Open mrakitin opened 5 years ago
The labextensions seem to be enabled correctly:
$ jupyter labextension list
JupyterLab v0.35.5
Known labextensions:
app dir: /opt/conda_envs/collection-2019-3.0-amx/share/jupyter/lab
@jupyter-widgets/jupyterlab-manager v0.38.1 enabled OK
jupyter-matplotlib v0.4.2 enabled OK
We basically tried to follow https://github.com/bluesky/tutorial/blob/master/binder/postBuild, but with the wrong order (@jupyter-widgets/jupyterlab-manager
wasn't installed first, but was installed eventually, and then everything was rebuilt.)
That's how the server is started from the collection-2019-3.0-amx
conda environment (via st.cmd
and a systemd service):
jupyter lab --no-browser --notebook-dir=/epics/iocs/notebook/notebooks/ --ip=0.0.0.0 --port=17000 --NotebookApp.token='' --debug > $jupyter_log 2>&1
The relevant kernel config is:
{
"argv": ["/opt/conda_envs/collection-2019-3.0-amx/bin/python","-m","IPython.kernel","-f","{connection_file}","--profile","collection"],
"display_name": "Data Collection (2019-3.0)"
}
Here are some debug logs from Firefox web console:
A side note about puppet-managed /opt/conda/.condarc
file: https://controlsweb.nsls2.bnl.gov/trac/ticket/3627 (that was observed with @jklynch during the deployment visit).
@danielballan, what versions of the libraries would you suggest?
I suggest upgrade to JupyterLab 1.0.x and trying again.
Thanks @danielballan! We will work with @jklynch on that this Thursday afternoon. Will report back after that.
This was partially resolved by creating a new environment, installing all the latest jupyterlab and friends and by changing 00-base.py
to use the configuration from https://github.com/bluesky/tutorial/blob/master/scripts/beamline_configuration.py (avoiding nslsii.configure_base(...)
) via https://github.com/NSLS-II-AMX/profile_collection/commit/ab426a41cc37918e51e0c8d394b5d75e1bc6235e.
However, we found a common issue also observed in our tutorials -- see https://github.com/bluesky/tutorial/issues/37 for details.
Tried to figure it out with @jklynch on 7/10/2019, were able to enable the live plotting in the classic notebook, but it still failed for the JupyterLab.
Here is the environment on
xf17id2-ca
:attn @danielballan