync" to be used by the logging,
 # localization, and delocalization actions.
-GSUTIL_RSYNC_FN = textwrap.dedent("""\
-  function gsutil_rsync() {
+GCLOUD_RSYNC_FN = textwrap.dedent("""\
+  function gcloud_rsync() {
     local src="${1}"
     local dst="${2}"
     local user_project_name="${3}"
 
     local user_project_flag=""
     if [[ -n "${user_project_name}" ]]; then
-      user_project_flag="-u ${user_project_name}"
+      user_project_flag="--billing-project=${user_project_name}"
     fi
 
     local attempt
     for ((attempt = 0; attempt < 4; attempt++)); do
-      log_info "gsutil ${user_project_flag} -mq rsync -r \"${src}\" \"${dst}\""
-      if gsutil ${user_project_flag} -mq rsync -r "${src}" "${dst}"; then
+      log_info "gcloud storage rsync -r ${user_project_flag} \"${src}\" \"${dst}\""
+      if gcloud storage rsync -r ${user_project_flag} "${src}" "${dst}"; then
         return
       fi
       if (( attempt < 3 )); then
-        log_warning "Sleeping 10s before the next attempt of failed gsutil command"
-        log_warning "gsutil ${user_project_flag} -mq rsync -r \"${src}\" \"${dst}\""
+        log_warning "Sleeping 10s before the next attempt of failed gcloud command"
+        log_warning "gcloud storage rsync -r ${user_project_flag} \"${src}\" \"${dst}\""
         sleep 10s
       fi
     done
 
-    log_error "gsutil ${user_project_flag} -mq rsync -r \"${src}\" \"${dst}\""
+    log_error "gcloud storage rsync -r ${user_project_flag} \"${src}\" \"${dst}\""
     exit 1
   }
 """)
@@ -198,9 +198,9 @@ def make_runtime_dirs_command(script_dir: str, tmp_dir: str,
 
     log_info "Localizing ${!INPUT_VAR}"
     if [[ "${!INPUT_RECURSIVE}" -eq "1" ]]; then
-      gsutil_rsync "${!INPUT_SRC}" "${!INPUT_DST}" "${USER_PROJECT}"
+      gcloud_rsync "${!INPUT_SRC}" "${!INPUT_DST}" "${USER_PROJECT}"
     else
-      gsutil_cp "${!INPUT_SRC}" "${!INPUT_DST}" "" "${USER_PROJECT}"
+      gcloud_cp "${!INPUT_SRC}" "${!INPUT_DST}" "" "${USER_PROJECT}"
     fi
   done
 """)
@@ -218,9 +218,9 @@ def make_runtime_dirs_command(script_dir: str, tmp_dir: str,
 
     log_info "Delocalizing ${!OUTPUT_VAR}"
     if [[ "${!OUTPUT_RECURSIVE}" -eq "1" ]]; then
-      gsutil_rsync "${!OUTPUT_SRC}" "${!OUTPUT_DST}" "${USER_PROJECT}"
+      gcloud_rsync "${!OUTPUT_SRC}" "${!OUTPUT_DST}" "${USER_PROJECT}"
     else
-      gsutil_cp "${!OUTPUT_SRC}" "${!OUTPUT_DST}" "" "${USER_PROJECT}"
+      gcloud_cp "${!OUTPUT_SRC}" "${!OUTPUT_DST}" "" "${USER_PROJECT}"
     fi
   done
 """)
@@ -274,7 +274,7 @@ def make_runtime_dirs_command(script_dir: str, tmp_dir: str,
   {mk_runtime_dirs}
 
   echo "${{{script_var}}}" \
-    | python -c '{python_decode_script}' \
+    | python3 -c '{python_decode_script}' \
     > "{script_path}"
   chmod a+x "{script_path}"
 
diff --git a/dsub/providers/google_v2_base.py b/dsub/providers/google_v2_base.py
index 4147653..781fc16 100644
--- a/dsub/providers/google_v2_base.py
+++ b/dsub/providers/google_v2_base.py
@@ -529,8 +529,8 @@ def _build_pipeline_request(self, task_view):
                 '-c',
                 google_utils.LOCALIZATION_CMD.format(
                     log_msg_fn=google_utils.LOG_MSG_FN,
-                    recursive_cp_fn=google_utils.GSUTIL_RSYNC_FN,
-                    cp_fn=google_utils.GSUTIL_CP_FN,
+                    recursive_cp_fn=google_utils.GCLOUD_RSYNC_FN,
+                    cp_fn=google_utils.GCLOUD_CP_FN,
                     cp_loop=google_utils.LOCALIZATION_LOOP,
                 ),
             ],
@@ -566,8 +566,8 @@ def _build_pipeline_request(self, task_view):
                 '-c',
                 google_utils.LOCALIZATION_CMD.format(
                     log_msg_fn=google_utils.LOG_MSG_FN,
-                    recursive_cp_fn=google_utils.GSUTIL_RSYNC_FN,
-                    cp_fn=google_utils.GSUTIL_CP_FN,
+                    recursive_cp_fn=google_utils.GCLOUD_RSYNC_FN,
+                    cp_fn=google_utils.GCLOUD_CP_FN,
                     cp_loop=google_utils.DELOCALIZATION_LOOP,
                 ),
             ],
diff --git a/dsub/providers/local.py b/dsub/providers/local.py
index b4b4814..3f3e71e 100644
--- a/dsub/providers/local.py
+++ b/dsub/providers/local.py
@@ -712,9 +712,9 @@ def _delocalize_logging_command(self, logging_path, user_project):
     elif logging_path.file_provider == job_model.P_GCS:
       mkdir_cmd = ''
       if user_project:
-        cp_cmd = 'gsutil -u {} -mq cp'.format(user_project)
+        cp_cmd = 'gcloud storage cp --billing-project={}'.format(user_project)
       else:
-        cp_cmd = 'gsutil -mq cp'
+        cp_cmd = 'gcloud storage cp'
     else:
       assert False
 
@@ -773,7 +773,7 @@ def _localize_inputs_recursive_command(self, task_dir, inputs):
     return '\n'.join(provider_commands)
 
   def _get_input_target_path(self, local_file_path):
-    """Returns a directory or file path to be the target for "gsutil cp".
+    """Returns a directory or file path to be the target for "gcloud storage cp".
 
     If the filename contains a wildcard, then the target path must
     be a directory in order to ensure consistency whether the source pattern
@@ -784,7 +784,7 @@ def _get_input_target_path(self, local_file_path):
       local_file_path: A full path terminating in a file or a file wildcard.
 
