Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

No module named 'fsspec.callbacks' thrown at horovod/spark/common/store.py ln 33 #3450

Closed
zyluo opened this issue Mar 4, 2022 · 6 comments · Fixed by #3451
Closed

No module named 'fsspec.callbacks' thrown at horovod/spark/common/store.py ln 33 #3450

zyluo opened this issue Mar 4, 2022 · 6 comments · Fixed by #3451
Labels

Comments

@zyluo
Copy link

zyluo commented Mar 4, 2022

Environment:

  1. Framework: TensorFlow
  2. Framework version: 2.6.2
  3. Horovod version: 0.24.1
  4. MPI version: 4.1.0
  5. CUDA version: N/A
  6. NCCL version: N/A
  7. Python version: 3.7
  8. Spark / PySpark version: 3.2
  9. Ray version: N/A
  10. OS and version: Ubuntu 18.04
  11. GCC version: 9.3.1
  12. CMake version: 2.8

Checklist: >>>>>>>>>>>>>>>>> all "YES"

  1. Did you search issues to find if somebody asked this question before?
  2. If your question is about hang, did you read this doc?
  3. If your question is about docker, did you read this doc?
  4. Did you check if you question is answered in the troubleshooting guide?

Bug report:
Please describe erroneous behavior you're observing and steps to reproduce it.

  1. Create a GCP Dataproc cluster with 2.0.27-ubuntu18 image
  2. install TF 2.6.2 and Horovod 0.24.1
  3. >> from horovod.spark.common.store import HDFSStore
    output:
    Traceback (most recent call last):
    File "/opt/conda/default/lib/python3.8/site-packages/horovod/spark/common/store.py", line 33, in
    from fsspec.callbacks import _DEFAULModuleNotFoundError: No module named 'fsspec.callbacks'

The fsspec.callback module was introduced in https://github.com/fsspec/filesystem_spec/releases/tag/2021.07.0

The line

spark_require_list = ['numpy', 'petastorm>=0.11.0', 'pyarrow>=0.15.0', 'fsspec']

should be
spark_require_list = ['numpy', 'petastorm>=0.11.0', 'pyarrow>=0.15.0', 'fsspec>=2021.07.0']

@zyluo zyluo added the bug label Mar 4, 2022
@zyluo
Copy link
Author

zyluo commented Mar 5, 2022

fsspec.callback commit id

fsspec/filesystem_spec@47f72bc

@zyluo
Copy link
Author

zyluo commented Mar 5, 2022

@kamalsharma2 FYI

@chongxiaoc
Copy link
Collaborator

@zyluo Would you like to create a PR? Thanks for finding it out.

@zyluo
Copy link
Author

zyluo commented Mar 5, 2022

@chongxiaoc It would take quite a while for me to submit a PR. I'll wait for @kamalsharma2 to create one.

@chongxiaoc
Copy link
Collaborator

@chongxiaoc It would take quite a while for me to submit a PR. I'll wait for @kamalsharma2 to create one.

I will create a one for you then.

@EnricoMi
Copy link
Collaborator

EnricoMi commented Mar 5, 2022

As a workaround you can do:

pip install --force horovod[spark]

This should upgrade all dependencies to the latest available version, hence upgrade fsspec if it is already installed.

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

Successfully merging a pull request may close this issue.

3 participants