Using the GVL as a multiuser elastic slurm cluster

Hello all,

I want to use the cloudlaunch GVL appliance to set up an AWS server that will host Galaxy and a MULTI-USER HPC environment that is intended to be a permanent asset to my company’s comp team. I want users that get created to be able to submit sbatch scripts and take advantage of the resources dynamically. So far I have left a chain of failed “archived” appliances in my wake.

I can get everything to work fine for the ubuntu and research user but I have only had limited success adding new users. The add-cluster-users script that cloudman’s docs say to use does not work. Then I found the GVL ansible playbook on github and managed to get a new user created with all the nice galaxy links and home dir set up. But this new user can not use slurm and I cant ssh into the worker nodes.

I want this server to basically mimic the experience of the academic HPC clusters that many of us are used to.

Can anyone help me find the right set of tutorials to do this? I feel like I am so close.

I am not married to the GVL via cloudlaunch mechanism, it just seemed to promise exactly what I was looking for without me having to become an expert in a couple different areas first. But if the best way to get what I want lies somewhere else, I am glad to explore other options.


1 Like

Hi Gus,
Sorry to hear about the troubles. Try the following repo:

The instructions in the readme show how to get going; I’ve just tested and updated the scripts so the script runs to completion on a fresh launch of the GVL 4.4.0.

Hopefully that gives you the functionality you’re looking for and let us know if you have more questions.

1 Like

Thank you for this. I have found that tool very useful and it did indeed solve the problem.

1 Like