     Returns:
-      The path to use as the "gsutil cp" target.
+      The path to use as the "gcloud storage cp" target.
     """
 
     path, filename = os.path.split(local_file_path)
@@ -806,21 +806,25 @@ def _localize_inputs_command(self, task_dir, inputs, user_project):
 
       commands.append('mkdir -p "%s"' % os.path.dirname(local_file_path))
 
-      if i.file_provider in [job_model.P_LOCAL, job_model.P_GCS]:
-        # The semantics that we expect here are implemented consistently in
-        # "gsutil cp", and are a bit different than "cp" when it comes to
-        # wildcard handling, so use it for both local and GCS:
-        #
-        # - `cp path/* dest/` will error if "path" has subdirectories.
-        # - `cp "path/*" "dest/"` will fail (it expects wildcard expansion
-        #   to come from shell).
+      if i.file_provider == job_model.P_GCS:
+        # `cp path/* dest/` will error if "path" has subdirectories, and
+        # `cp "path/*" "dest/"` will fail (it expects wildcard expansion
+        # to come from shell), so use "gcloud storage cp" instead.
         if user_project:
-          command = 'gsutil -u %s -mq cp "%s" "%s"' % (
+          command = 'gcloud storage cp --billing-project=%s "%s" "%s"' % (
               user_project, source_file_path, dest_file_path)
         else:
-          command = 'gsutil -mq cp "%s" "%s"' % (source_file_path,
+          command = 'gcloud storage cp "%s" "%s"' % (source_file_path,
                                                  dest_file_path)
         commands.append(command)
+      elif i.file_provider == job_model.P_LOCAL:
+        # "gcloud storage cp" does not support local-to-local copies, so use
+        # rsync (without "-r", since this path is non-recursive) instead.
+        # It matches the wildcard semantics we need: like the old "gsutil
+        # cp", it silently skips subdirectories rather than erroring like
+        # plain "cp" does.
+        command = 'rsync "%s" "%s"' % (source_file_path, dest_file_path)
+        commands.append(command)
 
     return '\n'.join(commands)
 
@@ -865,13 +869,17 @@ def _delocalize_outputs_commands(self, task_dir, outputs, user_project):
       if o.file_provider == job_model.P_LOCAL:
         commands.append('mkdir -p "%s"' % dest_path)
 
-      # Use gsutil even for local files (explained in _localize_inputs_command).
-      if o.file_provider in [job_model.P_LOCAL, job_model.P_GCS]:
+      if o.file_provider == job_model.P_GCS:
         if user_project:
-          command = 'gsutil -u %s -mq cp "%s" "%s"' % (user_project, local_path,
+          command = 'gcloud storage cp --billing-project=%s "%s" "%s"' % (user_project, local_path,
                                                        dest_path)
         else:
-          command = 'gsutil -mq cp "%s" "%s"' % (local_path, dest_path)
+          command = 'gcloud storage cp "%s" "%s"' % (local_path, dest_path)
+        commands.append(command)
+      elif o.file_provider == job_model.P_LOCAL:
+        # "gcloud storage cp" does not support local-to-local copies (see
+        # _localize_inputs_command), so use rsync instead.
+        command = 'rsync "%s" "%s"' % (local_path, dest_path)
         commands.append(command)
 
     return '\n'.join(commands)
diff --git a/examples/custom_scripts/README.md b/examples/custom_scripts/README.md
index fd1d965..c471591 100644
--- a/examples/custom_scripts/README.md
+++ b/examples/custom_scripts/README.md
@@ -55,7 +55,7 @@ To run a Bash script to decompress the VCF file, type:
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
   --regions "us-central1" \
   --logging "gs://MY-BUCKET/get_vcf_sample_ids.sh/logging" \
@@ -87,7 +87,7 @@ Because the `--wait` flag was set, `dsub` will block until the job completes.
 To list the output, use the command:
 
 ```
-gsutil ls gs://MY-BUCKET/get_vcf_sample_ids.sh/output
+gcloud storage ls gs://MY-BUCKET/get_vcf_sample_ids.sh/output
 ```
 
 Output should look like:
@@ -99,7 +99,7 @@ gs://MY-BUCKET/get_vcf_sample_ids.sh/output/sample_ids.txt
 To see the first few lines of the sample IDs file, run:
 
 ```
-gsutil cat gs://MY-BUCKET/get_vcf_sample_ids.sh/output/sample_ids.txt | head -n 5
+gcloud storage cat gs://MY-BUCKET/get_vcf_sample_ids.sh/output/sample_ids.txt | head -n 5
 ```
 
 Output should look like:
@@ -134,7 +134,7 @@ To run a Python script to decompress the VCF file, type:
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
   --regions "us-central1" \
   --logging "gs://MY-BUCKET/get_vcf_sample_ids.py/logging" \
@@ -166,7 +166,7 @@ Because the `--wait` flag was set, `dsub` will block until the job completes.
 To list the output, use the command:
 
 ```
-gsutil ls gs://MY-BUCKET/get_vcf_sample_ids.py/output
+gcloud storage ls gs://MY-BUCKET/get_vcf_sample_ids.py/output
 ```
 
 Output should look like:
@@ -178,7 +178,7 @@ gs://MY-BUCKET/get_vcf_sample_ids.py/output/sample_ids.txt
 To see the first few lines of the sample IDs file, run:
 
 ```
-gsutil cat gs://MY-BUCKET/get_vcf_sample_ids.py/output/sample_ids.txt | head -n 5
+gcloud storage cat gs://MY-BUCKET/get_vcf_sample_ids.py/output/sample_ids.txt | head -n 5
 ```
 
 Output should look like:
@@ -221,7 +221,7 @@ Run either of the following commands:
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
   --regions "us-central1" \
   --logging "gs://MY-BUCKET/get_vcf_sample_ids/logging" \
@@ -234,7 +234,7 @@ dsub \
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
   --regions "us-central1" \
   --logging "gs://MY-BUCKET/get_vcf_sample_ids/logging" \
@@ -265,7 +265,7 @@ When all tasks for the job have completed, `dsub` will exit.
 To list the output objects, use the command:
 
 ```
-gsutil ls gs://MY-BUCKET/get_vcf_sample_ids/output
+gcloud storage ls gs://MY-BUCKET/get_vcf_sample_ids/output
 ```
 
 Output should look like:
diff --git a/examples/custom_scripts/submit_list.sh b/examples/custom_scripts/submit_list.sh
index 4168dc3..754a05b 100755
--- a/examples/custom_scripts/submit_list.sh
+++ b/examples/custom_scripts/submit_list.sh
@@ -57,7 +57,7 @@ echo
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
   --regions "us-central1" \
   --logging "${LOGGING}" \
diff --git a/examples/custom_scripts/submit_one.sh b/examples/custom_scripts/submit_one.sh
index 2886bb9..f6c15a3 100755
--- a/examples/custom_scripts/submit_one.sh
+++ b/examples/custom_scripts/submit_one.sh
@@ -60,7 +60,7 @@ echo
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
   --regions "us-central1" \
   --logging "${LOGGING}" \
@@ -73,5 +73,5 @@ dsub \
 
 # Check output
 echo "Check the head of the output file:"
-2>&1 gsutil cat "${OUTPUT_FILE}" | head
+2>&1 gcloud storage cat "${OUTPUT_FILE}" | head
 
diff --git a/examples/decompress/README.md b/examples/decompress/README.md
index 7fa2f81..f8750ef 100644
--- a/examples/decompress/README.md
+++ b/examples/decompress/README.md
@@ -33,9 +33,9 @@ To run a command to decompress the VCF file, type:
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "gs://MY-BUCKET/decompress_one/logging" \
   --disk-size 200 \
   --image ubuntu:14.04 \
@@ -65,7 +65,7 @@ Because the `--wait` flag was set, `dsub` will block until the job completes.
 To list the output, use the command:
 
 ```
-gsutil ls gs://MY-BUCKET/decompress_one/output
+gcloud storage ls gs://MY-BUCKET/decompress_one/output
 ```
 
