Building the image would involve copying the "static" data from the repo into the "data" dir in the image. The dynamic data would be mounted as sub-directories of the static data directories the same way it's currently done for dynamic legends (legends/dynamic/.)
This would make it clearer that the data you see in the repo directory is the set of data managed by version control, not stuff that could change on a daily basis.
It's possible some or all of our static data may go away in favor of an approach that transfers the whole configuration from the supercomputer. Or a shared database, but I doubt we'll have the time to develop something that robust.
Building the image would involve copying the "static" data from the repo into the "data" dir in the image. The dynamic data would be mounted as sub-directories of the static data directories the same way it's currently done for dynamic legends (
legends/dynamic/.
)This would make it clearer that the data you see in the repo directory is the set of data managed by version control, not stuff that could change on a daily basis.
It's possible some or all of our static data may go away in favor of an approach that transfers the whole configuration from the supercomputer. Or a shared database, but I doubt we'll have the time to develop something that robust.