def _set_up_single_archive(self):
key_name = "test-key"
test_file_path = os.path.join(
test_dir_path, "test-orignal"
)
test_file_size = 1024 ** 2
test_data = os.urandom(test_file_size)
with open(test_file_path, "wb") as output_file:
output_file.write(test_data)
# create the bucket
bucket = self._s3_connection.create_unique_bucket()
self.assertTrue(bucket is not None)
# create an empty key
write_key = Key(bucket)
# set the name
write_key.name = key_name
self.assertFalse(write_key.exists())
# upload some data
with open(test_file_path, "rb") as archive_file:
write_key.set_contents_from_file(archive_file)
self.assertTrue(write_key.exists())
return test_data, write_key
def test_key_with_files(self):
"""
test simple key 'from_file' and 'to_file' functions
"""
log = logging.getLogger("test_key_with_files")
bucket_name = "com-dougfort-test-key-with-files"
key_name = "A" * 1024
test_file_path = os.path.join(
_test_dir_path, "test_key_with_files-orignal"
)
test_file_size = 1024 ** 2
buffer_size = 1024
log.debug("writing %s bytes to %s" % (
test_file_size, test_file_path,
))
bytes_written = 0
with open(test_file_path, "w") as output_file:
while bytes_written < test_file_size:
output_file.write(_random_string(buffer_size))
bytes_written += buffer_size
# create the bucket
bucket = self._s3_connection.create_bucket(bucket_name)
self.assertTrue(bucket is not None)
self.assertEqual(bucket.name, bucket_name)
# create an empty key
write_key = Key(bucket)
# set the name
write_key.name = key_name
self.assertFalse(write_key.exists())
# upload some data
with open(test_file_path, "r") as archive_file:
write_key.set_contents_from_file(archive_file)
self.assertTrue(write_key.exists())
# create another key with the same name
read_key = Key(bucket, key_name)
# read back the data
retrieve_file_path = os.path.join(
_test_dir_path, "test_key_with_files-orignal"
)
with open(retrieve_file_path, "w") as retrieve_file:
read_key.get_contents_to_file(retrieve_file)
self.assertTrue(
filecmp.cmp(test_file_path, retrieve_file_path, shallow=False)
)
# delete the key
read_key.delete()
self.assertFalse(write_key.exists())
# delete the bucket
self._s3_connection.delete_bucket(bucket_name)
def _archive_one_file(
self,
bucket,
key_name,
replace,
size,
):
key = Key(bucket)
key.name = key_name
self._log.info("_archive_one_file ({0} {1} ...) versioning={2}".format(
bucket.name,
key.name,
bucket.versioning,
))
bucket_accounting = self._bucket_accounting[bucket.name]
retry_count = 0
force_error = random.randint(0, 99) < self._archive_failure_percent
while not self._halt_event.is_set():
bucket_accounting.increment_by("archive_request", 1)
input_file = MockInputFile(size, force_error)
try:
key.set_contents_from_file(input_file, replace=replace)
except MockInputFileError:
bucket_accounting.increment_by("archive_error", 1)
self._log.info("MockInputFileError")
return
except LumberyardRetryableHTTPError, instance:
bucket_accounting.increment_by("archive_error", 1)
if retry_count >= _max_archive_retries:
raise
self._log.warn("%s: retry in %s seconds" % (
instance,
instance.retry_after,
))
self._halt_event.wait(timeout=instance.retry_after)
retry_count += 1
self._log.warn("retry #%s" % (retry_count, ))
continue
verification_key = (
bucket.name,
key.name,
key.version_id,
)
self._log.info("archived {0}".format(verification_key))
if verification_key in self.key_verification:
self._log.error("_archive_one_file duplicate key %s" %
(verification_key, ))
bucket_accounting.increment_by("archive_success", 1)
# we count this as 'bytes in' because that's what the server counts
bucket_accounting.increment_by("success_bytes_in", size)
self.key_verification[verification_key] = \
(size, input_file.md5_digest, )
break
def test_uninterrupted_resumable(self):
"""
test get_contents_to_file without any interruption.