 Output should look like:
@@ -77,7 +77,7 @@ gs://MY-BUCKET/decompress_one/output/ALL.ChrY.Cornell.20130502.SNPs.Genotypes.vc
 To see the first few lines of the decompressed file, run:
 
 ```
-gsutil cat gs://MY-BUCKET/decompress_one/output/*.vcf | head -n 5
+gcloud storage cat gs://MY-BUCKET/decompress_one/output/*.vcf | head -n 5
 ```
 
 Output should look like:
@@ -122,9 +122,9 @@ output file name.
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "gs://MY-BUCKET/decompress_list/logging" \
   --disk-size 200 \
   --image ubuntu:14.04 \
@@ -153,7 +153,7 @@ when all tasks for the job have completed, `dsub` will exit.
 To list the output objects, use the command:
 
 ```
-gsutil ls gs://MY-BUCKET/decompress_list/output
+gcloud storage ls gs://MY-BUCKET/decompress_list/output
 ```
 
 Output should look like:
diff --git a/examples/decompress/submit_list.sh b/examples/decompress/submit_list.sh
index 334d6d9..4a880c1 100755
--- a/examples/decompress/submit_list.sh
+++ b/examples/decompress/submit_list.sh
@@ -34,9 +34,9 @@ readonly SCRIPT_DIR="$(dirname "${0}")"
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "${MY_BUCKET}/decompress_list/logging/" \
   --disk-size 200 \
   --image ubuntu:14.04 \
diff --git a/examples/decompress/submit_one.sh b/examples/decompress/submit_one.sh
index 0870e61..ff85417 100755
--- a/examples/decompress/submit_one.sh
+++ b/examples/decompress/submit_one.sh
@@ -35,9 +35,9 @@ readonly SCRIPT_DIR="$(dirname "${0}")"
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "${MY_BUCKET_PATH}"/logging/ \
   --disk-size 200 \
   --image ubuntu:14.04 \
diff --git a/examples/fastqc/README.md b/examples/fastqc/README.md
index ef02fb7..a50e227 100644
--- a/examples/fastqc/README.md
+++ b/examples/fastqc/README.md
@@ -77,9 +77,9 @@ To run FastQC on the BAM file, type:
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "gs://MY-BUCKET/fastqc/submit_one/logging" \
   --disk-size 200 \
   --name "fastqc" \
@@ -99,9 +99,9 @@ You should see output like:
 Job: fastqc--<userid>--170619-105212-67
 Launched job-id: fastqc--<userid>--170619-105212-67
 To check the status, run:
-  dstat --provider google-cls-v2 --project MY-PROJECT --jobs fastqc--<userid>--170619-105212-67 --status '*'
+  dstat --provider google-batch --project MY-PROJECT --jobs fastqc--<userid>--170619-105212-67 --status '*'
 To cancel the job, run:
-  ddel --provider google-cls-v2 --project MY-PROJECT --jobs fastqc--<userid>--170619-105212-67
+  ddel --provider google-batch --project MY-PROJECT --jobs fastqc--<userid>--170619-105212-67
 Waiting for job to complete...
 Waiting for: fastqc--<userid>--170619-105212-67.
 ```
@@ -113,7 +113,7 @@ Because the `--wait` flag was set, `dsub` will block until the job completes.
 To list the output, use the command:
 
 ```
-gsutil ls -l gs://MY-BUCKET/fastqc/submit_one/output
+gcloud storage ls -l gs://MY-BUCKET/fastqc/submit_one/output
 ```
 
 Output should look like:
@@ -156,9 +156,9 @@ output file name.
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "gs://MY-BUCKET/samtools/submit_list/logging" \
   --disk-size 200 \
   --name "fastqc" \
@@ -175,9 +175,9 @@ Job: fastqc--<userid>--170522-154943-70
 Launched job-id: fastqc--<userid>--170522-154943-70
 3 task(s)
 To check the status, run:
-  dstat --provider google-cls-v2 --project MY-PROJECT --jobs fastqc--<userid>--170522-154943-70 --status '*'
+  dstat --provider google-batch --project MY-PROJECT --jobs fastqc--<userid>--170522-154943-70 --status '*'
 To cancel the job, run:
-  ddel --provider google-cls-v2 --project MY-PROJECT --jobs fastqc--<userid>--170522-154943-70
+  ddel --provider google-batch --project MY-PROJECT --jobs fastqc--<userid>--170522-154943-70
 Waiting for job to complete...
 Waiting for: fastqc--<userid>--170522-154943-70.
 ```
@@ -189,7 +189,7 @@ when all tasks for the job have completed, `dsub` will exit.
 To list the output objects, use the command:
 
 ```
-gsutil ls -l gs://MY-BUCKET/fastqc/submit_list/output
+gcloud storage ls -l gs://MY-BUCKET/fastqc/submit_list/output
 ```
 
 Output should look like:
diff --git a/examples/fastqc/submit_list.sh b/examples/fastqc/submit_list.sh
index 2421cf3..8163aac 100755
--- a/examples/fastqc/submit_list.sh
+++ b/examples/fastqc/submit_list.sh
@@ -40,9 +40,9 @@ gcloud builds submit "${SCRIPT_DIR}" \
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "${OUTPUT_ROOT}/logging/" \
   --disk-size 200 \
   --name "fastqc" \
diff --git a/examples/fastqc/submit_one.sh b/examples/fastqc/submit_one.sh
index bb848fb..9b4141e 100755
--- a/examples/fastqc/submit_one.sh
+++ b/examples/fastqc/submit_one.sh
@@ -41,9 +41,9 @@ gcloud builds submit "${SCRIPT_DIR}" \
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "${OUTPUT_ROOT}/logging" \
   --disk-size 200 \
   --name "fastqc" \
diff --git a/examples/samtools/README.md b/examples/samtools/README.md
index 5f968ad..93e30de 100644
--- a/examples/samtools/README.md
+++ b/examples/samtools/README.md
@@ -38,9 +38,9 @@ To run a command to index the BAM file, type:
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "gs://MY-BUCKET/samtools/submit_one/logging" \
   --disk-size 200 \
   --name "samtools index" \
@@ -63,9 +63,9 @@ You should see output like:
 Job: samtools-i--<userid>--170522-153810-14
 Launched job-id: samtools-i--<userid>--170522-153810-14
 To check the status, run:
-  dstat --provider google-cls-v2 --project MY-PROJECT --jobs samtools-i--<userid>--170522-153810-14 --status '*'
+  dstat --provider google-batch --project MY-PROJECT --jobs samtools-i--<userid>--170522-153810-14 --status '*'
 To cancel the job, run:
-  ddel --provider google-cls-v2 --project MY-PROJECT --jobs samtools-i--<userid>--170522-153810-14
+  ddel --provider google-batch --project MY-PROJECT --jobs samtools-i--<userid>--170522-153810-14
 Waiting for job to complete...
 Waiting for: samtools-i--<userid>--170522-153810-14.
 ```
@@ -77,7 +77,7 @@ Because the `--wait` flag was set, `dsub` will block until the job completes.
 To list the output, use the command:
 
 ```
-gsutil ls -l gs://MY-BUCKET/samtools/submit_one/output
+gcloud storage ls -l gs://MY-BUCKET/samtools/submit_one/output
 ```
 
 Output should look like:
@@ -119,9 +119,9 @@ output file name.
 
