Skip to content

Inconsistent WebHDFS path handling over Hadoop versions #193

Description

@MaicoTimmerman

Given the following test-case:

# test_client.py
class TestAppend(_IntegrationTest):
  def test_simple_with_plus(self):
    self.client.write('ap+q', b'hello,', overwrite=True)
    self.client.write('ap+q', b' world!', append=True)
    eq_(self._read('ap+q'), b'hello, world!')

We are failing on hadoop 3.1.1 with the following error:

  File "./hdfs/test/test_client.py", line 274, in test_simple_with_plus
    self.client.write('ap+q', b' world!', append=True)
  File "./hdfs/hdfs/client.py", line 501, in write
    res = self._append(hdfs_path, buffersize=buffersize)
  File ".//hdfs/hdfs/client.py", line 118, in api_handler
    raise err
Exception: File /user/<user>/.hdfscli/ap+q not found.

and succeeding on 3.1.4, due to changes made in:

As upgrading HDFS is a non-trivial operation, I'm wondering if this client library would be able to special-case this inconsistency to allow for smooth upgrading over HDFS versions.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions