DASK distributed import Client error (ImportError: cannot import name future_set_exc_info)
See original GitHub issueHi! Along the line of “Import error in dask-scheduler
#1120”, there is another problem. I use Datalab instance on Google cloud platform (GCP). For DASK installation I use the following command in a cell:
!pip install dask[complete] distributed --upgrade
Installation goes well and I can verify it:
!dask-scheduler
which produce an output:
distributed.scheduler - INFO - ----------------------------------------------- distributed.scheduler - INFO - Clear task state distributed.scheduler - INFO - Scheduler at: tcp://172.17.0.2:8786 distributed.scheduler - INFO - Local Directory: /tmp/scheduler-XAWu9a distributed.scheduler - INFO - -----------------------------------------------
On the other hand, I get
ImportError: cannot import name future_set_exc_info
pointing to
100 from tornado.concurrent import (Future, is_future, chain_future, future_set_exc_info, 101 future_add_done_callback, future_set_result_unless_cancelled)
when I call from distributed import Client
…
Any ideas how to solve this???
Issue Analytics
- State:
- Created 5 years ago
- Comments:7 (3 by maintainers)
Top GitHub Comments
It seems like we found rather stable solution which doesn’t interfere with existing datalab dask installation. We run an external script doing:
!pip install tornado==4.5.1 distributed==1.21 dask-ml[complete]
and restarting current server on GCP and then verifying installation. After that all packages need for development are available. It’s not a best way of solving the problem, but it works for us.Dask.distributed requires tornado 4.5 or greater. But no, the error that you’re seeing is likely something different. I recommend looking through datalab docs on how to install or upgrade libraries in your user environment. They should have a nice solution to this somewhere, but unfortunately I’m not very familiar with their service.
On Wed, Apr 18, 2018 at 9:07 AM, NatLun091238 notifications@github.com wrote: