diff --git a/endpoints/v2/blob.py b/endpoints/v2/blob.py
index 42439650a..56d6dbd10 100644
--- a/endpoints/v2/blob.py
+++ b/endpoints/v2/blob.py
@@ -124,8 +124,10 @@ def start_blob_upload(namespace, repo_name):
     return accepted
   else:
     # The user plans to send us the entire body right now
-    uploaded = _upload_chunk(namespace, repo_name, new_upload_uuid)
+    uploaded, error = _upload_chunk(namespace, repo_name, new_upload_uuid)
     uploaded.save()
+    if error:
+      _range_not_satisfiable(uploaded.byte_count)
 
     return _finish_upload(namespace, repo_name, uploaded, digest)
 
@@ -210,16 +212,16 @@ def _upload_chunk(namespace, repo_name, upload_uuid):
     input_fp = wrap_with_handler(input_fp, found.sha_state.update)
 
     try:
-      length_written, new_metadata = storage.stream_upload_chunk(location_set, upload_uuid,
-                                                                 start_offset, length, input_fp,
-                                                                 found.storage_metadata,
-                                                                 content_type=BLOB_CONTENT_TYPE)
+      length_written, new_metadata, error = storage.stream_upload_chunk(location_set, upload_uuid,
+                                                                        start_offset, length, input_fp,
+                                                                        found.storage_metadata,
+                                                                        content_type=BLOB_CONTENT_TYPE)
     except InvalidChunkException:
       _range_not_satisfiable(found.byte_count)
 
   found.storage_metadata = new_metadata
   found.byte_count += length_written
-  return found
+  return found, error
 
 
 def _finish_upload(namespace, repo_name, upload_obj, expected_digest):
@@ -263,9 +265,12 @@ def _finish_upload(namespace, repo_name, upload_obj, expected_digest):
 @require_repo_write
 @anon_protect
 def upload_chunk(namespace, repo_name, upload_uuid):
-  upload = _upload_chunk(namespace, repo_name, upload_uuid)
+  upload, error = _upload_chunk(namespace, repo_name, upload_uuid)
   upload.save()
 
+  if error:
+    _range_not_satisfiable(upload.byte_count)
+
   accepted = make_response('', 204)
   accepted.headers['Location'] = _current_request_path()
   accepted.headers['Range'] = _render_range(upload.byte_count, with_bytes_prefix=False)
@@ -282,7 +287,12 @@ def monolithic_upload_or_last_chunk(namespace, repo_name, upload_uuid):
   if digest is None:
     raise BlobUploadInvalid()
 
-  found = _upload_chunk(namespace, repo_name, upload_uuid)
+  found, error = _upload_chunk(namespace, repo_name, upload_uuid)
+
+  if error:
+    found.save()
+    _range_not_satisfiable(found.byte_count)
+
   return _finish_upload(namespace, repo_name, found, digest)
 
 
diff --git a/storage/basestorage.py b/storage/basestorage.py
index 37d473177..dee4937c9 100644
--- a/storage/basestorage.py
+++ b/storage/basestorage.py
@@ -100,15 +100,11 @@ class BaseStorage(StoragePaths):
       if size_to_read < 0:
         size_to_read = self.buffer_size
 
-      try:
-        buf = in_fp.read(size_to_read)
-        if not buf:
-          break
-        out_fp.write(buf)
-        bytes_copied += len(buf)
-      except IOError as err:
-        logger.error('Failed to stream_write_to_fp: %s', err)
+      buf = in_fp.read(size_to_read)
+      if not buf:
         break
+      out_fp.write(buf)
+      bytes_copied += len(buf)
 
     return bytes_copied
 
diff --git a/storage/cloud.py b/storage/cloud.py
index 9b6c41eaa..4c0ee9d60 100644
--- a/storage/cloud.py
+++ b/storage/cloud.py
@@ -166,17 +166,16 @@ class _CloudStorage(BaseStorageV2):
                                                         **self._upload_params)
 
   def stream_write(self, path, fp, content_type=None, content_encoding=None):
-    return self._stream_write_internal(path, fp, content_type, content_encoding)
+    self._stream_write_internal(path, fp, content_type, content_encoding)
 
   def _stream_write_internal(self, path, fp, content_type=None, content_encoding=None,
                              cancel_on_error=True, size=filelike.READ_UNTIL_END):
+    error = None
     mp = self.__initiate_multipart_upload(path, content_type, content_encoding)
 
     # We are going to reuse this but be VERY careful to only read the number of bytes written to it
     buf = StringIO.StringIO()
 
-    chunk_sizes = []
-
     num_part = 1
     total_bytes_written = 0
     while size == filelike.READ_UNTIL_END or total_bytes_written < size:
@@ -192,26 +191,24 @@ class _CloudStorage(BaseStorageV2):
         if bytes_staged == 0:
           break
 
-        chunk_sizes.append(bytes_staged)
-
         buf.seek(0)
         mp.upload_part_from_file(buf, num_part, size=bytes_staged)
         total_bytes_written += bytes_staged
         num_part += 1
-      except IOError:
+      except IOError as ex:
+        logger.warn('stream write error: %s', ex)
+        error = ex
         app.metric_queue.put('MultipartUploadFailure', 1)
         if cancel_on_error:
           mp.cancel_upload()
-          return 0
+          return 0, error
+        else:
+          break
 
     if total_bytes_written > 0:
       app.metric_queue.put('MultipartUploadSuccess', 1)
-      try:
-        mp.complete_upload()
-      except:
-        logger.error('complete_upload failed: chunk_sizes: %s', chunk_sizes)
-        raise
-    return total_bytes_written
+      mp.complete_upload()
+    return total_bytes_written, error
 
   def list_directory(self, path=None):
     self._initialize_cloud_conn()
@@ -307,8 +304,8 @@ class _CloudStorage(BaseStorageV2):
 
