Issue on Starting HPC Dask clusters with dask-jobqueue

[cxiao13]

Dear tutorial organizers,

I am trying to reproduce the results in 05-dask-hpc.ipynb using NCAR HPC JupyterHub on Casper.

When running

# Block progress until workers have spawned (typically only in demos and benchmarks!)
client.wait_for_workers(2)

The kernel was running forever and did not respond.

I did not change any YAML config file.

Do you have any idea what could be the reason? Thank you!

[vanderwb]

Hi @cxiao13. You probably need to change interface = 'ib0' to interface = 'ext'. This is something that has changed on Casper since we first created this tutorial. I am going to push a commit that will fix this for anyone else also using this tutorial.