 ```
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project MY-PROJECT \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "gs://MY-BUCKET/samtools/submit_list/logging" \
   --disk-size 200 \
   --name "samtools index" \
@@ -141,9 +141,9 @@ Job: samtools-i--<userid>--170522-154943-70
 Launched job-id: samtools-i--<userid>--170522-154943-70
 3 task(s)
 To check the status, run:
-  dstat --provider google-cls-v2 --project MY-PROJECT --jobs samtools-i--<userid>--170522-154943-70 --status '*'
+  dstat --provider google-batch --project MY-PROJECT --jobs samtools-i--<userid>--170522-154943-70 --status '*'
 To cancel the job, run:
-  ddel --provider google-cls-v2 --project MY-PROJECT --jobs samtools-i--<userid>--170522-154943-70
+  ddel --provider google-batch --project MY-PROJECT --jobs samtools-i--<userid>--170522-154943-70
 Waiting for job to complete...
 Waiting for: samtools-i--<userid>--170522-154943-70.
 ```
@@ -155,7 +155,7 @@ when all tasks for the job have completed, `dsub` will exit.
 To list the output objects, use the command:
 
 ```
-gsutil ls -l gs://MY-BUCKET/samtools/submit_list/output
+gcloud storage ls -l gs://MY-BUCKET/samtools/submit_list/output
 ```
 
 Output should look like:
diff --git a/examples/samtools/submit_list.sh b/examples/samtools/submit_list.sh
index 82ca243..77efdaf 100755
--- a/examples/samtools/submit_list.sh
+++ b/examples/samtools/submit_list.sh
@@ -36,9 +36,9 @@ readonly SCRIPT_DIR="$(dirname "${0}")"
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "${OUTPUT_ROOT}/logging/" \
   --disk-size 200 \
   --name "samtools index" \
diff --git a/examples/samtools/submit_one.sh b/examples/samtools/submit_one.sh
index c3e46ed..4a322a8 100755
--- a/examples/samtools/submit_one.sh
+++ b/examples/samtools/submit_one.sh
@@ -37,9 +37,9 @@ readonly SCRIPT_DIR="$(dirname "${0}")"
 
 # Launch the task
 dsub \
-  --provider google-cls-v2 \
+  --provider google-batch \
   --project "${MY_PROJECT}" \
-  --zones "us-central1-*" \
+  --regions us-central1 \
   --logging "${OUTPUT_ROOT}"/logging \
   --disk-size 200 \
   --name "samtools index" \
diff --git a/examples/split_process/README.md b/examples/split_process/README.md
index ad71d35..32fd1a5 100644
--- a/examples/split_process/README.md
+++ b/examples/split_process/README.md
@@ -44,6 +44,6 @@ rm "${WORKSPACE}/temp/*"
 ```
 WORKSPACE=gs://mybucket/someprefix
 ./demo_split_process.sh input.txt "${WORKSPACE}"
-gsutil ls "${WORKSPACE}/output/"
-gsutil rm "${WORKSPACE}/temp/*"
+gcloud storage ls "${WORKSPACE}/output/"
+gcloud storage rm "${WORKSPACE}/temp/*"
 ```
diff --git a/examples/split_process/demo_split_process.sh b/examples/split_process/demo_split_process.sh
index e33a1af..012745d 100755
--- a/examples/split_process/demo_split_process.sh
+++ b/examples/split_process/demo_split_process.sh
@@ -11,10 +11,10 @@
 # example:
 #   WORKSPACE=gs://mybucket/someprefix
 #   ./demo_split_process.sh input.txt "${WORKSPACE}"
-#   gsutil ls "${WORKSPACE}/output/"
-#   gsutil rm "${WORKSPACE}/temp/*"
+#   gcloud storage ls "${WORKSPACE}/output/"
+#   gcloud storage rm "${WORKSPACE}/temp/*"
 #
-# You need dsub, docker, and gsutil installed.
+# You need dsub, docker, and gcloud installed.
 # Change WORKSPACE to point to a bucket you have write permission to.
 #
 # Since this uses the local provider, you can set WORKSPACE to a local path,
diff --git a/test/integration/e2e_accelerator.google-batch.sh b/test/integration/e2e_accelerator.google-batch.sh
index 9a9a1c6..4f6ae4b 100755
--- a/test/integration/e2e_accelerator.google-batch.sh
+++ b/test/integration/e2e_accelerator.google-batch.sh
@@ -87,7 +87,7 @@ echo
 echo "Checking GPU detection output..."
 
 # Check that GPU was detected and accessible
-RESULT="$(gsutil cat "${STDOUT_LOG}")"
+RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
 
 # Validate GPU hardware was detected
 if ! echo "${RESULT}" | grep -qi "Tesla T4"; then
diff --git a/test/integration/e2e_accelerator.google-cls-v2.sh b/test/integration/e2e_accelerator.google-cls-v2.sh
index 0960efc..3c13a56 100755
--- a/test/integration/e2e_accelerator.google-cls-v2.sh
+++ b/test/integration/e2e_accelerator.google-cls-v2.sh
@@ -45,7 +45,7 @@ echo
 echo "Checking output..."
 
 # Check the results
-RESULT="$(gsutil cat "${STDOUT_LOG}")"
+RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
 if ! echo "${RESULT}" | grep -qi "GPU Memory"; then
   1>&2 echo "GPU Memory not found in the dsub output!"
   1>&2 echo "${RESULT}"
diff --git a/test/integration/e2e_accelerator_vpc_sc.google-batch.sh b/test/integration/e2e_accelerator_vpc_sc.google-batch.sh
index b798992..496c6f7 100755
--- a/test/integration/e2e_accelerator_vpc_sc.google-batch.sh
+++ b/test/integration/e2e_accelerator_vpc_sc.google-batch.sh
@@ -155,7 +155,7 @@ echo
 echo "Checking GPU detection output..."
 
 # Check that GPU was detected and accessible
-RESULT="$(gsutil cat "${STDOUT_LOG}")"
+RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
 
 # Validate GPU hardware was detected
 if ! echo "${RESULT}" | grep -qi "Tesla T4"; then
diff --git a/test/integration/e2e_after.py b/test/integration/e2e_after.py
index 719ede7..a5e09ef 100644
--- a/test/integration/e2e_after.py
+++ b/test/integration/e2e_after.py
@@ -56,7 +56,7 @@
 
 print('\nChecking output...')
 
-RESULT = test_util.gsutil_cat(TEST_FILE_PATH_2)
+RESULT = test_util.gcloud_cat(TEST_FILE_PATH_2)
 if 'hello world' not in RESULT:
   print('Output file does not match expected', file=sys.stderr)
   sys.exit(1)
