From 51acef802eb1de0b32673191b9b42d9f12a709c6 Mon Sep 17 00:00:00 2001 From: Michael Weisner Date: Fri, 16 Jan 2026 12:44:45 -0500 Subject: [PATCH] Update 07_jupyter_with_conda_singularity.mdx updated instructions for users to start compute jobs before compiling environments and packages. --- .../09_ood/07_jupyter_with_conda_singularity.mdx | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/docs/hpc/09_ood/07_jupyter_with_conda_singularity.mdx b/docs/hpc/09_ood/07_jupyter_with_conda_singularity.mdx index 819f095f7d..970c1892f2 100644 --- a/docs/hpc/09_ood/07_jupyter_with_conda_singularity.mdx +++ b/docs/hpc/09_ood/07_jupyter_with_conda_singularity.mdx @@ -55,6 +55,16 @@ The above code automatically makes your environment look for the default shared ::: ### Prepare Overlay File +First, start a job to work on the compute nodes. +```bash +[NetID@log-1 my_env]$ sbatch --cpus-per-task=2 --mem=10GB --time=04:00:00 --wrap "sleep infinity" + +# wait to be assigned a node then SSH to the node +``` +Once you SSH to the node, you can begin setting up the environment. + +:::note Files will not properly compile on the login nodes as they have different Red Hat OS images than the compute nodes. All compilation and package installation must be done on the compute nodes. + ```bash [NetID@log-1 ~]$ mkdir /scratch/$USER/my_env [NetID@log-1 ~]$ cd /scratch/$USER/my_env @@ -143,9 +153,9 @@ To install larger packages, like Tensorflow, you must first start an interactive ```bash Singularity> exit -[NetID@log-1 my_env]$ srun --cpus-per-task=2 --mem=10GB --time=04:00:00 --pty /bin/bash +[NetID@log-1 my_env]$ sbatch --cpus-per-task=2 --mem=10GB --time=04:00:00 --wrap "sleep infinity" -# wait to be assigned a node +# wait to be assigned a node then SSH to the node [NetID@cm001 my_env]$ singularity exec --fakeroot --overlay /scratch/$USER/my_env/overlay-15GB-500K.ext3:rw /share/apps/images/cuda12.3.2-cudnn9.0.0-ubuntu-22.04.4.sif /bin/bash