[Feature request]: Generic Cluster Config

[TimothyWillard]

Label

good first issue, meta/workflow

Priority Label

low priority

Is your feature request related to a problem? Please describe.

Currently we have hard coded knowledge about rockfish and to a lesser extent longleaf throughout the repo. Which makes updating said info or adding new HPC environments painful. Has been described in other detail here: #329 (comment).

Is your feature request related to a new application, scenario round, pathogen? Please describe.

n/a

Describe the solution you'd like

Place config files, maybe yaml or something else, in a central place keyed on environment name. Where these should go is still an open question, viable options are:

1. Place somewhere in this repository, or
2. Create a new repository.

I have a preference for (2) since this repository is already a bit bloated and updates to config would be small and need to be quickly done. Some information that would be helpful to know about each HPC environment in these config files would be:

1. Preferred module versions for gcc, git, anaconda, etc.
2. Some light information the the directory structure (i.e. where to normally expect flepiMoP and a project repo).

Further work would be required to incorporate this info into our various scripts.

[jcblemai]

All of that is very good, thanks. I wonder if we should even go to the route of having this configuration provide the ability to put a user-defined script for each cluster that is sourced at the start of the batch job (in case e.g LD_LIBRARY_PATH changes are made, or to change the shwordsplit for zsh/bash incompatibility.

I'm in favor of having it in this repo, but only slightly. Easier oversight.

[jcblemai]

#191 wanted to link this issue to it as well.