diff --git a/test/integration/e2e_after.sh b/test/integration/e2e_after.sh
index 9b9c6b7..bb414d6 100755
--- a/test/integration/e2e_after.sh
+++ b/test/integration/e2e_after.sh
@@ -55,7 +55,7 @@ fi
 echo
 echo "Checking output..."
 
-readonly RESULT="$(gsutil cat "${TEST_FILE_PATH_2}")"
+readonly RESULT="$(gcloud storage cat "${TEST_FILE_PATH_2}")"
 if [[ "${RESULT}" != "hello world" ]]; then
   echo "Output file does not match expected"
   echo "Expected: hello world"
diff --git a/test/integration/e2e_block_external_network.google-cls-v2.sh b/test/integration/e2e_block_external_network.google-cls-v2.sh
index feb4889..157549a 100755
--- a/test/integration/e2e_block_external_network.google-cls-v2.sh
+++ b/test/integration/e2e_block_external_network.google-cls-v2.sh
@@ -31,9 +31,8 @@ echo "Launching pipeline..."
 
 set +o errexit
 
-# Run gsutil with Boto:num_retries=0 option. Otherwise, gsutil will retry up to
-# 24 times due to the network error
-# https://stackoverflow.com/questions/44459685/sql-server-agent-job-and-gsutil
+# Run gcloud storage with --no-user-output-enabled and max retries set to 0.
+# Otherwise, gcloud storage will retry due to the network error
 JOB_ID="$(run_dsub \
   --image 'gcr.io/google.com/cloudsdktool/cloud-sdk:327.0.0-slim' \
   --block-external-network \
@@ -54,9 +53,9 @@ readonly ATTEMPT_1_STDERR_LOG="$(dirname "${LOGGING}")/${TEST_NAME}.1-stderr.log
 readonly ATTEMPT_2_STDERR_LOG="$(dirname "${LOGGING}")/${TEST_NAME}.2-stderr.log"
 
 for STDERR_LOG_FILE in "${ATTEMPT_1_STDERR_LOG}" "${ATTEMPT_2_STDERR_LOG}" ; do
-  RESULT="$(gsutil cat "${STDERR_LOG_FILE}")"
+  RESULT="$(gcloud storage cat "${STDERR_LOG_FILE}")"
   if ! echo "${RESULT}" | grep -qi "Unable to find the server at storage.googleapis.com"; then
-    1>&2 echo "Network error from gsutil not found in the dsub stderr log!"
+    1>&2 echo "Network error from gcloud not found in the dsub stderr log!"
     1>&2 echo "${RESULT}"
     exit 1
   fi
diff --git a/test/integration/e2e_cleanup.local.sh b/test/integration/e2e_cleanup.local.sh
index 1677cf9..5064299 100755
--- a/test/integration/e2e_cleanup.local.sh
+++ b/test/integration/e2e_cleanup.local.sh
@@ -26,7 +26,7 @@ readonly SCRIPT_DIR="$(dirname "${0}")"
 source "${SCRIPT_DIR}/test_setup_e2e.sh"
 
 # Stage a test file.
-date | gsutil cp - "${INPUTS}/recursive/deep/today.txt"
+date | gcloud storage cp - "${INPUTS}/recursive/deep/today.txt"
 
 readonly TGT_1="${OUTPUTS}/testfile_1.txt"
 readonly TGT_2="${OUTPUTS}/testfile_2.txt"
@@ -82,7 +82,7 @@ JOB_ID=$(run_dsub \
 check_jobid "${JOB_ID}"
 
 for out in "${TGT_1}" "${TGT_2}" "${TGT_3}"; do
-  if ! gsutil ls "${out}" > /dev/null; then
+  if ! gcloud storage ls "${out}" > /dev/null; then
     echo "Missing output: ${out}"
     exit 1
   fi
diff --git a/test/integration/e2e_command_flag.sh b/test/integration/e2e_command_flag.sh
index da66abc..bc03140 100755
--- a/test/integration/e2e_command_flag.sh
+++ b/test/integration/e2e_command_flag.sh
@@ -57,7 +57,7 @@ VAR5=VAL5
 EOF
 )
 
-readonly RESULT="$(gsutil cat "${STDOUT_LOG}")"
+readonly RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
 if ! diff <(echo "${RESULT_EXPECTED}") <(echo "${RESULT}"); then
   echo "Output file does not match expected"
   exit 1
diff --git a/test/integration/e2e_env_list.py b/test/integration/e2e_env_list.py
index 3099526..fed80d3 100644
--- a/test/integration/e2e_env_list.py
+++ b/test/integration/e2e_env_list.py
@@ -73,7 +73,7 @@
 VAR5=VAL5
 """.lstrip()
 
-RESULT = test_util.gsutil_cat(test.STDOUT_LOG)
+RESULT = test_util.gcloud_cat(test.STDOUT_LOG)
 if not test_util.diff(RESULT_EXPECTED, RESULT):
   print('Output file does not match expected')
   sys.exit(1)
diff --git a/test/integration/e2e_env_tasks.sh b/test/integration/e2e_env_tasks.sh
index 512a287..c1c559e 100755
--- a/test/integration/e2e_env_tasks.sh
+++ b/test/integration/e2e_env_tasks.sh
@@ -101,7 +101,7 @@ for ((TASK_ID=1; TASK_ID <= NUM_TASKS; TASK_ID++)); do
     sed -e 's#^ *##'
   )"
 
-  RESULT="$(gsutil cat "${LOGGING}.${TASK_ID}-stdout.log")"
+  RESULT="$(gcloud storage cat "${LOGGING}.${TASK_ID}-stdout.log")"
   if ! diff <(echo "${RESULT_EXPECTED}") <(echo "${RESULT}"); then
     echo "Output file does not match expected"
     exit 1
diff --git a/test/integration/e2e_image.sh b/test/integration/e2e_image.sh
index 6f38c84..f085307 100755
--- a/test/integration/e2e_image.sh
+++ b/test/integration/e2e_image.sh
@@ -58,7 +58,7 @@ for image in ${IMAGE_ARRAY[@]}; do
   echo "Checking output..."
 
   # Check the results
-  RESULT="$(gsutil cat "${STDOUT_LOG}")"
+  RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
   if ! diff <(echo "${RESULT_EXPECTED}") <(echo "${RESULT}"); then
     echo "Output file does not match expected"
     exit 1
diff --git a/test/integration/e2e_input_wildcards.sh b/test/integration/e2e_input_wildcards.sh
index d7c8d04..638ea14 100755
--- a/test/integration/e2e_input_wildcards.sh
+++ b/test/integration/e2e_input_wildcards.sh
@@ -43,7 +43,7 @@ function exit_handler() {
   # Only cleanup on success
   if [[ "${code}" -eq 0 ]]; then
     rm -rf "${TEST_TMP}"
-    gsutil -mq rm "${INPUTS}/**"
+    gcloud storage rm "${INPUTS}/**"
   fi
 
   return "${code}"
@@ -64,7 +64,7 @@ for INPUT_DIR in "${INPUT_BASIC}" "${INPUT_WITH_SPACE}"; do
   done
 done
 
-gsutil -m rsync -r "${INPUT_ROOT}" "${INPUTS}/"
+gcloud storage rsync -r "${INPUT_ROOT}" "${INPUTS}/"
 
 echo "Launching pipeline..."
 
@@ -94,7 +94,7 @@ FILE_NAME=file.3.txt
 EOF
 )
 
-readonly RESULT="$(gsutil cat "${STDOUT_LOG}")"