"""
log = logging.getLogger("test_uninterrupted_resumable")
key_name = "test-key"
test_file_path = os.path.join(test_dir_path, "test-orignal")
test_file_size = 1024**2
buffer_size = 1024
log.debug("writing {0} bytes to {1}".format(test_file_size,
test_file_path))
bytes_written = 0
with open(test_file_path, "wb") as output_file:
while bytes_written < test_file_size:
output_file.write(os.urandom(buffer_size))
bytes_written += buffer_size
# create the bucket
bucket = self._s3_connection.create_unique_bucket()
self.assertTrue(bucket is not None)
# create an empty key
write_key = Key(bucket)
# set the name
write_key.name = key_name
self.assertFalse(write_key.exists())
# upload some data
with open(test_file_path, "rb") as archive_file:
write_key.set_contents_from_file(archive_file)
self.assertTrue(write_key.exists())
# create a ResumableDownloadHandler
tracker_file_path = os.path.join(test_dir_path, "tracker-file")
download_handler = ResumableDownloadHandler(
tracker_file_name=tracker_file_path)
# read back the data
retrieve_file_path = os.path.join(test_dir_path,
"test_key_with_files-orignal")
with open(retrieve_file_path, "wb") as retrieve_file:
write_key.get_contents_to_file(retrieve_file,
res_download_handler=\
download_handler)
self.assertTrue(
filecmp.cmp(test_file_path, retrieve_file_path, shallow=False))
# delete the key
write_key.delete()
self.assertFalse(write_key.exists())
# delete the bucket
self._s3_connection.delete_bucket(bucket.name)
def test_well_behaved_upload(self):
"""
an upload that sends sequences of 10mb
"""
key_name = "test-key"
sequence_size = 10 * 1024 * 1024
sequence_count = 3
total_seconds = 3 * 301
write_key = Key(self._s3_connection.default_bucket)
write_key.name = key_name
test_file = MockInputFile(sequence_size, sequence_count, total_seconds)
write_key.set_contents_from_file(test_file)
self.assertTrue(write_key.exists())
def test_fast_upload(self):
"""
the fastest upload we can manage
"""
key_name = "test-key"
sequence_size = 3 * 10 * 1024 * 1024
sequence_count = 1
total_seconds = 0.0
write_key = Key(self._s3_connection.default_bucket)
write_key.name = key_name
test_file = MockInputFile(sequence_size, sequence_count, total_seconds)
write_key.set_contents_from_file(test_file)
self.assertTrue(write_key.exists())
def test_well_behaved_upload(self):
"""
an upload that sends sequences of 10mb
"""
key_name = "test-key"
sequence_size = 10 * 1024 * 1024
sequence_count = 3
total_seconds = 3 * 301
write_key = Key(self._s3_connection.default_bucket)
write_key.name = key_name
test_file = MockInputFile(sequence_size, sequence_count, total_seconds)
write_key.set_contents_from_file(test_file)
self.assertTrue(write_key.exists())
def test_fast_upload(self):
"""
the fastest upload we can manage
"""
key_name = "test-key"
sequence_size = 3 * 10 * 1024 * 1024
sequence_count = 1
total_seconds = 0.0
write_key = Key(self._s3_connection.default_bucket)
write_key.name = key_name
test_file = MockInputFile(sequence_size, sequence_count, total_seconds)
write_key.set_contents_from_file(test_file)
self.assertTrue(write_key.exists())
def _archive_one_file( self, bucket, key_name, replace, size, ):
key = Key(bucket)
key.name = key_name
self._log.info("_archive_one_file ({0} {1} ...) versioning={2}".format(
bucket.name,
key.name,
bucket.versioning,
))
bucket_accounting = self._bucket_accounting[bucket.name]
retry_count = 0
force_error = random.randint(0, 99) < self._archive_failure_percent
while not self._halt_event.is_set():
bucket_accounting.increment_by("archive_request", 1)
input_file = MockInputFile(size, force_error)
try:
key.set_contents_from_file(input_file, replace=replace)
except MockInputFileError:
bucket_accounting.increment_by("archive_error", 1)
self._log.info("MockInputFileError")
return
except LumberyardRetryableHTTPError, instance:
bucket_accounting.increment_by("archive_error", 1)
if retry_count >= _max_archive_retries:
raise
self._log.warn("%s: retry in %s seconds" % (
instance, instance.retry_after,
))
self._halt_event.wait(timeout=instance.retry_after)
retry_count += 1
self._log.warn("retry #%s" % (retry_count, ))
continue
verification_key = (bucket.name, key.name, key.version_id, )
self._log.info("archived {0}".format(verification_key))
if verification_key in self.key_verification:
self._log.error("_archive_one_file duplicate key %s" % (
verification_key, ))
bucket_accounting.increment_by("archive_success", 1)
# we count this as 'bytes in' because that's what the server counts
bucket_accounting.increment_by("success_bytes_in", size)
self.key_verification[verification_key] = \
