[Bug]: Errors Using hpc_install_or_update.sh On UNC's Longleaf Cluster

[anjalika-nande]

Label

bug

Priority Label

high priority

Describe the bug/issue

Hi @TimothyWillard , I'm trying to set up longleaf and when I run the line source hpc_install_or_update.sh longleaf it asks if I want to install flepiMoP and if I say yes then it gives the following error and logs me out of longleaf. Guessing it has to do with the github access key or something but not sure where to input this?

Did not find flepiMoP at /work/users/a/n/anande/flepiMoP, do you want to clone the repo? (y/n) y
Cloning on your behalf.
Cloning into '/work/users/a/n/anande/flepiMoP'...
Warning: Permanently added 'github.com' (ECDSA) to the list of known hosts.
[email protected]: Permission denied (publickey).
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.
Connection to longleaf.unc.edu closed.

To Reproduce

No response

Environment, if relevant

No response

[pearsonca]

I think this can be a future PR against the install script. I suspect best we can do here is catch the error and translate it to a more informative message (likely akin to: "Did you add your id_rsa.pub to github => settings => etc?". Maybe emit the public key for easy copy-paste?)

[TimothyWillard]

This is still in development in GH-329. And I think these steps that you are using are out of date. What you must do to use in the meantime is clone the flepiMoP repository yourself, switch to the correct branch, and then run the script as an executable (although, I expect these to change again soon). I suspect this has more to do with your git setup on longleaf then the script itself. Could you try cloning another git repository to see what happens? Can you try following the steps described here: https://github.com/ACCIDDA/ACCIDDA-KeyDocAndDiscussion/wiki/HPC-Setup.

[anjalika-nande]

yep, following those steps now!

[anjalika-nande]

I did the key generation and now it does try to pull flepi but I get this error:

EnvironmentFileNotFound: '/work/users/a/n/anande/flepiMoP/environment.yml' file not found

[TimothyWillard]

Yeah, that sounds like the error I expected. To use this installation script right now you have to clone the flepiMoP repository yourself, switch to the correct branch, and then run the script as an executable. Although, I think that will change today as we are in the process of merging that PR.

[anjalika-nande]

okay, thanks!

[anjalika-nande]

Hi Tim, I saw that you merged with main so I tried repeating the process (editing all the file paths accordingly) but I still got this error:
EnvironmentFileNotFound: '/work/users/a/n/anande/flepiMoP/environment.yml' file not found

I checked in /work/users/a/n/anande/flepiMoP/ folder and it didn't have the file. Not sure why when cloning happens in the hpc_install_or_update.sh file, this file is missing? Anyway I git pulled again from the folder and then all the files got added and it worked.

[TimothyWillard]

Was the clone of flepiMoP at /work/users/a/n/anande/flepiMoP/ placed there by the hpc_install_or_update.sh script or did you clone that yourself? If it wasn't placed there by the script it sounds like maybe the clone was just behind main.

[anjalika-nande]

It was placed there by the hpc_install_or_update.sh script!

[TimothyWillard]

I see, that is concerning, but I'm unable to reproduce the issue with the details given. I've attached what I did as an example of what is expected. Could you try following https://iddynamics.gitbook.io/flepimop/how-to-run/advanced-run-guides/running-on-a-hpc-with-slurm, and then report the output of the error you encounter?

longleaf_install.txt

[anjalika-nande]

Thanks! Since I have it working now I'm just running some RSV related simulations as it's due next week and I'm traveling tomorrow. Is it okay if I get to this later?

[TimothyWillard]

@anjalika-nande have you had a chance to follow up on this or had any more issues with this install script?