+readonly RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
 if ! diff <(echo "${RESULT_EXPECTED}") <(echo "${RESULT}"); then
   echo "Output file does not match expected"
   exit 1
diff --git a/test/integration/e2e_io_auto.sh b/test/integration/e2e_io_auto.sh
index 35b6f94..813b841 100755
--- a/test/integration/e2e_io_auto.sh
+++ b/test/integration/e2e_io_auto.sh
@@ -64,7 +64,7 @@ readonly EXPECTED_FS_OUTPUT_ENTRIES=(
 )
 
 # Get the results- "env" and "find" output is bounded by "BEGIN" and "END"
-readonly RESULT=$(gsutil cat "${STDOUT_LOG}")
+readonly RESULT=$(gcloud storage cat "${STDOUT_LOG}")
 readonly ENV=$(echo "${RESULT}" | sed -n '/^BEGIN: env$/,/^END: env$/p')
 readonly FIND=$(echo "${RESULT}" | sed -n '/^BEGIN: find$/,/^END: find$/p')
 
diff --git a/test/integration/e2e_io_gcs_tasks.sh b/test/integration/e2e_io_gcs_tasks.sh
index 43948f6..a3a4515 100755
--- a/test/integration/e2e_io_gcs_tasks.sh
+++ b/test/integration/e2e_io_gcs_tasks.sh
@@ -44,11 +44,11 @@ io_tasks_setup::write_tasks_file
 
 # Copy the script to GCS to test loading the script remotely
 echo "Copying script to ${DSUB_PARAMS}"
-gsutil cp "${SCRIPT_DIR}/script_io_test.sh" "${DSUB_PARAMS}/"
+gcloud storage cp "${SCRIPT_DIR}/script_io_test.sh" "${DSUB_PARAMS}/"
 
 # Copy the TASKS_FILE to GCS to test loading the tasks file remotely
 echo "Copying tasks file to ${DSUB_PARAMS}"
-gsutil cp "${TASKS_FILE}" "${DSUB_PARAMS}/"
+gcloud storage cp "${TASKS_FILE}" "${DSUB_PARAMS}/"
 
 echo "Launching pipelines..."
 
@@ -62,4 +62,4 @@ io_tasks_setup::check_output
 io_tasks_setup::check_dstat "${JOB_ID}"
 
 # Clean up what we uploaded after the test is done.
-gsutil rm "${DSUB_PARAMS}"/**
+gcloud storage rm "${DSUB_PARAMS}"/**
diff --git a/test/integration/e2e_io_mount_dir.local.sh b/test/integration/e2e_io_mount_dir.local.sh
index 746a2d0..3fb8779 100755
--- a/test/integration/e2e_io_mount_dir.local.sh
+++ b/test/integration/e2e_io_mount_dir.local.sh
@@ -18,7 +18,7 @@ set -o errexit
 set -o nounset
 
 # This test verifies that mounting a local directory works. The test will copy
-# input files via gsutil to the local disk.
+# input files via gcloud to the local disk.
 #
 # The actual operation performed here is to download a BAM and compute
 # the md5, writing it to <filename>.bam.md5.
diff --git a/test/integration/e2e_io_recursive.sh b/test/integration/e2e_io_recursive.sh
index 3bca94c..9b6128b 100755
--- a/test/integration/e2e_io_recursive.sh
+++ b/test/integration/e2e_io_recursive.sh
@@ -64,7 +64,7 @@ echo "Setting up test inputs"
 echo "Setting up pipeline input..."
 build_recursive_files "${INPUT_DEEP}" "${INPUT_SHALLOW}"
 
-gsutil -m rsync -r "${LOCAL_INPUTS}" "${INPUTS}/"
+gcloud storage rsync -r "${LOCAL_INPUTS}" "${INPUTS}/"
 
 echo "Launching pipeline..."
 
@@ -87,7 +87,7 @@ setup_expected_fs_output_entries "${DOCKER_GCS_OUTPUTS}"
 setup_expected_remote_output_entries "${OUTPUTS}"
 
 # Verify in the stdout file that the expected directories were written
-readonly RESULT=$(gsutil cat "${STDOUT_LOG}")
+readonly RESULT=$(gcloud storage cat "${STDOUT_LOG}")
 
 readonly FS_FIND_IN=$(echo "${RESULT}" | sed -n '/^BEGIN: find$/,/^END: find$/p' \
   | grep --fixed-strings /mnt/data/input/"${DOCKER_GCS_INPUTS}")
@@ -134,11 +134,15 @@ echo
 echo "On-disk output file list matches expected"
 
 # Verify in GCS that the DEEP directory is deep and the SHALLOW directory
-# is shallow. Gsutil prints directories with a trailing "/:" marker that is
-# stripped using sed in order to match the output format of the `find` utility.
-readonly GCS_FIND="$(gsutil ls -r "${OUTPUTS}" \
+# is shallow. For each directory, `gcloud storage ls -r` prints both a
+# redundant header line ending in "/:" and a self-referencing entry line
+# ending in "/". The header lines are dropped and the trailing "/" is
+# stripped from the remaining lines to match the output format of the
+# `find` utility.
+readonly GCS_FIND="$(gcloud storage ls -r "${OUTPUTS}" \
                      | grep -v '^ *$' \
-                     | sed -e 's#/:$##')"
+                     | grep -v ':$' \
+                     | sed -e 's#/$##')"
 
 for REC in "${EXPECTED_REMOTE_OUTPUT_ENTRIES[@]}"; do
   if ! echo "${GCS_FIND}" | grep --quiet --fixed-strings "${REC}"; then
diff --git a/test/integration/e2e_io_tasks.py b/test/integration/e2e_io_tasks.py
index 85e422b..604b421 100644
--- a/test/integration/e2e_io_tasks.py
+++ b/test/integration/e2e_io_tasks.py
@@ -104,7 +104,7 @@
       '--output OUTPUT_PATH')
   OUTPUT_FILE = '%s/%s.md5' % (OUTPUT_PATH[:-len('/*.md5')],
                                os.path.basename(INPUT_BAM))
-  RESULT = test_util.gsutil_cat(OUTPUT_FILE)
+  RESULT = test_util.gcloud_cat(OUTPUT_FILE)
 
   if not test_util.diff(RESULT_EXPECTED.strip(), RESULT.strip()):
     print('Output file does not match expected')
@@ -119,7 +119,7 @@
 RESULT_EXPECTED = POPULATION_MD5
 for i in range(TASKS_COUNT):
   OUTPUT_FILE = '%s/TASK_%s.md5' % (test.OUTPUTS, (i + 1))
-  RESULT = test_util.gsutil_cat(OUTPUT_FILE)
+  RESULT = test_util.gcloud_cat(OUTPUT_FILE)
 
   if not test_util.diff(RESULT_EXPECTED.strip(), RESULT.strip()):
     print('Output file does not match expected')
diff --git a/test/integration/e2e_logging_content.sh b/test/integration/e2e_logging_content.sh
index 1aabec0..4e51f9d 100755
--- a/test/integration/e2e_logging_content.sh
+++ b/test/integration/e2e_logging_content.sh
@@ -74,7 +74,7 @@ echo "Checking output..."
 # Check the results
 readonly STDOUT_RESULT_EXPECTED="$(echo -n "${STDOUT_MSG%.}")"
 
