Stream: dask

Topic: Error importing dask_jobqueue after updating xarray package

view this post on Zulip Mira Berdahl (Jan 12 2024 at 19:29):

Hi, I just updated my xarray package, after which I started a new session on JupyterHub and am receiving a new error when trying to import PBS Cluster

Here is what I get:

 dask_jobqueue import PBSCluster
---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
Input In [1], in <module>
     13 from distributed import wait
     14 #from ncar_jobqueue import NCARCluster
---> 16 from dask_jobqueue import PBSCluster
     17 import dask

File /glade/work/mberdahl/miniconda/envs/pangeo/lib/python3.9/site-packages/dask_jobqueue/__init__.py:3, in <module>
      1 # flake8: noqa
      2 from . import config
----> 3 from .core import JobQueueCluster
      4 from .moab import MoabCluster
      5 from .pbs import PBSCluster

File /glade/work/mberdahl/miniconda/envs/pangeo/lib/python3.9/site-packages/dask_jobqueue/core.py:20, in <module>
     18 from distributed.deploy.local import nprocesses_nthreads
     19 from distributed.scheduler import Scheduler
---> 20 from distributed.utils import tmpfile
     22 logger = logging.getLogger(__name__)
     24 job_parameters = """
     25     cores : int
     26         Total number of cores per job
   (...)
     63         Name of Dask worker.  This is typically set by the Cluster
     64 """.strip()

ImportError: cannot import name 'tmpfile' from 'distributed.utils' (/glade/work/mberdahl/miniconda/envs/pangeo/lib/python3.9/site-packages/distributed/utils.py)

Thoughts? Thanks

view this post on Zulip Katie Dagon (Jan 12 2024 at 22:41):

My initial thought would be to try updating dask and/or dask-jobqueue. Sometimes xarray updates have breaking changes. Looking at the versions of things in your environment can be helpful too, I usually do this with conda list.

Last updated: May 16 2025 at 17:14 UTC