def run(self):
log.info('initiating snakebite hdfs client')
try:
client = AutoConfigClient()
except krbV.Krb5Error as _: # pylint: disable=no-member
if self.verbose:
print('', file=sys.stderr)
print(_, file=sys.stderr)
start_time = time.time()
dir_count = 0
file_count = 0
repl1_count = 0
for path in self.path_list:
try:
result_list = client.ls([path],
recurse=True,
include_toplevel=True,
include_children=True)
for result in result_list:
if self.verbose and (dir_count + file_count) % 100 == 0:
print('.', file=sys.stderr, end='')
if result['block_replication'] == 0:
dir_count += 1
continue
file_count += 1
if result['block_replication'] == 1:
file_path = result['path']
repl1_count += 1
if self.verbose:
print('', file=sys.stderr)
print(file_path)
if self.replication_factor:
log.info('setting replication factor to %s on %s',
self.replication_factor, file_path)
# returns a generator so must evaluate in order to actually execute
# otherwise you find there is no effect on the replication factor
for _ in client.setrep([file_path],
self.replication_factor,
recurse=False):
if 'result' not in _:
print(
'WARNING: result field not found in setrep result: {}'
.format(_),
file=sys.stderr)
continue
if not _['result']:
print(
'WARNING: failed to setrep: {}'.format(
_))
except (snakebite.errors.FileNotFoundException,
snakebite.errors.RequestError) as _:
if self.verbose:
print('', file=sys.stderr)
print(_, file=sys.stderr)
if self.verbose:
print('', file=sys.stderr)
secs = int(time.time() - start_time)
print('\nCompleted in {} secs\n'.format(secs), file=sys.stderr)
print('{} files with replication factor 1 out of {} files in {} dirs'\
.format(repl1_count, file_count, dir_count), file=sys.stderr)
class HdfsFileManager(FileManagerBase):
"""A wrapper of snakebite client."""
def can_handle(self, path):
return path.startswith('hdfs://')
def __init__(self):
self._client = AutoConfigClient()
def ls(self, path: str, recursive=False) -> List[File]:
files = []
for file in self._client.ls([path], recurse=recursive):
if file['file_type'] == 'f':
files.append(File(
path=file['path'],
size=file['length']))
return files
def move(self, source: str, destination: str) -> bool:
return len(list(self._client.rename([source], destination))) > 0
def remove(self, path: str) -> bool:
return len(list(self._client.delete([path]))) > 0
def copy(self, source: str, destination: str) -> bool:
# TODO
raise NotImplementedError()
def mkdir(self, path: str) -> bool:
return next(self._client.mkdir([path], create_parent=True))\
.get('result')
def main():
hadoop_conf_dir = "/media/d2/code-sky/dockers/hadoop/etc/hadoop"
os.environ['HADOOP_CONF_DIR'] = hadoop_conf_dir
file_dict = {}
cli = AutoConfigClient()
target_hdfs_path = "/"
for element in cli.ls([target_hdfs_path]):
print("Result: " + str(element))
def ls(hdfs_path, recurse=False, include_toplevel=True, include_children=False):
"""
Parameters:
paths (list) : Paths to list
recurse (boolean) : Recursive listing
include_toplevel (boolean) : Include the given path in the listing. If the path is a file, include_toplevel is always True.
include_children (boolean) : Include child nodes in the listing.
Returns:
(list) path listings with attributes
"""
client = AutoConfigClient()
path_info = list(client.ls([hdfs_path], recurse, include_toplevel, include_children))
return LsObject(path_info)
class HdfsFileManager(FileManagerBase):
"""A wrapper of snakebite client."""
def can_handle(self, path):
return path.startswith('hdfs://')
def __init__(self):
self._client = AutoConfigClient()
def ls(self, path: str, recursive=False) -> List[str]:
files = []
for file in self._client.ls([path], recurse=recursive):
if file['file_type'] == 'f':
files.append(file['path'])
return files
def move(self, source: str, destination: str) -> bool:
return len(list(self._client.rename([source], destination))) > 0
def remove(self, path: str) -> bool:
return len(list(self._client.delete([path]))) > 0
# don't nuke this; hbase uses it for bulk loading.
re.compile("^/tmp/hbase-staging/?"),
# let's try to make sure we're not matching against a top-level path
re.compile("^/[-_.a-zA-Z0-9]+/?$"),
re.compile("cloudera_health_monitoring_canary_files"),
# let's bail out explicitly on anything in our data path
re.compile("^/data/production/?"),
]
if client.test(args.path, exists=True):
for x in client.ls([args.path], recurse=args.recurse_filesystem):
if any(regex.search(x['path']) for regex in donotdelete_whitelist):
logger.info("Matched banned thing, not attempting to delete it: %s", x['path'])
else:
f_timestamp = datetime.datetime.fromtimestamp(x['modification_time']/1000)
if f_timestamp < older_than:
logger.info("I might delete this: %s %s", x['path'], f_timestamp)
if args.actually_delete:
logger.info("Issuing delete of %s", list(client.delete([x['path']], recurse=True)))
if client.test(x['path'], exists=True):
logger.info("Removed %s", x['path'])
else:
logger.info( "I would have deleted this: %s ", x['path'])
else:
logger.warn("%s is not found on hdfs", args.path)