-readonly STDOUT_RESULT="$(gsutil cat "${STDOUT_LOG}")"
+readonly STDOUT_RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
 if ! diff <(echo "${STDOUT_RESULT_EXPECTED}") <(echo "${STDOUT_RESULT}"); then
   echo "STDOUT file does not match expected"
   exit 1
@@ -82,7 +82,7 @@ fi
 
 readonly STDERR_RESULT_EXPECTED="$(echo -n "${STDERR_MSG%.}")"
 
-readonly STDERR_RESULT="$(gsutil cat "${STDERR_LOG}")"
+readonly STDERR_RESULT="$(gcloud storage cat "${STDERR_LOG}")"
 if ! diff <(echo "${STDERR_RESULT_EXPECTED}") <(echo "${STDERR_RESULT}"); then
   echo "STDERR file does not match expected"
   exit 1
diff --git a/test/integration/e2e_runtime.sh b/test/integration/e2e_runtime.sh
index fb276b9..4bd9e30 100755
--- a/test/integration/e2e_runtime.sh
+++ b/test/integration/e2e_runtime.sh
@@ -64,7 +64,7 @@ TMPDIR:
 EOF
 )
 
-readonly RESULT="$(gsutil cat "${STDOUT_LOG}")"
+readonly RESULT="$(gcloud storage cat "${STDOUT_LOG}")"
 if ! diff <(echo "${RESULT_EXPECTED}") <(echo "${RESULT}"); then
   echo "Output file does not match expected"
   exit 1
diff --git a/test/integration/e2e_skip.sh b/test/integration/e2e_skip.sh
index 20eb2c0..99d48ed 100755
--- a/test/integration/e2e_skip.sh
+++ b/test/integration/e2e_skip.sh
@@ -29,9 +29,9 @@ source "${SCRIPT_DIR}/test_setup_e2e.sh"
 TEST_FILE_PATH_1="${OUTPUTS}/testfile_1.txt"
 TEST_FILE_PATH_2="${OUTPUTS}/testfile_2.txt"
 
-echo "hello world" | gsutil cp - "${TEST_FILE_PATH_1}"
+echo "hello world" | gcloud storage cp - "${TEST_FILE_PATH_1}"
 
-if gsutil ls "${TEST_FILE_PATH_2}" &> /dev/null; then
+if gcloud storage ls "${TEST_FILE_PATH_2}" &> /dev/null; then
   echo "Unexpected: the output file '${TEST_FILE_PATH_1}' already exists."
   exit 1
 fi
@@ -44,7 +44,7 @@ JOB_ID="$(
     --skip \
     --wait)"
 
-RESULT="$(gsutil cat "${TEST_FILE_PATH_1}")"
+RESULT="$(gcloud storage cat "${TEST_FILE_PATH_1}")"
 if [[ "${RESULT}" != "hello world" ]]; then
   echo "Output file does not match expected (from step 4)"
   echo "Expected: hello world"
@@ -60,7 +60,7 @@ JOB_ID="$(
     --skip \
     --wait)"
 
-RESULT="$(gsutil cat "${TEST_FILE_PATH_2}")"
+RESULT="$(gcloud storage cat "${TEST_FILE_PATH_2}")"
 if [[ "${RESULT}" != "hello from the job" ]]; then
   echo "Output file does not match expected (from step 2)"
   echo "Expected: hello world"
diff --git a/test/integration/e2e_skip_tasks.sh b/test/integration/e2e_skip_tasks.sh
index 4359c78..0ef9ecb 100755
--- a/test/integration/e2e_skip_tasks.sh
+++ b/test/integration/e2e_skip_tasks.sh
@@ -30,7 +30,7 @@ TEST_FILE_PATH_1="${OUTPUTS}/testfile_1.txt"
 TEST_FILE_PATH_2="${OUTPUTS}/testfile_2.txt"
 TEST_FILE_PATH_3="${OUTPUTS}/testfile_3.txt"
 
-if gsutil ls "${TEST_FILE_PATH_1}" &> /dev/null; then
+if gcloud storage ls "${TEST_FILE_PATH_1}" &> /dev/null; then
   echo "Unexpected: the output file '${TEST_FILE_PATH_1}' already exists."
   exit 1
 fi
@@ -51,12 +51,12 @@ JOB_ID="$(
     --skip \
     --wait)"
 
-if ! gsutil ls "${TEST_FILE_PATH_1}" &> /dev/null; then
+if ! gcloud storage ls "${TEST_FILE_PATH_1}" &> /dev/null; then
   echo "Unexpected: the output file '${TEST_FILE_PATH_1}' was not created."
   exit 1
 fi
 
-RESULT="$(gsutil cat "${TEST_FILE_PATH_1}")"
+RESULT="$(gcloud storage cat "${TEST_FILE_PATH_1}")"
 if [[ "${RESULT}" != "hello world" ]]; then
   echo "Output file does not match expected (from step 1)"
   echo "Expected: hello world"
@@ -72,7 +72,7 @@ JOB_ID="$(
     --skip \
     --wait)"
 
-RESULT="$(gsutil cat "${TEST_FILE_PATH_1}")"
+RESULT="$(gcloud storage cat "${TEST_FILE_PATH_1}")"
 if [[ "${RESULT}" != "hello world" ]]; then
   echo "Output file does not match expected (from step 2)"
   echo "Expected: hello world"
@@ -96,14 +96,14 @@ JOB_ID="$(
     --skip \
     --wait)"
 
-RESULT="$(gsutil cat "${TEST_FILE_PATH_1}")"
+RESULT="$(gcloud storage cat "${TEST_FILE_PATH_1}")"
 if [[ "${RESULT}" != "hello world" ]]; then
   echo "Output file does not match expected (from step 3)"
   echo "Expected: hello world"
   echo "Got: ${RESULT}"
   exit 1
 fi
-RESULT="$(gsutil cat "${TEST_FILE_PATH_2}")"
+RESULT="$(gcloud storage cat "${TEST_FILE_PATH_2}")"
 if [[ "${RESULT}" != "hello again from row 2" ]]; then
   echo "Output file does not match expected (from step 3)"
   echo "Expected: hello again from row 2"
diff --git a/test/integration/io_setup.sh b/test/integration/io_setup.sh
index 6cf7caf..db62459 100644
--- a/test/integration/io_setup.sh
+++ b/test/integration/io_setup.sh
@@ -54,10 +54,10 @@ readonly TEST_LOCAL_MOUNT_PARAMETER="file://${TEST_TMP_PATH}"
 function io_setup::mount_local_path_setup() {
   mkdir -p "${TEST_TMP_PATH}"
   if [[ ! -f "${TEST_TMP_PATH}/${POPULATION_FILE}" ]]; then
-    gsutil cp "${POPULATION_FILE_FULL_PATH}" "${TEST_TMP_PATH}/${POPULATION_FILE}"
+    gcloud storage cp "${POPULATION_FILE_FULL_PATH}" "${TEST_TMP_PATH}/${POPULATION_FILE}"
   fi
   if [[ ! -f "${TEST_TMP_PATH}/${INPUT_BAM_FILE}" ]]; then
-    gsutil cp "${INPUT_BAM_FULL_PATH}" "${TEST_TMP_PATH}/${INPUT_BAM_FILE}"
+    gcloud storage cp "${INPUT_BAM_FULL_PATH}" "${TEST_TMP_PATH}/${INPUT_BAM_FILE}"
   fi
 }
 readonly -f io_setup::mount_local_path_setup
@@ -190,7 +190,7 @@ function io_setup::_check_output() {
   local output_file="${1}"
   local result_expected="${2}"
 
-  local result=$(gsutil cat "${output_file}")
+  local result=$(gcloud storage cat "${output_file}")
   if ! diff <(echo "${result_expected}") <(echo "${result}"); then
     echo "Output file does not match expected"
     exit 1
diff --git a/test/integration/io_tasks_setup.sh b/test/integration/io_tasks_setup.sh