     # We are going to upload each chunk to a separate key
     chunk_path = self._rel_upload_path(str(uuid4()))
-    bytes_written = self._stream_write_internal(chunk_path, in_fp, cancel_on_error=False,
-                                                size=length, content_type=content_type)
+    bytes_written, error = self._stream_write_internal(chunk_path, in_fp, cancel_on_error=False,
+                                                       size=length, content_type=content_type)
 
     new_metadata = copy.deepcopy(storage_metadata)
 
@@ -316,7 +313,7 @@ class _CloudStorage(BaseStorageV2):
     if bytes_written > 0:
       new_metadata[_CHUNKS_KEY].append(_PartUploadMetadata(chunk_path, offset, bytes_written))
 
-    return bytes_written, new_metadata
+    return bytes_written, new_metadata, error
 
   def _chunk_generator(self, chunk_list):
     for chunk in chunk_list:
@@ -480,9 +477,12 @@ class GoogleCloudStorage(_CloudStorage):
       fp = filelike.StreamSlice(fp, 0, size)
 
     # TODO figure out how to handle cancel_on_error=False
-    key.set_contents_from_stream(fp)
+    try:
+      key.set_contents_from_stream(fp)
+    except IOError as ex:
+      return 0, ex
 
-    return key.size
+    return key.size, None
 
   def complete_chunked_upload(self, uuid, final_path, storage_metadata):
     self._initialize_cloud_conn()
diff --git a/storage/fakestorage.py b/storage/fakestorage.py
index e9f6305ea..12440a32c 100644
--- a/storage/fakestorage.py
+++ b/storage/fakestorage.py
@@ -57,7 +57,10 @@ class FakeStorage(BaseStorageV2):
   def stream_upload_chunk(self, uuid, offset, length, in_fp, _, content_type=None):
     upload_storage = _FAKE_STORAGE_MAP[uuid]
     upload_storage.seek(offset)
-    return self.stream_write_to_fp(in_fp, upload_storage, length), {}
+    try:
+      return self.stream_write_to_fp(in_fp, upload_storage, length), {}, None
+    except IOError as ex:
+      return 0, {}, ex
 
   def complete_chunked_upload(self, uuid, final_path, _):
     _FAKE_STORAGE_MAP[final_path] = _FAKE_STORAGE_MAP[uuid]
diff --git a/storage/local.py b/storage/local.py
index 7a4c0d10d..d574293bc 100644
--- a/storage/local.py
+++ b/storage/local.py
@@ -106,9 +106,12 @@ class LocalStorage(BaseStorageV2):
     return new_uuid, {}
 
   def stream_upload_chunk(self, uuid, offset, length, in_fp, _, content_type=None):
-    with open(self._init_path(self._rel_upload_path(uuid)), 'r+b') as upload_storage:
-      upload_storage.seek(offset)
-      return self.stream_write_to_fp(in_fp, upload_storage, length), {}
+    try:
+      with open(self._init_path(self._rel_upload_path(uuid)), 'r+b') as upload_storage:
+        upload_storage.seek(offset)
+        return self.stream_write_to_fp(in_fp, upload_storage, length), {}, None
+    except IOError as ex:
+      return 0, {}, ex
 
   def complete_chunked_upload(self, uuid, final_path, _):
     content_path = self._rel_upload_path(uuid)
diff --git a/storage/swift.py b/storage/swift.py
index 71631a608..472a8c92f 100644
--- a/storage/swift.py
+++ b/storage/swift.py
@@ -269,15 +269,22 @@ class SwiftStorage(BaseStorage):
     return random_uuid, metadata
 
   def stream_upload_chunk(self, uuid, offset, length, in_fp, storage_metadata, content_type=None):
+    error = None
+
     if length == 0:
-      return 0, storage_metadata
+      return 0, storage_metadata, error
 
     # Note: Swift limits segments to a maximum of 5GB, so we keep writing segments until we
     # are finished hitting the data limit.
     total_bytes_written = 0
     while True:
-      bytes_written, storage_metadata = self._stream_upload_segment(uuid, offset, length, in_fp,
-                                                                    storage_metadata, content_type)
+      try:
+        bytes_written, storage_metadata = self._stream_upload_segment(uuid, offset, length, in_fp,
+                                                                      storage_metadata, content_type)
+      except IOError as ex:
+        logger.warn('stream write error: %s', ex)
+        error = ex
+        break
 
       if length != filelike.READ_UNTIL_END:
         length = length - bytes_written
@@ -285,7 +292,9 @@ class SwiftStorage(BaseStorage):
       offset = offset + bytes_written
       total_bytes_written = total_bytes_written + bytes_written
       if bytes_written == 0 or length <= 0:
-        return total_bytes_written, storage_metadata
+        return total_bytes_written, storage_metadata, error
+
+    return total_bytes_written, storage_metadata, error
 
   def _stream_upload_segment(self, uuid, offset, length, in_fp, storage_metadata, content_type):
     updated_metadata = copy.deepcopy(storage_metadata)