Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

checkout: partial directory cache

See original GitHub issue

Currently, if, for some reason, you’ve lost some of the cache files for a part of your directory, dvc will throw assertion error:

ERROR: failed to pull data from the cloud
------------------------------------------------------------
Traceback (most recent call last):
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/command/data_sync.py", line 46, in do_run
    recursive=self.args.recursive,
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/repo/pull.py", line 27, in pull
    target=target, with_deps=with_deps, force=force, recursive=recursive
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/repo/checkout.py", line 54, in checkout
    stage.checkout(force=force, progress_callback=progress_callback)
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/stage.py", line 822, in checkout
    force=force, tag=self.tag, progress_callback=progress_callback
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/output/base.py", line 228, in checkout
    progress_callback=progress_callback,
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/remote/base.py", line 370, in checkout
    progress_callback=progress_callback,
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/remote/local.py", line 353, in do_checkout
    self.link(c, p)
  File "/var/akonshin/.virtualenvs/tr/lib/python3.5/site-packages/dvc/remote/local.py", line 155, in link
    assert os.path.isfile(cache)
AssertionError
------------------------------------------------------------

This is because unlike for regular data files, we don’t check for cache file existence before linking it. With standalone data files we print a warning that “cache file doesn’t exist and file is not going to be created”, so we need to do something similar here.

Some backstory: user on ODS was getting this error and turned out that he had ran dvc gc -c for some time and then interrupted it, so parts of data were removed and so when he tried to dvc pull later, he got this error.