index 9b93627..f78d4ec 100644
--- a/test/integration/io_tasks_setup.sh
+++ b/test/integration/io_tasks_setup.sh
@@ -70,7 +70,7 @@ function io_tasks_setup::check_output() {
 
     output_path="$(grep "${input_bam}" "${TASKS_FILE}" | cut -d $'\t' -f 3)"
     output_file="${output_path%/*.md5}/$(basename "${input_bam}").md5"
-    result="$(gsutil cat "${output_file}")"
+    result="$(gcloud storage cat "${output_file}")"
 
     if ! diff <(echo "${expected}") <(echo "${result}"); then
       echo "Output file does not match expected"
@@ -88,7 +88,7 @@ function io_tasks_setup::check_output() {
   expected="${POPULATION_MD5}"
   for ((i=0; i < tasks_count; i++)); do
     output_file="${OUTPUTS}/TASK_$((i+1)).md5"
-    result="$(gsutil cat "${output_file}")"
+    result="$(gcloud storage cat "${output_file}")"
 
     if ! diff <(echo "${expected}") <(echo "${result}"); then
       echo "Output file does not match expected"
diff --git a/test/integration/script_block_external_network.sh b/test/integration/script_block_external_network.sh
index a2b20e5..052ace3 100755
--- a/test/integration/script_block_external_network.sh
+++ b/test/integration/script_block_external_network.sh
@@ -21,8 +21,8 @@ set -o nounset
 
 RC=0
 
-if ! gsutil -o 'Boto:num_retries=0' ls gs://genomics-public-data; then
-  1>&2 echo "\`gsutil ls\` should not have succeeded"
+if ! CLOUDSDK_STORAGE_MAX_RETRIES=0 gcloud storage ls gs://genomics-public-data; then
+  1>&2 echo "\`gcloud storage ls\` should not have succeeded"
   RC=1
 fi
 
diff --git a/test/integration/test_setup_e2e.py b/test/integration/test_setup_e2e.py
index 898d878..f5df9e8 100644
--- a/test/integration/test_setup_e2e.py
+++ b/test/integration/test_setup_e2e.py
@@ -100,9 +100,9 @@ def _environ():
 print("  Bucket detected as: %s" % DSUB_BUCKET)
 
 print("  Checking if bucket exists")
-if not test_util.gsutil_ls_check("gs://%s" % DSUB_BUCKET):
+if not test_util.gcloud_ls_check("gs://%s" % DSUB_BUCKET):
   print("Bucket does not exist: %s" % DSUB_BUCKET, file=sys.stderr)
-  print("Create the bucket with \"gsutil mb\".", file=sys.stderr)
+  print("Create the bucket with \"gcloud storage buckets create\".", file=sys.stderr)
   sys.exit(1)
 
 # Set standard LOGGING, INPUTS, and OUTPUTS values
@@ -130,10 +130,10 @@ def _environ():
 print("Output path: %s" % OUTPUTS)
 
 print("  Checking if remote test files already exists")
-if test_util.gsutil_ls_check("%s/**" % TEST_GCS_ROOT):
+if test_util.gcloud_ls_check("%s/**" % TEST_GCS_ROOT):
   print("Test files exist: %s" % TEST_GCS_ROOT, file=sys.stderr)
   print("Remove contents:", file=sys.stderr)
-  print("  gsutil -m rm %s/**" % TEST_GCS_ROOT, file=sys.stderr)
+  print("  gcloud storage rm %s/**" % TEST_GCS_ROOT, file=sys.stderr)
   sys.exit(1)
 
 if TASKS_FILE:
diff --git a/test/integration/test_setup_e2e.sh b/test/integration/test_setup_e2e.sh
index c5e7cc1..00d6be9 100755
--- a/test/integration/test_setup_e2e.sh
+++ b/test/integration/test_setup_e2e.sh
@@ -64,13 +64,12 @@ fi
 echo "  Bucket detected as: ${DSUB_BUCKET}"
 
 echo "  Checking if bucket exists"
-if ! gsutil ls "gs://${DSUB_BUCKET}" 2>/dev/null; then
+if ! gcloud storage ls "gs://${DSUB_BUCKET}" 2>/dev/null; then
   1>&2 echo "Bucket does not exist (or we have no access): ${DSUB_BUCKET}"
-  1>&2 echo "Create the bucket with \"gsutil mb\"."
+  1>&2 echo "Create the bucket with \"gcloud storage buckets create\"."
   1>&2 echo "Current gcloud settings:"
   1>&2 echo "  account: $(gcloud config get-value account 2>/dev/null)"
   1>&2 echo "  project: $(gcloud config get-value project 2>/dev/null)"
-  1>&2 echo "  pass_credentials_to_gsutil: $(gcloud config get-value pass_credentials_to_gsutil 2>/dev/null)"
   exit 1
 fi
 
@@ -155,10 +154,10 @@ echo "Output path: ${OUTPUTS}"
 readonly DSUB_PARAMS="${TEST_GCS_ROOT}/params"
 
 echo "  Checking if remote test files already exists"
-if gsutil ls "${TEST_GCS_ROOT}/**" 2>/dev/null; then
+if gcloud storage ls "${TEST_GCS_ROOT}/**" 2>/dev/null; then
   1>&2 echo "Test files exist: ${TEST_GCS_ROOT}"
   1>&2 echo "Remove contents:"
-  1>&2 echo "  gsutil -m rm ${TEST_GCS_ROOT}/**"
+  1>&2 echo "  gcloud storage rm ${TEST_GCS_ROOT}/**"
   exit 1
 fi
 
diff --git a/test/integration/test_util.py b/test/integration/test_util.py
index 8100261..3014e97 100644
--- a/test/integration/test_util.py
+++ b/test/integration/test_util.py
@@ -29,13 +29,13 @@ def to_string(stdoutbytes):
   return stdoutbytes.decode(encoding)
 
 
-def gsutil_ls_check(path):
-  return not subprocess.call('gsutil ls "%s" 2>/dev/null' % path, shell=True)
+def gcloud_ls_check(path):
+  return not subprocess.call('gcloud storage ls "%s" 2>/dev/null' % path, shell=True)
 
 
-def gsutil_cat(path):
+def gcloud_cat(path):
   return to_string(
-      subprocess.check_output('gsutil cat "%s"' % path, shell=True))
+      subprocess.check_output('gcloud storage cat "%s"' % path, shell=True))
 
 
 def diff(str1, str2):
diff --git a/test/integration/unit_skip.test-fails.sh b/test/integration/unit_skip.test-fails.sh
index e80a500..5fd629d 100755
--- a/test/integration/unit_skip.test-fails.sh
+++ b/test/integration/unit_skip.test-fails.sh
@@ -46,8 +46,8 @@ readonly NEWFILE="${OUTPUTS}/newfile"
 readonly OUT_FOLDER_2="${OUTPUTS}/newfolder"
 
 # Create pre-existing output
-echo "test output" | gsutil cp - "${EXISTING}"
-echo "test output" | gsutil cp - "${EXISTING_2}"
+echo "test output" | gcloud storage cp - "${EXISTING}"
+echo "test output" | gcloud storage cp - "${EXISTING_2}"
 
 echo "Job 1 ..."