Kernels on ARE with xp65

Hi all,

I am trying to select one of the conda environments on ARE with xp65, such as analysis3-25.07.

I specified /g/data/xp65/public/modules/ in the setting of Module directories conda/analysis3 in the setting of Modules. I still cannot find the NRI environments.

I took a look at Running xp65 analysis on ARE, and I thought there may be issues with .bashrc, .bash_profile, or .condarc, but I’m not sure how to make the necessary changes. Can you help me with?

Here are the .bashrc, .bash_profile, and .condarc,

.bashrc

.bash_profile

.condarc

Thanks

Hi @chenhui.jin,

Have you also added gdata/xp65 in the storage field? That is also required for the xp65 folder to be visible by the ARE Jupyterlab instance.

For more information, please refer to conda/analysis3 Python Environment on the Hive Docs.

Cheers
Davide

Hi @atteggiani,

Thanks for your reply. Yes, I added gdata/xp65 in the storage setting, but it didn’t show kernel options last week.

I just started a jupyter session this morning, and it works now.

1 Like

Hmm this was an odd one, i just had a similar issue. My notebook was using analysis3-25.07 and then I shutdown the kernel to reset my Dask client and suddenly all the xp65 kernels vanished. I had to restart my ARE session to get them back.

1 Like

The issue persisted for a week for me; it’s just back to normal today.

Huh this has happened again to me this morning and I now can’t get it to see any xp65 kernels

Ok so what I think has happened is yesterday my notebook was using analysis3-25.08 and that is no longer available in `/g/data/xp65/public/apps/med_conda_scripts/`. This seems to have made my session panic and decide there are no usable kernels in that path.

This happened to me this afternoon when trying to load the conda/analysis3 modules in an ARE JupyterLab instance. I’ve since tried 25.04, 25.07, and 25.08 with no luck. I had been using the default for hours with no issues until I restarted the kernel and everything disappeared.

Edited to add that I can’t seem to launch a JupyterLab instance from ARE using the analysis3 modules now :thinking:

I have also encountered the same issue this afternoon!

Any issues encountered in the last couple of hours (@meharas & @spencerC) should be resolved now - @rbeucher has fixed some issues with the conda environment.

Issues from yesterday backwards are probably still extant for now.

2 Likes