Lmod has detected the following error: The following module(s) are unknown

[rgas20]

All,

Having one more issue with loading modules on a new instance of Open OnDemand.

When I load a Jupyter notebook for instance:

Script starting…
Waiting for Jupyter Notebook server to open port 38242…
TIMING - Starting wait at: Fri Feb 19 15:04:51 CST 2021
TIMING - Starting main script at: Fri Feb 19 15:04:51 CST 2021
Lmod has detected the following error: The following module(s) are unknown:
“python/3.7”

Please check the spelling or version number. Also try “module spider …”
It is also possible your cache file is out-of-date; it may help to try:
$ module --ignore-cache load “python/3.7”

Also make sure that all modulefiles written in TCL start with the string
#%Module

This application template has been taken from a working instance of OOD and copied over to a fresh setup of OOD, looking through the docs I seem to have all the bases covered. In the form.yml we are asking it to load the python/3.7 module. I’ve also verified that the reverse proxy is working correctly from the new instance and tested it as a prerequisite to setting up the apps. When I check the submit node on our cluster, I can see the job starting in the queue (then failing) so it doesn’t seem to be a limitation in that regard. I’m not sure what to check next however or what else may be missing. Any help is appreciated.

[jeff.ohrstrom]

Does that module, python/3.7 actually exist on the compute node? There’s not much too this issue, proxies and all that. It’s an issue confined to the job itself. I’d hop on a compute node and verify that that module does indeed exist (through module load or module spider commands)

It seems like the module doesn’t in fact exist on the compute node. When you say you’re spinning a new OOD instance, is this new instance also interacting with new compute infrastructure?

I just ran this on our Slurm infrastructure to sort of replicate the issue and triage. This would create a job and just spit out all the modules.

sbatch --export=NONE --wrap 'module spider'

I’m guessing that you won’t see python/3.7 even if you submit this very simple job. Though if you can infact see it there and not through OOD - let me know because that would be very strange and indeed we’d have to look more at it.