(size, input_file.md5_digest, )
break
def test_slow_upload(self):
"""
an upload that sends less than 10mb in 5 min
"""
key_name = "test-key"
sequence_size = 1024
sequence_count = 3 * 10
total_seconds = 3 * 10 * 60
write_key = Key(self._s3_connection.default_bucket)
write_key.name = key_name
test_file = MockInputFile(sequence_size, sequence_count, total_seconds)
with self.assertRaises(LumberyardHTTPError) as context_manager:
write_key.set_contents_from_file(test_file)
self.assertEqual(context_manager.exception.status, REQUEST_TIMEOUT)
def test_slow_upload(self):
"""
an upload that sends less than 10mb in 5 min
"""
key_name = "test-key"
sequence_size = 1024
sequence_count = 3 * 10
total_seconds = 3 * 10 * 60
write_key = Key(self._s3_connection.default_bucket)
write_key.name = key_name
test_file = MockInputFile(sequence_size, sequence_count, total_seconds)
with self.assertRaises(LumberyardHTTPError) as context_manager:
write_key.set_contents_from_file(test_file)
self.assertEqual(context_manager.exception.status, REQUEST_TIMEOUT)
def archive_file(args, nimbusio_identity, file_name):
log = logging.getLogger("archive_file")
s3_emulator = motoboto.connect_s3(nimbusio_identity)
if args.collection_name is None:
bucket = s3_emulator.default_bucket
else:
bucket = s3_emulator.get_bucket(args.collection_name)
key_name = "".join([args.prefix, file_name])
archive_key = Key(bucket, key_name)
log.info("archiving {0} as key {1} to collection {2}".format(file_name,
archive_key,
bucket))
file_path = os.path.join(args.watch_path, file_name)
with open(file_path, "rb") as input_file:
archive_key.set_contents_from_file(input_file)
log.info("archive successful version identifier = {0}".format(
archive_key.version_id))
s3_emulator.close()
def test_uninterrupted_resumable(self):
"""
test get_contents_to_file without any interruption.
"""
log = logging.getLogger("test_uninterrupted_resumable")
key_name = "test-key"
test_file_path = os.path.join(
test_dir_path, "test-orignal"
)
test_file_size = 1024 ** 2
buffer_size = 1024
log.debug("writing {0} bytes to {1}".format(test_file_size,
test_file_path))
bytes_written = 0
with open(test_file_path, "wb") as output_file:
while bytes_written < test_file_size:
output_file.write(os.urandom(buffer_size))
bytes_written += buffer_size
# create the bucket
bucket = self._s3_connection.create_unique_bucket()
self.assertTrue(bucket is not None)
# create an empty key
write_key = Key(bucket)
# set the name
write_key.name = key_name
self.assertFalse(write_key.exists())
# upload some data
with open(test_file_path, "rb") as archive_file:
write_key.set_contents_from_file(archive_file)
self.assertTrue(write_key.exists())
# create a ResumableDownloadHandler
tracker_file_path = os.path.join(
test_dir_path, "tracker-file"
)
download_handler = ResumableDownloadHandler(
tracker_file_name=tracker_file_path
)
# read back the data
retrieve_file_path = os.path.join(
test_dir_path, "test_key_with_files-orignal"
)
with open(retrieve_file_path, "wb") as retrieve_file:
write_key.get_contents_to_file(retrieve_file,
res_download_handler=\
download_handler)
self.assertTrue(
filecmp.cmp(test_file_path, retrieve_file_path, shallow=False)
)
# delete the key
write_key.delete()
self.assertFalse(write_key.exists())
# delete the bucket
self._s3_connection.delete_bucket(bucket.name)
def test_interrupted_resumable(self):
"""
test get_contents_to_file with a simulated interruption.
"""
log = logging.getLogger("test_uninterrupted_resumable")
key_name = "test-key"
test_file_path = os.path.join(
test_dir_path, "test-orignal"
)
test_file_size = 1024 ** 2
interrupted_size = 1024 * 42
test_data = os.urandom(test_file_size)
log.debug("writing {0} bytes to {1}".format(test_file_size,
test_file_path))
with open(test_file_path, "wb") as output_file:
output_file.write(test_data)
# create the bucket
bucket = self._s3_connection.create_unique_bucket()
self.assertTrue(bucket is not None)
# create an empty key
write_key = Key(bucket)
# set the name
write_key.name = key_name
# self.assertFalse(write_key.exists())
# upload some data
with open(test_file_path, "rb") as archive_file:
write_key.set_contents_from_file(archive_file)
self.assertTrue(write_key.exists())
# create a ResumableDownloadHandler
tracker_file_path = os.path.join(
test_dir_path, "tracker-file"
)
download_handler = ResumableDownloadHandler(
tracker_file_name=tracker_file_path
)
retrieve_file_path = os.path.join(
test_dir_path, "test_key_with_files-orignal"
)
# copy some of the data to the retrieve file to simulate an
# interrupted retrieve
with open(retrieve_file_path, "wb") as output_file:
