Closed havok2063 closed 2 years ago
The --port=8000
argument is not shared with the pre heated kernel, it can be set by using env vars or voila.json
: https://voila.readthedocs.io/en/stable/customize.html#partially-pre-render-notebook
Or try it on the default port
I get the same error when I use the VOILA_APP_PORT envvar in my config. I've checked in a config file in the standalone repo. I also tried switching everything to the default port and I also get the same error.
Also when I set the VOILA_APP_PORT to "8000", voila still seems to start on the default port 8866. But voila --Voila.show_config=True
does not throw any errors and looks correct to me, compared to what is in the docs.
Turns out I was still using an older version of Voila in my testing. I fixed this issue in #9, which is released as v0.0.5.
Also when I set the VOILA_APP_PORT to "8000", voila still seems to start on the default port 8866. But voila --Voila.show_config=True does not throw any errors and looks correct to me, compared to what is in the docs.
The env variable only appears to set the port to use inside the kernel to contact Voila, so they should both be set. This works for me:
VOILA_APP_PORT=8000 voila --no-browser --template=embed --enable_nbextensions=True --Voila.tornado_settings="{'allow_origin': 'http://localhost:8080', 'allow_credentials':True}" --port=8000 --preheat_kernel=True --pool_size=3
ok, your patch plus the port fixed things. It's now working! Thanks Mario!
Testing this out with voila 0.3.4 and the new preheated kernels code. I'm getting the following error in the console.
with the following voila command. I see this both in the standalone example https://github.com/havok2063/standalone_jdaviz/, as well as our MAST-specific setup.
This is without any changes to the notebook. When I add the
wait_for_request
method to the notebook, just after the cell with the imports, I get a different error in the voila server, same as the one mentioned here, https://github.com/voila-dashboards/voila/issues/1123.