output_file.write(test_data[interrupted_size:])
# spoof the resumable handler into thinking it has a retrieve
# in progress
download_handler._save_tracker_info(write_key)
# resume the retrieve
with open(retrieve_file_path, "wb") as retrieve_file:
write_key.get_contents_to_file(retrieve_file,
res_download_handler=\
download_handler)
# read back the retrieved data
with open(retrieve_file_path, "rb") as retrieve_file:
retrieved_data = retrieve_file.read()
self.assertEqual(len(retrieved_data), len(test_data))
self.assertTrue(retrieved_data == test_data)
# delete the key
write_key.delete()
self.assertFalse(write_key.exists())
# delete the bucket
self._s3_connection.delete_bucket(bucket.name)
def test_key_with_files_and_callback(self):
"""
test simple key 'from_file' and 'to_file' functions
"""
def _archive_callback(bytes_sent, total_bytes):
print >> sys.stderr, "archived", str(bytes_sent), "out of", \
str(total_bytes)
def _retrieve_callback(bytes_sent, total_bytes):
print >> sys.stderr, "retrieved", str(bytes_sent), "out of", \
str(total_bytes)
log = logging.getLogger("test_key_with_files")
bucket_name = "com-dougfort-test-key-with-files-and-callback"
key_name = "A" * 1024
test_file_path = os.path.join(
_test_dir_path, "test_key_with_files-orignal"
)
test_file_size = 1024 ** 2
buffer_size = 1024
log.debug("writing %s bytes to %s" % (
test_file_size, test_file_path,
))
bytes_written = 0
with open(test_file_path, "w") as output_file:
while bytes_written < test_file_size:
output_file.write(_random_string(buffer_size))
bytes_written += buffer_size
# create the bucket
bucket = self._s3_connection.create_bucket(bucket_name)
self.assertTrue(bucket is not None)
self.assertEqual(bucket.name, bucket_name)
# create an empty key
write_key = Key(bucket)
# set the name
write_key.name = key_name
self.assertFalse(write_key.exists())
# upload some data
with open(test_file_path, "r") as archive_file:
write_key.set_contents_from_file(
archive_file, cb=_archive_callback
)
self.assertTrue(write_key.exists())
# create another key with the same name
read_key = Key(bucket, key_name)
# read back the data
retrieve_file_path = os.path.join(
_test_dir_path, "test_key_with_files-orignal"
)
# 2011-08-08 dougfort boto aborts if you don't tell it the size
read_key.size = test_file_size
with open(retrieve_file_path, "w") as retrieve_file:
read_key.get_contents_to_file(
retrieve_file, cb=_retrieve_callback
)
self.assertTrue(
filecmp.cmp(test_file_path, retrieve_file_path, shallow=False)
)
# delete the key
read_key.delete()
self.assertFalse(write_key.exists())
# delete the bucket
self._s3_connection.delete_bucket(bucket_name)
def test_interrupted_resumable(self):
"""
test get_contents_to_file with a simulated interruption.
"""
log = logging.getLogger("test_uninterrupted_resumable")
key_name = "test-key"
test_file_path = os.path.join(test_dir_path, "test-orignal")
test_file_size = 1024**2
interrupted_size = 1024 * 42
test_data = os.urandom(test_file_size)
log.debug("writing {0} bytes to {1}".format(test_file_size,
test_file_path))
with open(test_file_path, "wb") as output_file:
output_file.write(test_data)
# create the bucket
bucket = self._s3_connection.create_unique_bucket()
self.assertTrue(bucket is not None)
# create an empty key
write_key = Key(bucket)
# set the name
write_key.name = key_name
# self.assertFalse(write_key.exists())
# upload some data
with open(test_file_path, "rb") as archive_file:
write_key.set_contents_from_file(archive_file)
self.assertTrue(write_key.exists())
# create a ResumableDownloadHandler
tracker_file_path = os.path.join(test_dir_path, "tracker-file")
download_handler = ResumableDownloadHandler(
tracker_file_name=tracker_file_path)
retrieve_file_path = os.path.join(test_dir_path,
"test_key_with_files-orignal")
# copy some of the data to the retrieve file to simulate an
# interrupted retrieve
with open(retrieve_file_path, "wb") as output_file:
output_file.write(test_data[interrupted_size:])
# spoof the resumable handler into thinking it has a retrieve
# in progress
download_handler._save_tracker_info(write_key)
# resume the retrieve
with open(retrieve_file_path, "wb") as retrieve_file:
write_key.get_contents_to_file(retrieve_file,
res_download_handler=\
download_handler)
# read back the retrieved data
with open(retrieve_file_path, "rb") as retrieve_file:
retrieved_data = retrieve_file.read()
self.assertEqual(len(retrieved_data), len(test_data))
self.assertTrue(retrieved_data == test_data)
# delete the key
write_key.delete()
self.assertFalse(write_key.exists())
# delete the bucket
self._s3_connection.delete_bucket(bucket.name)
